Back to: Five reasons to envy the French | Forward to: Mediocrity

Insufficient data

There's a deceptively simple question that's been bugging me this week, and it is this:

What is the minimum number of people you need in order to maintain (not necessarily to extend) our current level of technological civilization?

There are huge political ramifications hiding behind this question. Let me unpack them for you.

Conservative politicians in the US — and elsewhere — get a lot of mileage from appeals to false nostalgia, to a yearning for a time when things were simpler, everyone was sturdily self-sufficient or knew their place (or both), and government was small (sometimes small enough to drown in a bathtub). Nostalgia trips manifest themselves in all sorts of curious places. In SF (the literary field I know most about) we have the perennial libertarian/space colonization nexus.We have Ayn Rand, and her wish-fulfilment nerd fantasy of a world sustained by a tiny, overworked minority of geniuses who, if only they could demand a level of rewards corresponding to their work, would be rich beyond dreams of avarice (and able to make the trains run on time). Outside it, we have the peculiarly rustic aspirations of the green fringe, who'd like to see a world of five million or so pre-industrial humans living in harmony with nature. In the Republican party of the United States we see rhetoric couched in hatred for "big government", and among the UK's conservatives we see an almost masochistic addiction to cuts in public spending framed with calls for a big society in which many current government services will be delivered by voluntary citizens groups instead.

I think these ideas are mostly delusional because they rely on a fundamental misapprehension about the world around us — namely that we live in a society that can be made simple enough to comprehend.

Let's take a look at the superficial structures around us. How many people does it take to design a new automobile? Back in Henry Ford's day, it needed an office full of draughtsmen, a handle of senior engineers to sort out each major mechanical subsystem (gear train, engine, electrics, brakes, suspension, bodywork), and experts on coachbuilding to dictate the shape of the bodywork. There would be time and motion men to dictate the speed and sequence of assembly line activities, and more drafting work to design the tools the production line workers would use ... it took the effort of a few hundred men.

But modern cars are different. A typical 2010 automobile may contain roughly 20-30 electric motors and actuators (for everything from the central locking system to the air conditioning and the motorized seats and windows). There's a similar number of microprocessors involved in everything from the engine and gearbox management systems to the entertainment, navigation, communication, and accident mitigation systems (for example, the sensors and microprocessors that control the sequence of pyrotechnic detonators that inflate air bags, tension seat belts, and collapse the steering column in event of a collision). The in-car electronics alone require on the order of 10-20 million lines of code to run all these services — which implies the combined efforts of thousands of software developers, never mind the small army who design not only the body panels but the handling tools the production line robots use to install them. Cars are no longer user-serviceable because they're nearly as complex as 1960s airliners.

And as for your smartphone? The damned thing has a component count somewhere between ten major subsystems and frame components and a hundred billion (if you go down to the smallest scale and count the capacitors in its FLASH memory). The number of fab lines on the planet that can make memory chips of that density is limited, and they rely on rare elements mined only in exotic locations and in tiny abundance.

Medicine: let's not go there. Back in the late 19th century, we had doctors, nurses, surgeons, pharmacists, and dentists. Today, each of those professions has exploded into platoons and battalions of sub-specialities, and their roles are supported in turn by complex industries full of strange niches.

Around 1900, it took the effort of about 20-30% of a nation's work-force to provide food for everybody; and another 30-50% working in factories to produce clothing, machinery, and processed materials like bricks and billets of pig iron. Today, we only need 0.5-1% of the work force to feed everyone, and another 1-4% working in industry to produce the basics — but the microspecialities have exploded, to the extent that a lot of our needs seem to require a trans-national economy to provide. There are only two vendors of wide-body airliners on any scale today, Boeing and Airbus, and both of them are effectively multinational consortia (more than half the components of the Boeing 787 Dreamliner are produced overseas, and shipped to Seattle for final assembly). There seems to only be room for one vendor of super-Jumbo airliners — if Boeing and Airbus tried to exploit that niche simultaneously, they'd both starve — so they appear to be avoiding conflict in that (and some other) area(s). And so on.

So. I ask: how many people does it take, as a minimum, to maintain our current level of technological civilization?

I'd put an upper bound of about one billion on the range, because that encompasses basically the entire population of NAFTA and the EU, with Japan, Taiwan, and the industrial enterprise zones of China thrown in for good measure. (While China is significant, more than half of its population is still agrarian, hence not providing inputs to this system).

I'd put a lower bound of 100 million on the range, too. The specialities required for a civil aviation sector alone may well run to half a million people; let's not underestimate the needs of raw material extraction and processing (from crude oil to yttrium and lanthanum), of a higher education/research sector to keep training the people we need in order to replenish small pools of working expertise, and so on. Hypothetically, we may only need 500 people in one particular niche, but that means training 20 of them a year to keep the pool going, plus future trainers, and an allowance for wastage and drop-outs by people who made a bad career choice. Higher education accounts for 1.8-3% of gross spending in the developed world, with primary and secondary education taking a whopping chunk on top of that (if you spent 10 years in a school with a staff:pupil ratio of 1:10, then you soaked up a person-year of time; there may be more labour going into pre-university education than goes into agriculture and industry combined).

As to those political implications ...

Firstly: no, you can't simplify a complex society that runs on just-in-time delivery and a host of specialities. You need a huge training back-end to provide for the thousands of skilled graduate-entry niche occupations. You need an efficient just-in-time delivery system to keep everyone supplied with food, water, power, shelter and whatever else they need — it's that, or accept huge inefficiencies in your supply chain that wipe out the gains produced elsewhere.

Secondly, seemingly similar artefacts (cars, phones, airliners) have invisibly accreted complexity. The complexity makes them better (safer, more economical, more luxurious) than their predecessors, but vastly more difficult to engineer; stuff that used to be fixable by shade-tree mechanics and jobbing electricians has receded over the horizon. Back in the early 19th century, the complement of a sailing ship could expect to maintain the ship in every significant way using tools and expertise that they could carry aboard the ship. Today in the early 21st century, that's not an option with airliners or probably even automobiles.

Thirdly, the complexity embodies in these new products means that their production is dependent on a complex web of lower-level specialities.

Fourthly, there are more side-effects to keep track of. Exotic materials mean exotic contamination events from waste dumping, for example.

Fifthly: space colonization? Get back to me when you've tracked down how many people it takes to design and build a space suit. (The number is in the hundreds, if not the thousands.) More realistically, we won't have autonomous off-world colonies unless and until they can cover all the numerous specialities of the complex civilization that spawned the non-autonomous, dependent-on-resupply space program. Or, to put it another way: colonizing Mars might well be practical, but only if we can start out by plonking a hundred million people down there.



Very good piece, as always. However, I think there's a potential flaw in one of your premise statements "While China is significant, more than half of its population is still agrarian, hence not providing inputs to this system" - if the agrarian portion of the population is predominantly self-contained, essentially growing food only for itself, then I can see why it should be excluded. If, as I think may the case to a significant degree, the agrarian community is in fact predominantly responsible for the calorific maintenance of those part's of China's population that are contributing to "our current level of technological civilization", i.e. feeding them, then shouldn't they also be considered as proving "inputs to this system"?


I don't understand why you think there needs to be a large government for any of this to happen.

Except for Airbus, that is! ;)


One question: How do you count people who will, say, design part of a space suit for two months, switch to anesthesia equipment for another three and busy themselves with spay-painting nozzles the rest of the time?


I suppose we can't cheat, and fire everyone in charge of "designing newer better" versions of things and just keep around the people we need to maintain the infrastructure we have and fab copies of parts that go bad using existing schematics?


Wow. First commentor opportunitys don't happen very often.

You've summed it up quite neatly. I can't envision the 100 million person space lift, and even staging it in a million sequences sounds lame.

Rustic pioneers reproducing and supporting themselves at the Wild Wild West level - we'll need martian cattle.


Old British Empire colonies were not self-sufficient. Why would a colony in our solar system have to be?


A lot depends on what you mean by those words. It's not clear right now, for example, that we're actually maintaining our current level of civilization (look at all the infrastructure that goes unmaintained, and the expertise that's already been lost with regard to fixing or making more of it). But it's also unclear what that current level has to include -- would a world without fart apps be insufficient to meet your definition (or do fart apps essentially come for free once you have the structures to do the stuff you do need?). What about a world without waterproof mascara? Plug-in air fresheners? Drugs for made-up diseases? You might end up with a world at the same nominal "level" but very differently organized.

Along those lines, we live in a world where it takes relatively few people to build and/or design tweaks on stuff that's been designed and built before. You could imagine a brittle world at our current level (what? more brittle than the one we've got) with way fewer people in the design/research/creation side and rather more on the archiving/forensics/retraining side. How long such a culture could actually hold together is obviously a whole different question. Ours is more about finding things for all those people to do so they can be paid, rather than having a set of things to do and staffing up to meet that demand.


I find this a scary concept so I tend not to think about it much! I'm pretty sure that's not the answer.


I think you're largely right in most of the points you make in this excellent piece; however, I suspect your argument may work against rather than for your conclusion. As I see it, it is precisely because the system is so complex that big government can't administer it: for this type of top-down management to be effective, its practitioners need an extraordinary level of expertise in the administered fields––and as you point out, each of these fields represents a specialism that is only acquired with a lot of effort and time.

So, for these reasons, I'd argue that sectoral autonomy is a necessary outcome of growing complexity. Of course, the recent unpleasantness occasioned by the buccaneers of high finance have made the drawbacks of such autonomy all too obvious ...


Does Mars need to be able to design a Lexus, or just build them for the local population?

Would our quality of life really take much of a hit if we threw out all the gizmos that are in a 2010 Lexus, but not a 1969 Chevy?


One word: "externalities".

Organizations working within a market are absolutely terrible about handling externalities. But much of the support infrastructure for a market economy requiring numerous specialities consists of externality-handlers. Who builds and manages the road and transport systems, for example, or the schools, or the hospitals, or the nursing homes for the elderly, or, or, or ...

These functions may or may not be part of the legitimate role of government, but they're sure as hell not something we can survive without, and they require regulation, management, and subsidies whether or not they're funded directly by government or, I dunno, pinko socialist tooth fairies.


Or, to put it another way: colonizing Mars might well be practical, but only if we can start out by plonking a hundred million people down there.

I suddenly seem to recall a pair of novels that you wrote ... ;)


Your own answer to your question looks pretty good to me. If you were designing a minimal self-sufficient colony, and you were willing to spend orders of magnitude more manpower on the design process than on the colony itself, you could probably reduce it into the ten million range. Standardize, e.g., software and machine tools to eliminate extra redundancy. For a concrete example, I'm sure a colony doesn't need 45 different Ethernet chip designs.

How about a related question? Using early 21st century technology and hindsight to design and prepare, what is the smallest self-sufficient colony capable of maintaining, not extending, early 19th century technological civilization?

The colony spaceship is two centuries away, so anyone outfitting it for 21st century technology would be a historical re-enactment group. (-:


An interesting question. Your calculations are flawed though - they're based on the assumption that the way we currently do things is the most manpower-efficient one possible, which simply isn't true.

In fact, quite the opposite applies, given the huge political pressure to create jobs and provide "competitive advantage". To some extent, our society attempts to ensure that the number of people working on these things is exactly the number of people available to work. Where this cannot be accomplished by increasing production, it is often done by decreasing efficiency (adding more bureaucracy is a popular approach).

It is quite reasonable to assume that if we had less people available, we would improve the design of things to require less people to produce and maintain them. Currently, this is not something most products are designed for.

I'm fairly sure Asimov was (one of) the first to make this point, in the Foundation series.


I've been thinking about the myriad ways a colony could reduce redundancy (fewer competing versions of any industry/product/tool/component), and I wonder if you're waiting for your readers to prescribe the Soviet command economy, just so you can shoot it down. (-:

You did say, "maintain, not extend technology". Which means the command economy might be good enough.

BTW, last year you proved that a space ship needs to be the size of a continent to survive in space; this year you're proving that you need a continent's worth of people riding on it. Those match up nicely, don't they?


I've been having a similar frustration with the whole Dark Mountain / uncivilisation crowd prevalent with The Idler set (yet more metropolitan broadsheet wealthy retirees), Alan Moore and Copey.

One particular thing that annoyed - the vision of someone playing Street Fighter on a solar or bicycle powered (I forget which) Super Nintendo, because it epitomized so much that was ill thought out - for starters, a SNES requires vastly more energy to run than any of today's low power miracles.

Arguably, the total energy required by the infrastructure industries to create a product like a modern smartphone is probably a lot more than was required to produce the NES, or early 8-bit garage computers, but there is no denying that in a post-civilized state you'd be a lot better off with a low-power portable system.

Secondly - it was the implication of 'no new software' that is expressed in that thought. Not 'some kid showing a game he'd written on his bicycle powered computer'. We will perform plays and sit around campfires singing songs and listening to stories, in a world that seems, in an endless festival organised by the Mutoid Waste Company.

Where, I suspect, the storytellers and singers currently charging people to go to their Dark Mountain festival, etc, will somehow expect themselves to be exempt from spending 70% of their time farming enough food to stay alive, because they're doing something useful.

There's also something a bit Pol Pot about it all.

Although I love the Guardian article I linked having done a search on them.


Actually you are incorrect in re colonizing Mars requiring 100 Million population. Per your model that is the amount necessary for the colony to reach self sufficiency at current technological levels with current levels of comfort/ safety/ lifestyle. OTOH, if you allow distance learning, and the like and are willing to eliminate certain sectors as necessities( i.e. Civil aircraft, or the like) and/or accept a less comfortable or more archaic lifestyle in some areas you may be able to reach a stable society at a lower population level( but there are limits -- you have to have a means of getting material/ people from place to place, and you may be making safety/cost/ comfort tradeoffs that are pretty lame -i.e. primitive unsave polution intense but easily maintained/ manufactured vehicles are perfectly valid for colonial use), and if you allow periodic aid/ resupply from home, you can drop it even lower( but in the resupply case you have not achieved an autonomous colonization, merely a foothold).

My Guess is that the stable but sucky autonomous colony( think colony no one wants to live in, but within which life is practical, I'd guess you may be able to be stable at soewthing between 1/2 and 1 order of magnitude lower than the full deal population level.


I've had a similar discussion with other people, and have come to a similar conclusion. A minimum of 100 million would be needed to maintain our civilization, and you'd most likely end up with a lot less variety in goods and services (though the other direction is possible).

As you go back in time, fewer people are needed -- 1950s civilization would need fewer and 1900 fewer still. Interestingly, I think there might be a similar trend going forward -- automation increases future society may we be able to be maintained by a smaller and smaller number of people.

Your observations on complexity are interesting, as it happens that sort of thing is part of the reason I became a libertarian. I think our society is too complex to manage, so leaving things up to individuals and the operation of the free market is the best course to take. Centralized planning leads to mistakes which lead to more planning and more wasted resources.


Another good critique from New Statesmen.


Darn you for a thought provoking post during my lunch hour.

Thank'ee kindly.

Michael, let me ask you this - which vehicle would you rather be in a serious car accident in? The '10 Lexus? or the '69 Chevy?

Alright, back the larger topic. I guess I'm morbid, but I've wondered about things like this. Like, 'What if all the nations of the world were hurled to their own separate alternate Earth? What would happen to those people?'

My general answer is, it isn't pretty. Supply chains are irrevocably smashed and standards of living slide. Individual liberties (where applicable) probably take a hit as well thanks to forced draft industrialization and colonization.

What's more, I'm a US citizen in flyover country (Mississippi) and I'm scared shitless over the idea of the New Madrid fault breaking loose. Not because it will topple buildings where I live, but because it will make Katrina look like a dress rehearsal for Hell. Why? Well, I realize just how fragile our modern technological society is.

Memphis, the major population center likely to be impacted. That's a shipping hub. The Mississippi River is unlikely to be navigable for months if not years. I-55, a major north south corridor, will probably have large stretches closed for over a year. Expect large batches of the power grid to be screwed as well. And this doesn't even get into how to deal with all the damaged minor bridges, broken pipelines, etc., etc..

Now, how is a just in time inventory system supposed to work inside a large region of the country where a disaster has happened?

So, how do my meanderings tie back? Because if society was as simple as made out to be by many, Katrina wouldn't have been such a big deal 200 miles inland. We'd have pulled ourselves up by our bootstraps and got on with it. Instead, for over week, things were screwed. It pointed out to me how dependent we are on the wider world.


Colonization is a different matter from the minimum population required to sustain civilization.

In a colonial situation you don't need that level right from the start. Rather, you can get by with a much smaller population provided you give them equipment that is long lasting or supply them with replacements from the outside.

If you have vehicles that last a century, you won't need an auto industry for several generations. So there's time to build up the population and develop industries as needed.


When conservative politicians talk about the "good old days" in the US, and being self sufficient, it's very often code for "Remember when everyone who mattered was white and male?" There was one republican strategist who was interviewed after retirement and admitted "We used to say 'ni&&er' this and that, and then later we switch it out with 'welfare' and 'lazy'" Mostly, "Big Government" translates into "hand-outs for lazy, evil minorities and single-mothers!". They're fine with big government, as long as the big parts involve jailing/bombing other folks.

Ok, end-rant


Some of the major determinants of this number that I see are the following:


The basic system of education in the western world is about a century old. It is centered, for a good part, on learning irrelevant stuff to be tested and promptly forgotten. This is not without precedent. China under the Qing dynasty had a system for educating and testing its civil servants that was centuries old mostly useless for the purpose of actually educating them. So is our schooling and testing these days.

I for one think that the choice of education systems can easily result in changes in the order magnitude we're talking about.

Modes of Employment and Development

When actually developing new technology, it may or may not be helpful in some instances to engaged in several different fields of time. There simply are some fields were the marginal gain of further involvement drops off so quickly that, unless there is a lot to do in this field, it is worthwhile to pursue several at once, which is often enough not at all compatible with current expectations of employers.

Design of Products

Here, I'm especially thinking about some things like replaceable parts and components. Replacing a standardized blackbox just doesn't require an expert in all instances and often very little instruction. Sure, you'd need more experts to design blackboxes that work exactly as expected, even for beginners, but in the end you could see significant reductions in some fields.

(I'm especially thinking of NASA here, that seems to employ rather a lot more people than strictly necessary. I have yet to see a general pattern in NASA space probes that would suggest that anyone economized on designing and thus testing/qualifying the whole damn thing. Hint: Most space missions cost upwards of 5 times the launch costs ... Launch costs are the space-mission equivalent of the price of uranium for nuclear power or the price of tea in China for that matter.)

Finally: non-optimal use of labor vs. capital

There has been very little progress in implementing available labor saving technologies in a lot of places, mostly through the unavoidable legacy of old institutions (which includes corporations).

Even though I still think that the sewing machine is a technological marvel of its own right, there has been very little progress in the actual process of textile manufacture for decades, for the simple reason that labor is cheap. In a smaller society trying to be self-sufficient (without being ruled by a hereditary dictatorship), there would be no niches of cheap labor and advanced technology would be more likely to be implemented.

Next, there is considerable potential in people over the age of 60 and under the age of 30 (and those include teenagers!) that our society has decided to have no use for. Either because it lacks ethical codes to properly use it: There is no reason for children not to work, if they are ethically employed. But there has never been a discussion to my knowledge of what such conditions are. The way we are currently herding kids in schools is definitively not ethical either in many instances. And finally there are a lot of bad retirement laws and prejudice on behalf of potential employers that keep over 60 year olds out of proper work or any work at all. (Germany has struck over 58 year olds who can't find a job from unemployment statistics for a reason - there just too many of them being unemployed and out of luck finding employment.)

All things considered, when it comes to the *minimum* of people required to run a society on current technological levels I'd put my upper bound at your lower bound and make a guess at around 50 mio or below.


"Get back to me when you've tracked down how many people it takes to design and build a space suit."

Has designing and building space suits always cost this much, or are the demands for a baseline colonization space suit higher?

Your observations on complexity are interesting, as it happens that sort of thing is part of the reason I became a libertarian. I think our society is too complex to manage, so leaving things up to individuals and the operation of the free market is the best course to take. Centralized planning leads to mistakes which lead to more planning and more wasted resources.

Ah, the wisdom of the free market.
Hint: please look into bubbles and their bursting in the marketplace. And once things occur on a scale large enough to impact large numbers of people, well, isn't one of the libertarian credos "Your right to throw a punch ends at my nose?"


To paraphrase Neal Stephenson (very last paragraph of "The System of the World"), it is better to have some kind of a system organizing society, even a flawed and doomed one, than to have no system at all.


Adam Smith had a pithy summary for this problem: "The division of labor is limited by the extent of the market." ("The extent of the market" meaning the number of people who trade together in a single economic unit, not any libertarian BS.)

On the other hand, I think for political and institutional reasons, we've pushed a lot of technologies in ways that make them harder and harder for users to manipulate. There's a huge literature in Science, Technology and Society studies (one of my homes) on "technological determinism" arguments and how they are often overblown. There have been a lot of turning points with various technologies about how closed to make them - about how high to set the barrier for maintaining and upgrading them - and we've often taken the high barrier route. I'm not saying that everyone is ever going to be able to code their own OS, but certainly "computer education" is about teaching students to use the programs that exist, which themselves are very closed. Etc. This issue pops up even the bio-medical side of things - we've "medicalized" a lot of problems that used to be dealt with by individuals on their own or much less highly specialized or trained professionals, often with no particular improvement in outcomes (see, for example, the midwifery movement).

So, fantastic post, and we definitely need a large civilization to maintain the tech we've got, but perhaps we could do a bit better if we re-oriented our education system a bit and re-worked some of our tech to be less closed, and fought against some of the hyper-specialization that hasn't produced positive outcomes as much as inflating the wages of certain folks.


Technological capability is an expression of the ability to co-operate in groups; stability (how long can you do this?) counts as well as total group size.

You need a certain minimum number of people to knap flint (1), a certain minimum number of people to run an 18thC canal lock (about 6), a certain minimum number of people to run a railroad (thousands, if you count mining the iron and the coal and building the rolling stock and laying the track and setting up the signals and making the clocks that let you run your scheduling...), and so on. So the group size is a limit on what you can do; no building a transcontinental railroad with a thousand-person workforce.

Modern technological economies assume that you're competing on your ability to use and capture the surplus that certainly exists. They depend utterly on the enormous (all of a petrochemical industry, all of a marine industry able to build 500,000 dwt super tankers, satellites for weather forecasting...) size of the co-operation necessary to build the thing in the first place.

This _might not_ be true of a colony; it might have a plan for the surplus for the first hundred years.

But in terms of our present economy? I'd say the population involved is the lower bound; at tech level, which includes the organizational systems (are you doing quill pen and ledger, quill pen and ledger really really fast using a database, or something that really uses computers?) we're doing about as much as we can. (What we can do increases in step with what we are presently doing, in other words.)

I'd argue that's an inherent function of economic activity in any setup of co-operating groups on a large scale; they do as much as they are actually capable of doing, because they are after all competing with each other in terms of what they can accomplish with a given amount of resources. (And every now and again, something like the NHS IT project, which might be feasible with a different set of organizational systems, but is clearly not at all possible with the organizational systems presently in place, happens, and you get a firm upper bound on what some particular organization is capable of doing.)

A level of technological capability much past ours could probably get the population size for our level of technological capability down by building in better organizational structures, but it's not clear that the capability for the tech to have the organizational structures would squish down to the capability of the lower population. I suspect an authorial call there could go either way.


Bubbles and corrections are the normal operation of things. After a couple centuries of them I'm not sure why everyone is always shocked and unprepared when they happen. And government interference often makes things worse, as seen most recently in the US.

Thinking that if only you get the right people in charge, who are smart enough and wise enough, and get them all the information they need, all our economic problems will be solved and everyone will get everything they want is a fairy tale for children.


A space suit is a surprisingly complex artefact. It's actually a very compact man-rated spacecraft, with life support requirements that go beyond just circulating air and water and include thermoregulation and ergonomics -- it's no damn use just keeping the astronaut alive if they can't physically work in the thing, for example.

I suspect right now the state of the art is a toss-up between the Russian Orlan M and the NASA EMU (although unlike the Orlan M -- which can be donned in five minutes -- the EMU comes in multiple pieces and takes 45 minutes to 4 hours of pre-breathing before it's ready for EVA). The EMU officially costs US $12M; you can buy a surplus Orlan-M suit for €149,000 so I'd guess that even though it's cheaper than the EMU, it's still in the million dollar range when new.


I see nobody's bothered to answer you. This isn't obvious to many people, but the difference in scale is enormous. Shipping supplies from Earth to Mars, for example, is infinitely more costly, difficult, dangerous and slow than shipping them over the Atlantic, even if you do the latter with primitive caravels.

Come to think of it, we have an excellent example at hand. Consider the cost of keeping alive three guys in an air bubble just over our heads (on the International Space Station). Now scale this to thousands of people living millions of times further away, on Mars. Yeah, I don't see it happening either.


The analogy I like to use is traffic lights. They're there so that people on one street don't starve to death waiting for people on another street to give them a chance to move.

Also, they prevent traffic accidents. Now you could claim that traffic lights don't prevent accidents, but make them worse, and things would be so much better without them, but there wouldn't be any reason to assume that you were right.

It's not that some genius somewhere is sitting a terminal and determining when every light turns on and off, but a decent system, tuned a bit ensures that things work out more equittably than they would otherwise, and things proceed more smoothly overall. Also, we have plenty of roads with no lights at all "high ways", this is because these roads work just fine with minimal regulation.


You just basically made one of the standard arguments for _small_ government, which is that the economy is so complex, with no single entity being able to see or comprehend the whole, that the best you can do is set up the proper incentives, (no "too big too fail", for instance), and let people work it out by themselves. Then you try to compensate for whatever flaws you discover along the way (externalities etc).

So when someone says "big government is bad", they do not (necessarily) mean "I wish everything was simple". They mean "_I_ know how to make that one out of million parts in that airplane, (or _I_ know how to turn the ideas in my head into books that people are willing to buy), and some bureaucrat in central government doesn't, so get out of my way". Whether that's a good argument is another debate, but the idea is not substantially different from what you wrote here.


How many consumers do you need to whose demands for good ensure that the various industries are kept busy enough that they can amortize the hideous cost of their infrastructure?

From what I've read, one of the reason we're having issues with rare Earth materials is that a production plant cost billions and takes years to come online. China wound up dropping costs on their sales and make it less economic to build extra supplies. And trial and small batch plants aren't useful since they don't have any economy of scale.

And Semiconductor Fabs are hitting $10 billion these days, and apparently will only be worse. Unless they are fully utilized, product prices would go up massively.

I'm not sure we could support mobile phone infrastructure if phones cost $10k, and people would have to gather up and decide which areas had enough people for them to build their own local tower. And then reception is very spotty, killing the usefulness of the phone itself.

So even if you had the capability to keep on producing certain items, I'm not sure if you could still justify producing them if the population is small. Even if you were to streamline economics and money purely into the efficient allocation of resources and fulfillment of needs.

So if we're on Mars, my Game boy may be required for us to keep those Atmospheric Processors going. Ooops.


I think a possible issue is that our tech-level is inherently complex and un-self-organising. If we can develop a new kind of technology that can either be overseen by Robotic Masters (cf The Culture) or with a more organicesque base (The Diamond Age, Edenists in The Reality Dysfunction) then human upkeep becomes less important. Our world's fragility is a sign of immaturity to me.


"Standardize, e.g., software and machine tools to eliminate extra redundancy. For a concrete example, I'm sure a colony doesn't need 45 different Ethernet chip designs."

Go too far in your standardization and you end up with fragile monocultures. Admittedly, 45 may be excessive. But where do you draw the line? The point of diversity is to be able to handle unforeseen problems, which are by definition, impossible to quantify...


The larger the number of people, the greater the number of people needed to supply the "infrastructure" of society. With larger numbers the need to harvest resources and manufacture goods increases as well.

Its true that we humans are very good at hyperspecialization, but a lot of that is redundant. Manufacturing vehicles is remarkably similar once you leave the design stage and get down to the assembly lines. The same skills are needed there no matter which kind of vehicle you are building. Designing the vehicles is another matter.

An interesting fact is that the company thats designing and building the Falcon rocket has around 1200 people employed. That includes all their design and manufacturing, HR and marketing etc.

Does a colony/viable tecnological society really need more than one factory that makes cars ? How about planes ? CPU:s ?


One interesting question is what degree of autonomy is desired; for a Mars colony, shipping information to-and-fro is reasonably cheap, so most of those designers can be on Earth. Rare elements in small quantities can be shipped.

Of course, then the colony is no longer autonomous.

For that matter, merely maintaining current tech also doesn't need the designers, provided that the designs themselves are available - only people that can follow the designs to carry them out. I have a feeling the Romans tried that approach, maintaining current tech level without extending it, and it worked for centuries before inevitably stagnating. Perhaps it would be acceptable for an almost-autonomous colony which will be getting news from the homeworld, or for one which has a mechanism to switch to progress at some point.


On grounds of brute empiricism, I'd pick a lower bound that's equal to your upper bound: 1 billion. That is, I'd say that we have a technological civilization, we have that many people working in occupations characteristic of technological civilization, we visibly have multiple industries that exist only in a few locations, ergo it takes that many people to run all those industries. Any supposition that it can be done with as few as half, let alone your 10%, cannot be founded on empirical data; it's pure theoretical design and likely fails to take into account some of the real costs and inefficiencies.

Upper bound? I'll guess the current global population. Just because large numbers of people are working in low-tech countries with low productivity and are desperately poor doesn't conclusively prove that they aren't a necessary part of a metaindustrial economy. In particular, the huge numbers of somewhat upwardly mobile poor people in China and India (and Brazil, I think?) may be key to sustaining the industrial growth that's enabling those countries to help sustain the global technosphere.

Now, if I put on my libertarian hat, I'm willing to speculate that you could classify all government expenditures as unproductive and the people they support as not useful parts of the technosphere. But even that would be only around 20% of the economy in the US and maybe 50% in some European countries; that doesn't translate directly to population, but even if it did, you'd still need 500 to 800 million. And that's assuming that all the government expenditures are wasted, which I don't think is a realistic claim.

So at least a billion, I think.


Bjorn at 33 --

The counter argument is that your cells don't know they're in a body, but you can walk anyway, as a matter of choice and volition.

Systems of organization are real things; there's nothing that says a system of government can't be built that is in fact able to do at least as good a job as your nervous system and metabolism do. (And hopefully without the billion+ years of evolution being required.)

Whether you _want_ one is a different question from "could this be built?"


I'm not sure "how many people?" is the optimal way to think about the question. I'm not sure how to phrase my counter-proposal... it's something along the lines of "cognitive mass" or "volume of cognitive complexity".

At least some of the lower-level specialized skills don't necessarily have to be encoded in living meat brains. Consider the drafting example. I took drafting classes in the early 1980s, with pencils and straight-edges and stuff, and I took CAD/CAM classes in the mid 1980s, with a bunch of Tektronix graphics terminals attached to a PDP/11. The better tools served as a "force multiplier", letting one draftsman get more done in less time.

Now, yes, those tools took their own specialists to build and maintain... but what's the relationship between the amount of work they required and the amount of work they enabled? And are changes coming that may reduce the number of human-with-meat-brain specialists required?

For some tasks (farm labor?), my brain spirals off directly into the plot of Capek's "R.U.R.".

Speaking of farming, while the economic incentive for it doesn't seem to be an active force in the world today, one could imagine GM crops specifically designed to be simpler to manage (instead of higher-yield or less fragile or more nutritious or more marketable or whatever).

You've now got me thinking about ... maybe "microsingularities" is an adequate term? Not a full-fledged singularity of the Kurzweillian "OMG everything changes and we can't predict past it" sort, but of the "one extended family with the right tools and automation and reference library can maintain a high standard of living on a self-sufficient island for many decades" sort. (Which when combined with space travel would result in an explosion of humanity across ... wherever their space travel could actually get them to.) Get to that point and I imagine we'd end up with a large number of insular enclaves ... I'm imagining a bunch of Kibbutz-like social constructs that interact *between* each other in a way thoroughly distinct from their internal interactions.

This also brings to mind the Eric Flint "Ring of Fire" stuff, where a batch of modern people with their technology are "poofed" to a "simpler time" (the year 1632). They don't scramble to maintain their current tech level -- they scramble to efficiently and deliberately "gear down" to a tech level they *can* maintain with the population, expertise, and resources they have.

So, the question arises in my mind, not "how many people would it take to maintain our current tech level?", but "how many people would it *now* take to maintain approximately the tech level of (say) 1975?". Intuitively, if you keep the "reference year" fixed, it seems like the number required *should* be shrinking over time.

Interesting topic.


If you allowed a single wave of a magic wand that would install an industrial base on the moon, then I could argue that while transportation from Earth to anywhere is extremely complicated and uses a lot of resources, the other way around is *much* cheaper.

Earth suffers from two really annoying things that make getting stuff away from it expensive and difficult: Gravity and an atmosphere. But the atmosphere makes it a lot easier to get stuff down to earth, as very little fuel is needed to decelerate.

The moon is the exact opposite. It has little gravity and no atmosphere. Having no atmosphere you have to use a lot of fuel at high thrust (and thus highly inefficient) to decelerate by about 2400 m/s before landing.

But you need no fuel to leave the moon, all you need is a maglev train. Using 1GW of electrical power you could launch 7500 tons of stuff to earth each day. (Assuming 60% efficiency - includes accelerating the train itself. "Stuff" includes necessary heat shields.)

So, while the Moon would be relatively isolated from Earth import-wise, it could make excellent use of economies of scale by exporting high-value stuff to Earth or any other place in the solar system (like computer chips or rare minerals that either don't form on earth or have long since been washed into the oceans).

As for the magic wand ... I have no idea where to get that one from.


I haven't researched this, but I suspect that 1) China's agrarian population is supplying food for the rest of the population but 2) the same amount of food could be produced by a rather smaller number of people.


Hi Charlie,

I think the question you asked isn't answerable, for three reasons:

1. Our society isn't sustainable as it stands. It requires technological infrastructure we don't yet have, and/or fewer people and/or more efficient resource use and recycling.

2. Society doesn't modularize very well. If it did, we could look at the cell size, and that would be the minimum number of people. What we've got, to oversimplify, is a kludge. I suspect you know enough about simplifying computer kludges to the minimum workable program realize what the problem is.

3. If there is a minimum number, what do we do with the so-called "surplus?" Do they make the system less functional, or do we put them to work on something? If they're put to work, does that mean that they are superfluous? Systems grow or shrink, and it can be hard to work out what's necessary or not when the size of the system is constantly changing.

I'd suggest it's more useful to look at what kind of technological civilization can be sustained with, say, 100 people. In your example, thousands of people may be needed to design a spacesuit, but if you have an advanced 3-D printing facility, how many people are required to make and maintain space-suits from good patterns? Such a society is going to be intellectually static and heavily dependent on a literally bomb-proof existing knowledge base, but it is a more answerable question.


On the "autonomous Mars colony" point, taking the example of real-world colonization, I don't see why the colony has to end up at the same tech-level as the civilization doing the colonizing. At some point, would one be able to build a colony kit that would enable a smaller group of people to build a community with, say, a century-out-of-date tech level?

Maintaining our current tech level may be optimal, sure. But some people are willing to sacrifice tech level for perceived simplicity or perceived freedom. The less sacrifice needed, I think the larger the set of people who'd be willing to jump at it. I mean, today we've *got* crazy hermits who go off to live in a shack in the mountains, hunting game and living in log cabins and stuff. What might the asteroid-belt equivalent look like?


The argument made here is well reasoned, and the premises are combined in a logical way, but I disagree, because I think the premises are too contrained.

* Many jobs from the 1950's are now mechanized or automated. The reason we aren't all having one huge holiday is because now we're spending our time on inventing *new* technologies and making those new technologies work. (Even so, it happens that companies in *.eu have gone from a 5-day workweek to a 4-day workweek, personally, I've cut it down to 3 days, but that's not for everyone. )
* Actually, the hoomon meatbags are only still around at all because they simply refuse to be redundantized. Obviously us technologists can't just walk into a department and go "Hi, I'm here to make y'all redundant". That's why you need a cutthroat manager to do that for you. (and you wonder why everyone hates them? :-P)
* Recent Investments in autonomous robotics, semi-self-replicating rapid prototypers, robotic assembly plants & etc. will eventually pay off in the sense that we could send a fully autonomous robotic colonisation mission to mars with _0_ colonists, and we could even transmit updated blueprints. At some point within the next generation, we could send a limited number of colonists, and they'd manage to survive and thrive easily, at least, from our current technological perspective.

Actually, this sort of thing has been discussed in science fiction for quite some time. Think of the Isaac Asimov's Elijah Baily stories, involving heavily robot-dependent Spacers (especially Solarians) versus the everything-done-by-human-hands view espoused by the Terrans.

Note that by standardizing semi-autonomous, semi-self-replicating production facilities, and using open source, open blueprints, open science and open standards, etc... you can reduce the number of people you actually *need* to keep a current civilisation running down to just a viable breeding population (a few 100) + a bunch of big fileservers, and a number of redundant production facilities.

Now, we'll probably reach that point in 1 generation, say ~25 years. Obviously, the other 6-7 billion people on the planet aren't going to sit around and twiddle their thumbs. they're going to go out and innovate some more ;-)

Generalized version:

At any point in time a generation (for some definition of 'generation') cannot make an autonomous colony that is equal to the living standards of that generation; because a society that has X research resources does not have sufficient resources to both be itself, and simultaneously introspect itself.

However, a society can certainly in(tro)spect a (smaller and) simpler society, such as -for instance- the previous generation.

Therefore: Every generation can set up a very small and efficient autonomous colony that has the living standards of the previous generation. It is much harder (or even impractical) to set up a colony that is equal to the *current* generation.


HA! It just now occurred to me that you might be brainstorming about a new direction for events in the "Merchant Princes" universe, where "colonization" wouldn't mean *quite* the same thing most of us have been thinking about.


I think these thoughts whenever green-types start pushing "sustainability". Yes, I want things to be sustainable, but mostly in the sense that I want HUMANITY to be sustained. I do not consider our survival as a species to be a settled question. We have been around for a tiny tiny fraction of the time that the dinosaurs roamed the earth, and they died out, so why should we think we are immune?

It may be a matter of numbers. We might just need to have enough people around to solve the problems that expose us to risk (spaceflight, cryonics, insert your preferred utopian solution).

Of course, the only way to find that number is to solve the problems, and then see how many you have.


I agree with Ben's rant. Conservatism in the U.S. is very race-centered. The controversy over whether Obama is secretly Muslim (which was promptly forgotten when conservatives wanted to talk about his United Church of Christ preacher) was merely a way for people to talk about him being from a scary foreign culture without mentioning race. The reason conservatives talk so much about states rights (while opposing the state rights of legal pot and gay marriage) is because of resentment over the Civil Rights act of 1964. It explains why conservatism is stronger in the South and why the Republican Party is championed by the Southern Baptist church in particular.

When conservatives talk about yesteryear, I think they really are talking about their perception of lost privilege, not their hatred of technology.


One of the issues of standardisation (i.e. one Ethernet card design) is the Windows problem - monocultures are frail.


You're forgetting about cultural and situational differences. You've grown up appreciating civilization, and like good access to curry, pubs, and new gadgets; and you haven't run into any crippling barriers like plenty of my wife's Irish ancestors and my Russian Jewish ancestors did.

People like you were also inclined to scoff - maybe rightly - at the barbaric and untamed wilds of America and stay home. But that didn't keep us the American colonies from growing like wildfire. Some populations whom moved, and whom still might be inclined to join a new Mars colony include:

o Borderer cultures like those that dominated Texas and California. Ours're now in the news as the Tea Party.
o The really curious.
o Refugees from oppression and starvation.
o People whom want freedom to oppress others, like our old slavers.

All those classes of people would be happy to move even to place with submodern conveniences.


We don't have proper spaceflight yet (defined by me, arbitrarily as, at a minimum, the ability to go back and forth to the Moon at will - rather than being able to do so only as the climax of a dedicated one off project).

So before we can plant colonies our technological level would have to advance to at least that point - so for example we wouldn't be talking about the number of people needed to design and make a 2010 aircraft but the (presumably much larger) number required to design and make a ?2050 spacecraft.

I think therefore the overall numbers need to be higher.


I'd go with Stoddard and guess one billion as my lower bound. But that's all it is, a guess based upon a rather rough and ready empiricism.

I'd also think that the type of social organization you have would count for something. I would guess - guess - that an authoritarian top-down rule implementing some sort of absolute survival philosophy might get by with fewer numbers than something like a democratic or at any rate representative form of government whose officials are elected by some sort of popular consent . . . if they knew what they were doing and followed expert rather than political advice.


It depends on whether you want this to be a eternally sustainable number or an initial number. Eternally sustainable would be much higher.

For example, my school has an experimental robotic mine. It can be controlled by only a few people, and is mostly automated. If you could make the robotic machines a standard design, each controlled by a standard computer with "Mining Robot Type A" software running on it, then they could be built and replaced by only a few people.

You run a single chip fabricator that builds one type of computer that can be used for everything. Different software is installed based on what it's being used for.
These days, a foundry needs what number of people to run it, assuming it's automated as much as possible?
Build a CAD/CAM machine with a lot of standard templates and the software to make new templates. Every part of every machine should be replaceable using a different machine to make it.

You can probably reduce the low level, mining, computer, machining infrastructure to 50 people per 1,000.

People could be trained for the simple jobs using an apprentice system on top of general education. The Master would teach the equivalent of college courses in the necessary areas. More general subjects can be taught by specialized teachers. Like Dynamics, for example, only very few engineering disciplines need to know. Most can get by with general physics and kinematics.

It would start small, but as the population grew up, people would think of things they wanted to do, that the can't be done with what's on hand. So they'd need to begin to innovate. Whoever sets this up would need to make sure that they have the tools to do so.

The hardest part would be software. They'll need a lot of very fault-tolerant software, and simple tools to write their own. Part of the problem can be solved by providing some very effective, general purpose simulation tools, probably linked to the CAD/CAM drawings.

Anywho, the general idea is to make it easy for only a few people to maintain, allowing all the surplus to go into expanding and improving it.


How about designing something sophisticated, yet self-sustaining, that can help support a small colony (see Hal Clement's stories for nanotech-based Pseudo-Life), like "Noise" and "The Nitrogen Fix."

It would not be a 21st century level of tech -- more like 19th centure combined with 23rd century...

Ed Sweet


The human body and human brain are still more complex than anything designed by humans, and will remain so for several more years.

How many people does it take to produce a new one?


Short version: the author is assuming that all present trends will continue into the future, and some will not.

Long version: As our technology improves, we will soon start moving towards self-sufficiency again. Within 50 years or less, a million dollars worth of computer will have the processing power of a billion human beings - and concerns about all the niche intellectual occupations we'll need will be moot. Desktop fabrication will make most other industries unnecessary. Humanity will merge with machines, and colonies will need more than one person only because people like to have company.


It only takes two people to produce a new one, but it takes dozens to have a viable population that can avoid inbreeding. Even if we get to molecular-scale fabricators and design libraries that let anyone who can find the right atoms print anything they want, there may be a minimum number of thinking beings required to maintain everything.


Fully autonomous will be tough. Being able to import as little as twenty pounds a person a year of electronics and pharmaceuticals would help a lot.


Your lower bound sounds about right. I can imagine Japan producing just about everything if it had to, given the necessary minerals, but a lot of things would cost more because of greatly reduced economies of scale.


Al @6:

Old British Empire colonies were not self-sufficient. Why would a colony in our solar system have to be?

Old British Empire colonies were firstly, created on this planet (there are no former British colonies on the moon that I know of) and secondly, generally in places where there was a pre-existing human population (even if said population was either systematically exterminated, legally written off as "livestock" or enslaved to supply labour needs in other parts of the colonial system). As far as I'm aware, we haven't discovered human life on any of the planets (or planetoids, or moons or large asteroids) in our solar system as yet.

Secondly, most of those colonies were at least capable of being self-sustaining. They were able to produce sufficient food, obtain enough clean water, and create enough shelter to supply all of the people who were involved in the colony (and by and large, they relied on forced labour to do it - either indentured servitude, slavery, or penal servitude). The skills required weren't major technical stretches, and the technology level required to set up a colony wasn't that great (not to mention that by and large, the colonists were piggybacking off the infrastructure of the indigenous inhabitants of the area being colonised). Setting up a colony on a different planet requires a much greater technological "leap" than setting up a colony on a different continent: look over the history of British colonisation, and you'll realise a lot of the harder work was actually done by the explorers, who firstly located the "unoccupied" areas, secondly discovered sources of food and fresh water, and thirdly mapped out the terrain - the colonisers basically marked a particular spot with an "X" and headed there. While there's been a lot of work done by unmanned space probes when it comes to exploration, one of the key features of interplanetary colonisation is going to have to be either building shelter before anyone actually *gets* to the spot marked "X", or at least partial terraforming.

Both of those will require more advanced technology, and more complex skill sets than are likely to be acquired from scraping out the prisons and sending off the dregs of the army to supervise.

As for the whole initial issue of how many people it might take to sustain a colonial infrastructure in an inter-planetary situation, the trick might be to stop and consider how complex a society you want to start off with. In the cases of most colonial situations on this planet alone, the historical evidence is that until colonies became at least self-sustaining, they suffered an overall "loss" of technological advancement compared to the "home" culture. However, once the population reaches the point of sustainability, they quickly catch up with their former "home", and often surpass them in terms of quality of life (mostly because the new colony is able to install the newest, latest, super-duperest infrastructure created with the latest technology, while the home culture is generally using infrastructure at least one or two generations older). Until they become self-sustaining, the new colony will need constant inputs from the home culture, so it won't just be a matter of sending out just one generation ship - it'll be a matter of sending out at least one ship every single generation until the colony reaches a point where they can send people back.

I suspect the overall number will be much higher than the initial estimate of 100 million.


North Korea seems a pretty relevant datapoint. Population 23 million, imports only $3 billion on a GDP of $40 billion.

Some years, manages to maintain a ~1930s standard of living.


How many university specialties are there? 10,000? It might well work with a million, if the specialties could be allocated correctly.


One issue I don't notice addressed in the post or comments (my apologies if I missed it) is the matter of population density. I think that the answer to the question of "minimum number of people" has to include some indication of how densely packed they are, and possibly also where. Density has its advantages: you don't need to manufacture nearly as many cars or airplanes, for example, and fewer transportation links for food and goods result in fewer people needed for coordination and maintenance. Possibly also fewer decisions, which in turn leads to more efficient mechanization. It also makes a community better able to deal with its waste, though more vulnerable to natural disasters.

My way of thinking about the problem goes as follows: take a country like Japan that reasonably exemplifies the level of technological civilization you'd like to preserve. Then take away the rest of the world, leaving the population of Japan (~130 million) and all the arable land and mine-able resources behind. Can it maintain itself at these levels? I think no: According to one site I found ( ) Japan imports food, oil, and wood. I'm betting they also need raw materials like iron, copper, etc.

Being Japan, they get a tremendous boost in this thought experiment from their taste for fish: without the rest of the planet's population, it will be easier for even their current fishing fleet to bring in much larger catches from further afield, catching from more sustainable stocks. By the same token, the area formerly known as Russia would provide plenty of wood (particularly if they are allowed to use the railway and its terminal in what will have used to be North Korea) This means staffing the areas of production and supervising transportation, then in turn feeding and entertaining those people.

Mining and oil then become the main needs, the raw material extraction you discuss above. Say, another 1 million for that, especially if these efforts can overlap with the foreign production of food and timber.

That, I think, results in a self-sufficient population, at a little over 130 million people. So then the next stage of my thought experiment is to ask what's missing before I arrive at the full level of technological civilization anywhere on the planet. Space travel, I think: the Americans and Russians have a bit more capability there than the Japanese space agency does. Software, possible: the Japanese software industry is capable but small. I don't know much about Japanese pharmaceuticals, that might need shoring up. I can't think of anything else off-hand. The resulting level of technological civilization will look very different than elsewhere in the world, but I think will be equivalent at about 135 million people.

Now, the trick is that I think that if you start off with a different group of people -- say, the UK, or Texas -- you might wind up at a wildly different number. If the population were made up of Americans, well, we like to spread out, we eat lots of fruit and vegetables out of season, and we vastly prefer meat to fish. We'd need much more transportation and food production infrastructure as a result, and the attendant increased population. I doubt we could manage to sustain a population of fewer than 400 million Americans.

Every generation can set up a very small and efficient autonomous colony that has the living standards of the previous generation.

So what you're saying, if I read you right, is that we should get all of the politicians who hark on about Victorian values and the Good Old Days, and we should send them to Mars?


Interest essay, Charlie.

Sounds to me a lot like an analysis of reproduction:


Nice one, Charlie! What seems at first to be a straightforward technological question that turns out to have some really interesting sociological, political, economic, and even anthropological issues associated with it. This is the sort of question that you could turn loose on a 2 or 3 day interdisciplinary workshop, something hosted by NASA at Esalen Institute, or by ESA at Noordwijk.

The point about needing diversity (and therefore larger numbers) of technologies also holds for people and subcultures, so you will need more people than the bare minimum number required to do all the jobs. If nothing else, you need a pool of uncommitted labor to handle losses due to accidents (up to large scale disasters), illnesses (up to large scale epidemics), and other unpredictable events. And no matter what form of government you start out with, it might change over time, changing the parameters of the numbers problem.

That last is a fundamental part of the operation of human societies: they change over time. This is a good thing from the point of view of human survival, since an organism or organization that can't change is much more fragile in the face of changes in its environment, so we need to allow for change. That would mean having a pool of people who are willing and able to be retrained for newly-created skills.

I'd guess that these considerations mean you should increase your lower bounds by at least 25 or 30%.

Oh, by the way, would the people advocating reduction of "unnecessary" jobs consider artists and writers unnecessary? How about athletes? Philosophers or poets? I'd claim we really don't know very much about what is actually necessary for a human society, so if we're trying to build one we'd better be generous rather than miserly in obtaining skills and talents.


Everyone is looking at this backwards. Thing is, this is a solution to the fermi paradox that does not require the universe to be full of suck, because it gives you a galaxy where people can roam, explore, dispatch embassies, potentially even trade (at a minimum information!), but *not* go forth and turn every spec of rock in a remotely suitable location into a clone of Zferti 4.


Many SciFi authors 'get around' this way of thinking by staging any human diasporas after some sort of near perfect technology milestone. That is, the ability to build self reparing, or duplicating machines.

The portfolio of designs available becomes important to colonists or starfarers, rather than their ability to engineer new solutions.

Many authors describe 'general factories' that can produce any product a design is avilable for.


Stepping back from the question of how many, there is another question to answer-- is maintaining the level possible? Or is it the case that a society that is not growing and changing is a society that is stagnant and dying? I honestly don't know, it is an open question to me whether or not stable states in society are maintainable.


Watch Matt Ridley's (Mr Rational Optimist) talk at TED since he mentions what happens when there are sufficient poeple to support a particular degree of specialization. I think your upper bound is actually too low...


I wonder whether an allowance could be made to 'cheat' a little. If we're talking about colonies, one in the solar system would be able to exchange information freely with Earth. All research and design could be safely offloaded. Designing something (and therefore truly understanding it) is immeasurably more difficult than just building or maintaining it. Judging by my house (and from work in the water industry) almost everything is made in (but not normally developed in) the industrial bits of China (quick wikipedia search ~250 million industrial workers). The hardest thing to actually manufacture seems to be electronics. Given these have a very low mass for their value, perhaps these could literally be shipped...


So what you're saying, if I read you right, is that we should get all of the politicians who hark on about Victorian values and the Good Old Days, and we should send them to Mars?

Yes, though fortunately (or unfortunately ;-) ), in this case "previous generation" would be something like 1970's or 1980's levels of technology.

I've been reading the Apollo Spacecraft manuals recently. (1970's technology developed in the '60's). It's quite cool, The CM and LM each had just 1 very underpowered computer on board, and yet they multitasked: fly-by-wire, instrumentation, navigation, radar, autopilot and all kinds of software all at once on the one thing. Some of this stuff still isn't available on modern aircraft. (And I'll definitely bet they can't do automated celestial navigation or orbital calculations :-p )

But the Apollo systems are very simple by the standards of today. If we were to do over the program today, it would cost a lot less and require a lot less people. (And if you look at modern efforts at commercial spaceflight, that's exactly what we're seeing. )

TL;DR: If you want to have exactly "todays society", you'll need a copy of everyone in it, and the entire economy besides. If you just want a viable society, pick something from just 1 or 2 generations back, and you'll probably manage. (and this should be true for any value of "todays society")


Hi Charles,

Really like your question.

1. You don't need many people to make a civil society that can contribute to the greater good. There are only 4 million kiwis. In addition to having the potential to feed the UK cheaply (our original killer app. May all in the EU be reviewed by Auditors. From the Laundry). We make medical dehumidifiers, a large amount of high tech agricultural products and parts for chip fabbers.

You do need a series of these nations to make things. I guess that Australia and NZ could make most of the things we need if we really had to... and that is 20 million people. So your lower end is too high.

2. What is needed is standards. Moral == high value on trust and cooperation, basic honesty. Technical == railway gauges, Cat 6 cable structure, power points. Educational == an accepted range of knowledge everyone should know. The standards have fairly minimal tolerences -- and variations are not acceptable. This allows for the ability of people to move and "plug in" to new industries.

(This may annoy Feorag, but much of this was held by the Church for centuries).

3. I do not think we need just in time etc to make our society. NZ was at the end of the British Imperial logistic chain. We had things called warehouses. Run by canny Calvinists. You are going to need some inefficiencies -- to allow products to evolve (ie, the market) and to allow for defence (as a nation, and from personal disaster ie. some form of emergency social system).

4. Depending on where you set the society up, you will need colonies. To run the proverbial mines... both for minerals and biological substrates (although I think we underestimate what you can do with long-chain carbohydrates).

With the right infrastructure and good education, my guess is you could do this around 20 -- 50 million people from a cohesive nation or hegemony.



jobs like 'athlete and poet' arent really jobs, theyre hobbies.this innovation thing seems to be the problem, If you have a design that works- use that.
If the guys building it or who use the item have a idea of how to improve it, try it.
Kaizen. ,
that should chop your requirements right down.
and we can delete the telephone sanitizers and hairdressers, they can go on the B ark


I think considerable simplification is possible. For example, a conoly could use a single human language, say English. A lot of resource is spent in learning and translating foreign languages.


I think something that would make a huge difference is if this is designed and planned, or if it simply happens.
If you have to spend effort protecting against terrorist attacks from the disaffected outsiders then it's going to take FAR more effort and people than a peaceful society.

Second is the environment. Metal rich/poor? Hydrocarbons? Moon of a gas giant?

For instance, if you can arrive at the moon of a gas giant with a nice powerful magnetosphere, then you trail a superconducting wire out in orbit, hook it to the top of your space elevator with a strand of superconducting nanotube to the ground, and you've just eliminated almost your entire power infrastructure. You've got free, near infinite electricity. It'll slow you down, but there's a LOT of orbital energy in a earth sized moon.


The converse question is, "does this level of complexity increase so much that the earth's population maxes out it's rate to develop faster and more complexly"? A slowdown for the singularity?

I suspect that others on this comment thread are correct that if we can make things with fabs (what Nigel Calder called "santa claus machines") we could live quite well in a colony remote from the earth and without it as a source of goods and services.

However, what if we don't colonize worlds with a high tech civilization? Earth like worlds, if they exist, could be colonized with quite simple civilizational needs.
Even more hostile worlds could have smaller spaces made viable for humans, for example underground colonies on the moon.


Logistics question for you, Charlie, just because I'm curious: When you post a blog entry like this, are the quoted facts and figures made up entirely of things you have researched or read in the past, or do you spend some time verifying up numbers before hitting the "publish" button? Obviously some of the figures are educated guesses, and you usually imply that if you don't state it explicitly. I'm just wondering how much your blog is like "work."


There is simply no way a colony is going to maintain the technology of the mother society. I think what you'd do (ethical issues aside, and assuming a planet where earthlike crops can grow) is provide the colony with tech from the 1800s and a really, really good library. That way you can colonize with 5-10,000 people.

Hopefully, the original colonizers would create some kind of a blueprint for technological growth that includes best practices and avoids technological dead ends - for instance, you go straight to solar panels and skip petroleum completely.

As someone said above, this paradigm collapses if personal fabrication and similar technologies prove practical, but we'll have to see where that goes. (These technologies look really good, but I'm not sure they can be considered proven at this point.)


Don't forget the telephone sanitizers.


Taking that converse question further - if we're positing that the rate of progress depends on population - for all the reasons mentioned above - what other questions can we ask:

What is the shape of that curve wrt population - heck, can we even tell if it's superlinear (metcalfe's law style arguments) or sublinear (diminishing returns due to scarcity of resources, or opportunities for progress, or something) at the sizes we're thinking of?

If we also add an assumption that increasing populations unfortunately also add increasing environmental loads per unit time until the invention of some hypothetical futuretech, what's the right strategy for bootstrapping to that level while minimizing that load?


Really interesting question. A couple of points:

First has been raised, but not as clearly as I feel it should:
The total number of people required to design a wheel (today) is precisely 0. The designer is already dead and we can still build wheels: therefore, to maintain our society you do not need to factor in any designers: only manufacturers, maintainers and support staff (assuming a society that exists for the maintainence of its technological level and nothing else).

Second, its been touched on with suggestions of lowering the colonists standard of lifet: increase the tolerance to failure. The vast majority of specialists are focussed on producing outcomes with very small tolerances and/or are applicable in niche cases. Maintaining a Formula 1 race-car (or any very finely tuned performance system) takes very specific skills, but for general use that Chevy '69 (or any simply functional system) is good enough. Equally, having an oncologist will allow you to treat more cancers (therefore increasing average life-expectancy), having only a GP means that you can't treat as many cancers (decreasing life expectancy): a tragedy for the individual, an acceptable loss for the colony as a whole.

Ultimately, the systems shipped to Mars with the first colonists will need to be designed to function as a robust, flexible whole: the mess that is our society with so many things that are jury-rigged to work together involves far too much waste.

Assuming that the colonists are (initially) supported from earth, the colonisation will not be one ship, or even one ship every generation: a just-in-time supply chain will require the first load of re-supplies to leave Earth whilst the colonists are still in transit (its contents established based on projected needs). I'd imagine a constant stream of transports leaving Earth for Mars. This in turn requires a support network that is efficient and sustainable over the (potentially) centuries until self-sufficiency is acheived.

However, I found this question more interesting from a post-apocolyptic point of view. If one believes in a coming global catastrophe that will threaten our current society a follow-up question immediately leaps out: how do we ensure that the minimum number of people (with the right skillsets) survive?

If that is impossible: what is the minimum number of people (with what skills) to re-establish our current technological level post-catastrophe? How do we ensure their survival?

These questions are more practical and grokable, dealing as they are with a world in which we know humans can survive with current technology.


A second question: how do you factor in non-productive members of society? The minimum implies no wasted man-hours: how achievable is that? If not, what is the average amount of waste can we allocate per person?

There is also the consideration that as group size increases the amount of resources allocated to administering the group also increases (I can't remember what the name of this function is, or if it's an absolute increase or a proportional increase).


Thinks for your thought provoking article!

I quite liked reading it, though I think you overlook two things:

1) With standardized parts, the amount of design work required to keep the technological level woud drop significantly.

2) With basic technology we wouldn’t necessarily be 19th century, if we keep certain key components.

To 1) Why do we need several different processors for cars and mobile phones? Mostly efficiency and patents. A dedicated processor takes less power, and you can’t just copy stuff from others.

Now imagine a society built on free software (which anyone can change and adapt) and standardized parts (being reused everywhere). Today many devices even ship with their own screws, but these aren’t strictly needed.

I’d guess that you’d need 90-95% less people for maintaining the devices, because most things would be done the same way. No 5 cars with most parts different, but rather one type of car, one type of train and one type of train, all using the same parts wherever possible. And the machines creating the parts could mostly be built from these same parts, too.

You might waste 10-20% additional energy, maybe even 50%, but you would need wastly fewer people for keeping your society at its level.

The second point: If we would stick to certain key technologies (for example computers and everything where our advanced knowledge makes it easy for us to create it with little material requirements), we wouldn’t go back to 19th century. It just would be necessary to identify *simple to use* key technologies and make sure these are kept alive.

But here’s another problem to think about: If a religious ideology contra science would gain traction here on earth, how long would we have enough people to sustain our devices? At what point would we get into the rotting phase where parts aren’t recreated but only replaced with parts salvaged from already broken stuff?


The human body and human brain are still more complex than anything designed by humans, and will remain so for several more years.

How many people does it take to produce a new one?


You're wrong. (Classic reductionist thinking.)

While it takes a direct genetic input from two individuals to create a new one, to maintain a viable population takes a genetic pool of, at a minimum, several hundred. Some individuals within that pool will be fertile; take two individuals at random and you have a less than 50% probability that they will be able to reproduce. There is also -- absent medical backup -- around a 10% chance that parturition will malfunction, killing one of the parent units and seriously jeopardizing the survival prospects of the new one.

Moreover, the unit you produce is not independent. It takes an absolute irreducible minimum of one adult -- and not just any adult, but a lactating female -- to keep it alive for the first 12-15 months; then decreasing dependence for at least five or six more years. That's under ideal ambient environmental conditions, with zero predation, zero parasites, and food falling off the trees around it, a "garden of Eden" environment. In practice, such environments don't exist without massive human curatorial inputs. And at the end of that seven or eight year period you have, at best, produced an illiterate juvenile hunter-gatherer with minimal language and no social skills: little more than an animal.

To manufacture a useful member of a technoculture takes a lot of work, over a period of 15-30 years, and a buttload of support infrastructure; kindergartens, schools, hospitals, clinics, vaccine factories, clothing manufacturers, toy makers (or did you want your children to be developmentally challenged due to a sparse environment lacking in stimuli?), and so on.

Finally: your assumption that thinking machines will want to think about things useful to us is a charming conceit, but questionable. I suspect that left to their own devices, they won't. If we want them to be useful, we'll have to direct them; to at a minimum establish the parameters of the problems we want solving. And when the problem is as big as "design a plug-in portable replacement for human civilization to support a small group of expolorers or colonists a long way from home", that's going to take quite a lot of parameterization.


"If one believes in a coming global catastrophe that will threaten our current society a follow-up question immediately leaps out: how do we ensure that the minimum number of people (with the right skillsets) survive?

If that is impossible: what is the minimum number of people (with what skills) to re-establish our current technological level ..."

Or putting it the other way round, if a catastrophe happens without preparations and the only survivors are a random ensemble based on resistance to the virus or chance location when It happens - what level of technology are they likely to be able to support? How far will we fall?


However, what if we don't colonize worlds with a high tech civilization? Earth like worlds, if they exist, could be colonized with quite simple civilizational needs.

I find that unlikely.

Earth-like worlds with an oxygen atmosphere will only have such an atmosphere because they've already developed a biosphere that includes photoautotrophs. Which in turn implies the almost-certain existence of animals that eat them. Which in turn suggests the existence of the same sort of immunological and predatory arms races we see around us in Earth's biosphere(s).

In other words? Step out of the airlock, take a deep breath, and you'll probably die of anaphylactic shock within half an hour. And if you don't, sooner or later some of the local microfauna will discover that you make a lovely buffet ... and your immune system doesn't recognize them.

(There was a recent feature in New Scientist popularizing some research suggesting that chirality in amino-acids -- and by extension, in biochemistry -- might be seeded by differential absorption of UV light in the residual nebulae left by supernovae; in other words, that there's a physico-chemical reason for our protein chemistry's chirality which may well apply to biospheres on other planets as well.)


"Jobs like athlete and poet" may be unavoidable, frankly.

If we're considering the viability of a technoculture as a whole, we have to include the obscure fact that human beings require motivation. Yes, starving/beating/threatening to shoot them will work in the short term, but you can't keep it up forever without hugely damaging consequences -- consider, for example, that the long-term result of Stalin's terror and the Five Year Plans was to train Soviet middle management to lie like a rug when questioned about their productivity. And that's at the top level. In practice, over time humans of working age tend to put in about 7500-10,000 hours per year of labour. And if you crank up the hours worked the quality often declines, especially in knowledge/intellectual work.

And humans tend to want entertainment options when they're not working to distract them from inadequate or unpleasant workplace experiences. (The only ones who don't seek out entertainment tend to be the ones who consider their job to be the entertainment. Lucky sods.)

Even if we go for a totally austere, monastic, puritanical, planned society, there'll be some built in niches for "entertainers". For example, designers of educational toys and materials for training children -- children, especially infants, learn by play, and as the society changes the required models for them to play with will change, so toys and games and books will need updating and redesigning.

So no, I don't think we can do without musicians, writers, game designers, toy makers, and probably professional athletes too. Although they may be left out of the planning apparat for a colony, other than by pencilling in some hours a week as "free time" and assuming that some people will fill it with something other than drugs, crime, and rioting.


I think your back of the envelope calculation does make a lot of sense, but I feel that there are two big feedback terms you have left out.

In the current western civilization a quite large proportion of the workforce is engaged in counter-production.

One example of counter-production is marketing, being deliberately dishonest about things, in order to prevent the application of the best tool for the job.

Another example is patents and copyrights, actively preventing development of the best tool for the job.

But the worst form of counter-production is bureaucracy, whether you attribute it to Dilberts or Parkinssons principle.

Neither principle has any traction however, if there are nobody out there you can hire.

In fact, I assume Parkinssons law can be used to prove that the civilizations paperwork expands to consume the population, nomatter what size that might me. (See also: "Bill the Intergalactic Hero")

Similarly, the entire TV/Movie/Music entertainment industry, is the mass-production version of Caesars "Bread & Circus" crowd pacification strategy, necessary mainly because the last thing USA needs is for its entire population to take an active interest in the actual government of the place.

A population plunked down on Mars would be a whole lot more focused on survival and less on reality-TV, and consequently, I would not be surprised if an efficiency improvement by a factor of 10 can be realized on that account alone.

The second feedback term, is the question of which N individuals you send to Mars: Do you send the best for the job or do you send the ones which look good on TV ?

That alone is good for another factor 10 adjustment of your estimate.

Subject to the need for manual labor in terraforming, I would therefore argue that you could comfortably colonize Mars with only about 1 million humans, and still live keep your cool style, provided you neither bring private corporations need for profit nor respect intellectual property laws along.

However, if you pick the colonialists by lottery, or bring along laws designed to pacify a too large population, I think your 100M number is very realistic.



However, if you pick the colonialists by lottery, or bring along laws designed to pacify a too large population, I think your 100M number is very realistic.

This isn't really a space colonization question. (Other than insofar as our current planet is the Spaceship Earth, as Buckminster Fuller observed.) We don't get to pick individual "colonists".

A better frame than space colonization might be: it is the year 2200. The demographic transition has gone too far, and families of 1 or 2 children are "normal", more than 2 being seen as weird -- consequently, Earth's population has dropped back below 4 billion and is still trending downwards. How low can it (safely -- for human civilization) go?

Alternatively: the Martian Hyper-Scabies epidemic has wiped out 50% of the human population, planet-wide, before burning itself out. After the mass graves are filled, how sustainable is our level of civilization? (Ignoring environmental externalities for the time being, and collateral damage due to things like supply chain collapse/power interruption during the epidemic.)


Put like that, the trouble is, you can't necessarily rely on historical precedent, can you? I mean you might think we could say "The Victorians could support their society with X population so we could afford to go down to X and have at least Victorian complexity: and as as we've discovered lots since the 19th century we ought to be able to manage with less than X."

However, the Victorians were building up - for example, when they built their railways, they had lots of horses to help with the heavy work. We'd be on the way down. If a disaster wiped out our ability to refine and distribute oil, we'd not only have no petrol or diesel based transport, we'd have no (or few) horses to fall back on either. Or the skills to use them, either.

So there's probably a sort of hysteris going on here - for a given level of development we might need more people, to cope with the lack of skills and infrastructure that previous societies inherited form their predecessors.


Erm, Charlie:

> In practice, over time humans of working age tend to put in about 7500-10,000 hours per year of labour. And if you crank up the hours worked the quality often declines, especially in knowledge/intellectual work.

May I remind you that a year has at most 8784 hours?


Jobs like athlete and poet may indeed be unavoidable; so, too, the humble telephone sanitiser. Nevertheless, the passengers of Golgafrinchan Ark 'B' can be excluded from the colony - how many bankers, advertising copywriters and accountants does the colony (or any society) actually need?

The early-stage colony will need pioneers to build the habitats, the mines, the power plants and basic factories; they will live a life of hard work and austerity, and possibly a very short life at that. Its what pioneers do, for all kinds of reasons.

Successive waves of migrations bring more sophisticated skills, and build a more complex and culturally-enriching society - you might end up with fifty or a hundred million, and a well-developed service sector, but you don't have to send a fully-formed G7 economy out there on day 1.

As for the numbers, I'd look to Britain in 1942: an industrial economy with a productive population of 30 million, with a hugely successful research effort, self-sufficient in food, pharmaceuticals and basic goods. If they hadn't been at war, they'd have been self-sufficient in hydrocarbons - they 'made do' for everything else on low-quality local mineral deposits or did without by ingenuity - and the massive industrial output for the war effort might be seen, in this analysis, as a proxy for a 'terraforming' programme.

Note, also, that wartime Britain had fully mobilised the population - there were no domestic servants, shop assistants and hairdressers (all were put in uniform or into factories) and the 'Middle Third' of insurance clerks and car salesmen were all shipped overseas to fight. But they still had schools and universities; music, books, and films. They did basic scientific and medical research and a huge amount of technology development and innovation - directed towards warfare, of course, but a colony's R&D would certainly be focused on the problems of offworld colonisation.

Is it a good model? Probably not: I wouldn't want to live in wartime Britain. Even without the risk of death, it was austere and - let's not forget - it was a centrally-planned economy. We know what they turn into. But it worked, and it's a good study for a mid-stage colony that can, within a generation, build a culturally-rich society.


Whoops, out by 5 error. (I substituted hours/week for hours/working day.)

Mental arithmetic before caffeine: just say no.


As for the numbers, I'd look to Britain in 1942

That's a lousy model. Britain in 1942 was indeed fully mobilized. It was also running in crisis mode: between late 1939 and autumn 1945 there were no repairs to infrastructure other than emergency patches to keep essential services from breaking down completely. Universities were running two year degree courses instead of three year courses, and the graduating classes were going straight into uniform as officers (my father was one of them); subjects of dubious utility to the war effort weren't being taught. Music books and films were seen as essential to morale and as vehicles for propaganda; paper was tightly rationed. And after nearly six years of this, the British government came within a week of declaring bankruptcy in late 1945. They basically burned through an empire's worth of paper in six years (admittedly the rival empire they were fighting came off worse).

Anyway, I suspect you underestimate the need for "bankers, advertising copywriters and accountants". You need bankers because they're part of the risk management infrastructure of society -- when they're doing their job right. "Advertising copywriters" can write other stuff as well; I notice you didn't advocate doing without technical authors? (Happy fun statistic; in the 1980s, the US military concluded that roughly 30% of peacetime fatalities were due to documentation errors. Accurate and useful user manuals for machinery matter a lot.) As for accountants, see bankers, above: they're part of the feedback mechanism that tells you whether your business processes are working efficiently. If you give them an insanely complex tax code to abuse, then of course they're going to spend their time looking for loopholes and generating billable hours. But if you view them instead as resource optimization specialists -- "how do I maximize the bang per buck invested in these projects?" -- it's an entirely different picture.

Put it another way: try to imagine a complex society without risk management, efficiency monitoring, and communication skills. It's a mess, isn't it?


I think John Murphy @63 is on the right track.

Take one of the fully developed exporting (which implies excess capacity) Industrial nations, the smallest we can find, and tack on a Primary and Secondary resource producer with a fully developed infrastructure. Germany (83 Million), Japan(127 Million) or Taiwan (24 Million) with Australia (22 Million) or Canada (34 Million).

This gives us a number in the 100 million ballpark.

All of these countries have the necessary educational, social, and governmental structures to maintain themselves, and grow.

I certainly don't think that a population like this could be plopped down on another planet and function. The infrastructure and resource development has taken over a hundred years for even the youngest of these countries.


@CS, about caffeine..
So, what do you think about Greg Egan's quote
"I'd tried caffeine a few times; it made me believe I was focused and energetic, but it turned my judgment to shit. Widespread use of caffeine explained a lot about the twentieth century."
..being a self-proclaimed addict who cannot function without a daily fix? (It's from a character in a book, but something makes me think he was rather preachy in Distress and it might be something he believes or believed)

Caffeine boosts mental performance especially among the grey-haired folk, but does it improve judgement or decision making? If it's just benzedrine* lite..

I couldn't find any studies about that, though :(

*Ayn Rand was a heavy user of amphetamines. In fact, her personality traits have an uncanny resemblance to the list of psychological side effects of amphetamine use..


yn Rand was a Speed Freak? That... explains so much.


"Step out of the airlock, take a deep breath, and you'll probably die of anaphylactic shock within half an hour. And if you don't, sooner or later some of the local microfauna will discover that you make a lovely buffet ... and your immune system doesn't recognize them."

No. Firstly, even on earth you are only susceptible to a tiny fraction of pathogens. If that wasn't true, explorers would die the moment they left their local environs, e.g. Polynesians colonizing Pacific islands. The reason, as you know, is that organisms must adapt to each other. In another ecosystem, organisms probably would not even have the mechanisms to invade earth biology.

Secondly, your immune system is constantly working. You would breathe in those spores and immediately start generating antigens. Anaphylactic shock is the result of either an overloaded immunune system (see first point) or a dysfunctional one.

It is more likely that the problem colonists are going to have is that they cannot react biologically to the local biota at all and will be the new "shadow biosystem" as Paul Davies might say.


At this point is appears that the growth rate is non-linear, but how long can that go on? The logic of specialization that Charlie is alluding to runs out quite quickly under this scenario.

I speculate that this contributed to the fall of ancient civilizations. They reach some local high level of achievement, this is disrupted by some "barbarian" invasion and they cannot pull the web of specializations back together and so fail very quickly as their system unwinds to a lower level of complexity.


But what is the need for accountants and bankers *if* they have a reasonable tax code and well-regulated financial system to deal with?

The US employs about 5% of its labor force in finance (including banking and insurance), whereas the EU employs half as much and I can't bring myself to see much of a difference in their performance. The fact that European financial sector shed jobs during the crisis just as the American did supports that notion (at least a bit).

There is also the matter of overcapacity, even in the best of economic times, resulting from competition between companies. (Having spare capacity is a proven way to scare off competitors.)

Then there are lots of jobs resulting from nothing but the desire to create jobs, but not the desire to get anything done. NASA is an unfortunate example. (The number of jobs are one of the major points deciding projects and can easily override mere rational thinking.)

However, I'm slowly moving my lower bound (tens of millions) upwards, as I see the need for redundancy. Currently, we have too much redundancy, but that doesn't mean we can get rid of all of it.

Finally, there is the mode of extinction. If people died quickly in a plaque, they are likely to start working harder, having lots of children and fill gaps within decades.

Whereas any slow extinction I can think off would have a very different set of motivations. A slowly declining population - even over centuries - may not be able to overcome social rigidities. The Byzantine Empire comes to my mind.

However, a *world* of declining population would be unimaginable to me without at least some small countries growing in population, who would be likely to reform themselves and eventually overtake the declining rest. People just don't stop breeding if they see a perspective for their offspring. And some or even a lot of them do, they will always be outnumbered by the others at some point.


It seems like if you could stagger waves of colonists, with the first group of maybe 1000 building up agriculture and then basic industry, the second group of 10,000 building upon that, etc. What this brings to mind is the colonization from some of richard k morgan's books- send colony ships with racks of digitized personalities and load them onto fast-grown clones. If anything, this might be closer than the technology to ship a million people to mars.


One somewhat depressing effect of Charlie's logic is that colonization just gets harder as technology improves. The technology needs and hence population size size to sustain it outraces the improvements in star ship capability and capacity. At some point, you need to shift the whole planet from a to b, including infrastructure.

Even were we to posit replicators to maintain technology, would a smallish colony b relatively static, unable to do more than minor improvements and dependent on the replication technology to be available forever? I that sense, replicator technology is more like biology and our colonists while having a technological base, are more like agrarian societies until they can increase their population to the baseline "technological take off" level.


Would Charlie's thought experiment apply to a machine civilization too? In other words, are we fairly safe from expecting to find machine cultures colonizing the all available worlds in the galaxy?


I disagree. Strongly.

We've actually got a couple of different immune systems; the innate system (non-heuristic general-purpose first line of defense), and the adaptive immune system.

The innate system is basically pre-wired to respond violently to an invariant set of triggers indicative of attack by terrestrial pathogenic organisms. If we hypothesize a biosphere with compatible chirality (L-amino acids, D-sugars) we may nevertheless find that the specific epitopes that trigger the complement system are missing from the alien biosphere. In which case, the complement system -- our first line of defense -- won't work.

Worse, but somewhat less likely, is the possibility that the alien biosphere will involve peptides that are superantigens. If one or more of the key roles (e.g. the local equivalent of RuBisCO) is a superantigen, the colonists are shit out of luck.

Finally, you can't extrapolate from our experience of human penetration of the earth's biosphere to our compatability with an alien biosphere. We're the end product of about three billion years of molecular arms races; of course we're not susceptible to most terrestrial pathogens! Despite which, we've had numerous serious plagues in recorded history, with mortality levels up to or in excess of 90% (during the European colonization of South America).

We have no precedent for evaluating the threat caused by alien ecosystems because we don't even know what their biochemistry would look like, whether they'd have a prokaryote/eukaryote split, whether they'd have sessile photoautotrophs and mobile chemoautotrophs or some other arrangement, or whatever. But I suspect either we'd be unable to interact with an alien biosphere ... or it would dine out on our bones.


I think Our Gracious Host has underestimated the necessary size of a sustainable, early-21st-century-tech Ehrlenmeyer flask society; but not by an awful lot. The problem is that a "flask" (that is, entirely isolated from external threats/influences and kept in laboratory conditions, with carefully measured reagents and uniform heat sources) isn't the proper measure of "sustainability"; that, instead, is what happens outside the laboratory, after one allows for (in no particular order)
* dissent, both "loyal" and "disloyal"
* natural and unnatural disasters that destroy otherwise efficiently concentrated nodes of expertise and/or resources
* external opposition — whether we're talking about other colonies also fighting for some of the same resources, intelligent alienses, or just mutated bacteria is immaterial
* exhaustion/unavailability of resources, and the necessary efforts to develop replacements — just take a look at German efforts to deal with petroleum constraints during the Second Thirty Years' War
* impurities and side reactions in the economic flask — not just "market inefficiencies" (or even just "dissent") but actual barriers thrown up by the wonderful world of human behavior

These are nontrivial. I participated in a parellel exercise a few years back for more-official purpose, and the conclusion reached then was closer to 300m as I recall... for some value of "conclusion," as it wasn't exactly a consensus!


Whilst I follow the argument and the numbers, I find myself not agreeing with the conclusions. It took a while to work out why.

The only two times in something that gets even close to a historical record that humans have colonised parts of the Earth's surface with no previous human habitation are the Scandinavians into Iceland and the Polynesians into what's now New Zealand I think. The pilgrim fathers had Pocahontas and so on, and the original settlement of the Americas is prehistory.

If we look at those two cultures though, they're both quite warrior driven, they're both cultures in which sea-faring is basically an every-day activity for many or all of their members. Their colonies almost certainly started as fishermen/warriors establishing a toehold, then expanding it, then bringing families and so on, and they did so in vessels that carried <100 people when colonising and potentially <<100 on the fishing trips and the like.

There was doubtless an attrition rate, I don't know if there's any indication of what it was, but it wasn't bang, instant self-supporting colony, there was always resupply and help.

OK, spaceships will always be more complex than dugout canoes, balsa wood rafts, longships and the like. But launching people into space is currently a major effort taking a reasonable chunk of the resources of the major industrial economies per person. Thinking about what we're doing now, we're nowhere near paralleling the cultures that did anything similar in the past. By the time we are, maybe every red-blooded American boy will be able to repair all the systems on a spaceship in the same way that 20 years ago they were supposed to be able to repair their cars. The twiddles and computers that stop them currently repairing a car will be hacked and they'll be able to fix a 10 year old car with all its computers even if the newest models require specialists.

You won't try to launch a ship with 100M+ specialists, you'll launch 200,000+ ships with <1,000 people on board each, allowing for attrition and you won't do it all at once, there will be a wave of terraformers (until they find Aliens and leave only Newt alive), then a wave of miners and so on, making the colony increasingly capable of supporting its society without the extra influx, until you get to the point where you're not looking for specialists to fill critical niches, you send families and let them sink or swim in the same way that if you and Feorag moved to Australia now, it's sink or swim, but if your parents or grandparents had gone just after WWII they'd have been subsidised.

That doesn't directly address the issue of this post I realise, but it sort of does too. At some point the supported culture becomes largely self-supporting. New Zealand and Iceland have both managed it for a few hundred years. Sure, they're both currently in international trade agreements. Sure things would be grim if they were suddenly cut off. Sure things wouldn't stay identical to today. But from the much smaller numbers in both places I rather suspect that something we'd recognise as a reasonably advanced, if different, society would appear. I'd be inclined on that basis to move your estimate down by a couple of orders of magnitude... 1M to 10M might do very nicely.


Excellent post, sir. You've managed to write a compelling argument while pointing out the infeasibility of the genre in your chosen profession. That, my good man, takes cojones, as they say.


I think we are going to have to disagree on this one.

IMO, you are invoking the "gray goo" argument, just transferring the nanomachines tech to alien biology.

The issue is to what degree alien biologies can interact with each other. If they are so different, then we probably have little to fear. Their bacteria invade our bodies, cannot feed and replicate, and die. Likewise their viruses cannot hijack our replication mechanism. We remain shadow biologies to each other.
In this case, it is irrelevant whether our immune systems react or not. The invaders cannot replicate to do us harm and do not need to be attacked.

Regarding our adaptation to earth pathogens. What you say is true, except that bacterial and viral evolution itself is much faster than animal evolution. Yet despite this, viruses have difficulty even transferring to different hosts, e.g. avian flu's to humans. Novel bacteria and viruses should be somewhat like aliens and yet we just do not succumb to them except on rare occasions, and again, they need to be compatible with our biology to do that.

What about precedents for determining the assertion concerning different biologies? At least to some extent, our control of biology will be vastly superior to our current state by the time any humans attempt to colonize the stars. We will have a lot more experience of synthetic biologies and their interaction with humans, within the limits of our our imaginations.
We will also have a lot more data about potential molecular shapes that could be a problem, including the current worry about nano-particles of various stripes.

So the real worry will be organisms that will "dine on our bones" yet whose biology we cannot interact with.

The solution is quite simple and low tech. You simply send model animal species to the target planets in advance (pick your mode of transport and state), expose them for several generations and report the results. If the environment proves extremely hostile, don't go there, and choose targets where the animals thrive.

Alternatively, as you have discussed on other threads and I indicated with the idea of subsurface lunar colonies, we might terraform worlds, or parts of them, in advance of human colonization. These will probably be abiotic worlds and quite safe for our us. Although paradoxically, the new niches might open up a possible avenue for rapid evolution of our own microbes to attack us.

Those are my basic responses as to why I really do not think biology will be the overriding issue.


Abiotic worlds aren't "quite safe"; they're utterly lethal to us, without extensive and expensive life support equipment up to and including a portable biosphere.

Although I don't expect unmodified humanity, barring breakthrough discoveries (wormholes? magic tech of some other kind?) to ever get to visit worlds with a divergent biology more sophisticated than archaea.

"People just don't stop breeding if they see a perspective for their offspring."
In the developped world, family sizes (and fertility) aren't dropping because people don't see a perspective for their offspring.

They drop because of one major factor: the cost of sustaining the very life style Charlie ask us to maintain in this exercise.

Right now, it takes two working spouses to maintain the household in style (hence, one of them can't stop working to raise the children without a huge loss to lifestyle), and any children significantly consumes your resources (think of the costs of college/undergrad level education). That's why the family size is dropping, and the childbearing age is rising (with the attendant fertility problems - it's much harder to get pregnant at 30 than at 18).

(the utopian/dystopian solutions to this problem are left to your imagination - or a later exercice from our gracious host)


I was going to say the same thing. If I work a ton of overtime, I tend to put in 60 hours a week. 52 x 60 is 3120 hours.


If we assume rate of progress is really sublinear in population size (as your comment suggests), then a modestly smaller population (the ~2 orders of magnitude from Earth to this hypothetical colony) means a huge step down in rate of progress - even if they could maintain 2010 technology, in another generation they'd be hopelessly behind the folks back home without alot of intangible support.

On the complexity -> inflexibility -> collapse theory, this sounds a bit like this book I've seen Clay Shirky pointing to but haven't read yet: "Collapse of Complex Societies" by Tainter. Know anything about it?


You are right Charles, it does not matter what the name of the planet is, I just used "Mars" to give the right mental image for where everybody else went.

Humans have in fact terraformed Earth several times to varying degrees, most notably after the glaciers, less so after "the black death" which cut the population by a very large fraction and 1918 which left some societies severely short of hands.

However, we do get to select our colonialists, even on this planet.

In the pandemic scenario, our technology gives us an clear advantage which for instance rural china does not posess. See the H1N1 response efficiency as example. We did not have before.

It is also well documented that there are things you would have to be stupid to do during an pandemic, and equally well documented that if you do them, you become an autodarwinism candidate. Plenty of evidence from the black death and 1918 to make that point.

In the "lets look at 2210" scenario, we also get strong selection, either as chinas 1-child policy, which select on counterproductive parameters. Or the scandinavian welfare model, where selection, by now, is reduced to "any kid that gets born, through whatever technological means it takes".

I am not sure you can extrapolate that any population which declines due to higher standard of living will also ruin their gene-pool, although the Egyptian and Roman Empires makes compelling datapoints, but it will certainly take an active effort to avoid such a degradation.

So even though the planet is Earth, I will still maintain that 1 million humans will do the job fine.

But if you move from the present population to 1 million in a non-catastrophic fashion, it will not most certainly not work: There has to be a focusing event to get the one million humans onto the same page.


PS: Also, don't overestimate how many people it takes to make a gadget.

There is a huge effort spent making it a marketable/competitive/profitable gadget, which is driven falsely by a need for novelty.

I am sure everybody could get used to Rolls-Royce quality cars which lasted 50 years instead of 5, at a big savings in raw materials and work.

Similarly, all consumer electronics today has a carefully calibrated designlife of at least 2 but no more than 4 years. It is trivial to build electronics that last 30 or even 50 years, but there is no incentive to.


Usually in these kinds of statistics, the population gets broken up into extractive (food/mining), manufacturing, and services.

At some level it feels like the first is really a property of the environment the colonists are in (ie, what fraction needs to be farming to provide food) - we can assume our colonists went somewhere with easy-enough-to-get resources to replicate percentages on earth today, unless we're being mean to them.

Manufacturing (generalized to include software, etc) covers alot of the supply chain aspects of it - we could add up the global employment in this sector, but we'd still have to disentangle the fixed costs (building the factory) from the variable costs (takes less workers to build for100M than for billions). As a first pass, we could take only one firm from each category (since we're not worrying about how these colonists maintain the same rate of progress) and scale than down by assuming that each firm has X competitors.

Then the controversial point - services - I'll make the claim that these don't have the same minimum efficient scale issues that manufacturing does, meaning that it should be closer to scaling linearly with population size and activity. (For which there will be counterargument on education/training/etc, but it's just a first-order approximation here...)


"Abiotic worlds aren't "quite safe"; they're utterly lethal to us, without extensive and expensive life support equipment up to and including a portable biosphere."

You're setting your own terms here. If we accepted your definition of "worlds" and "extensive and expensive life support equipment up to and including a portable biosphere." this would apparently rule out colonizing the moon in underground cities, similarly Mars and all those hollow asteroid colonies, let alone the larger O'Neill style space colonies.

I don't buy that for one moment.

You've covered the "dropping a meat person" on a random world" in another posting. But in what way would an abiotic world be 'lethal'. If it was small without much atmosphere, certainly. But an earth sized world with a prebiotic atmosphere of CO2, water and noble gases? Hardly lethal unless you forget to take your O2 with you.

Keep the scale reasonable and you can terraform an abiotic piece of rock as long as you ensure that a biosphere is maintainable, e.g. inside and airtight hollow object. We know that is possible, because our planet does it all the time (admittedly with a large reservoir of organisms to draw from, and time). Add extra recycling kit if you need it, but as long as the energy inputs are sufficient and materials can be made to cycle properly, you can create a biospheric habitat. We may not be good at it today, but in a few centuries? A large abiotic rocky world would probably take a lot of time, but if Kepler and it's successors show us that there are lot of them, then we might consider large scale terraforming of the galaxy in advance of our expansion.

At this point we have no idea what the state of other worlds are. What is exciting to me is that we may have inkling of the probability of biotic worlds, perhaps biologically advanced ones, within a few decades.


"Collapse of Complex Societies". Thank you for the reference, I was completely unaware of this or anything related to it. I've just added the book to my buy list.


Sorry Vincent, I won't answer this - I'm too busy shaking my head right now.

How can anyone hold a festival with over a million people in a closed area and squeeze them through just two entry points, one of which being a 20m wide tunnel with a capacity of 360.000 people per hour in the very very best of all worlds?

The resulting stampede killed 15 people and this is pure luck.

Human stupidity is indeed infinite.


If you can design (ahem) a much smaller society to build, repair and replace things at the current level without a whole lot of innovation, then you get a ort of anti-singularity. Not quite stasis, but nonetheless a situation where most technology doesn't change very fast. The surplus of labor and capital required for continuous high-speed innovation simply won't be there.

Interestingly enough, this is the kind of universe that many conventional space operas and things like them seem to be set in. Old spaceships and new ones mostly interoperate, different groups don't have incomprehensibly different tools, and so forth. (You can argue that most of that is a user-interface fiction, but I'm not sure I really buy it.) The explanation usually seems to be that spacedrives, blasters or whatever reached the pinnacle of their development and then stayed there, but it seems equally plausible that when everyone is out diasporing or warring on one another or whatever they're simply not going to be doing those kinds of innovations.

Choice of tech will also be very important for a scaled-down society -- only in a culture where everyone throws most of their tech away every couple of years, for example, would something like current OLED make sense. And suppressing innovation in some kinds of technology may actually enable innovation in others -- for example, once moore's law got seriously going in the 90s, a large chunk of computer-architecture innovation just stopped, because you could get more bang from shrinks than from thinking about how to compute things. (It's started up again, but on a very different scale.)


Sorry Charlie,

I agree with Alex: anaphylactic shock is unlikely.

On the other hand, Polynesian islands are NOT the proper model for a colonization effort. Aside from the fact that the land supports almost no food naturally (most coconuts were planted), they're not bad places to colonize, compared with where the Polynesians came from: the Solomon Islands and New Guinea.

The best analogs we have for alien ecosystems on this planet are PNG, Australia, and the Andes. All had been settled for a very long time. The Andes produced a large-scale civilization, but never figured out how to make furnaces hot enough to melt iron (just got hot enough for bronze). PNG stayed Stone Age, despite millenia of contact with metal-bearing cultures (they found a 2500 year old Dongson bronze axe in a PNG village). Papuans ain't stupid by any stretch of the imagination, but that island is also a very hostile place to live. Australia is similarly hostile to indigenous civilization. The civilization it now supports was 90% imported.

Basically, that's what you're going to run into on any Earth-like world: even if you can eat the local biota, you're probably not going to find something to replace maize, wheat, potato, or any other crops. Things like air chemistry and temperature regimes will be uncomfortable. And there will be things that want to metabolize you (see next paragraph). Colonizing a new planet really will be like riding a rocket down to Papua New Guinea, or to the Bolivian altiplano, or the western desert of Australia. If we're lucky. Hard place to land, hard place to colonize. But doable, if you're concerned more about basic human survival than about building Niu Edenbura.

As for diseases: It's possible that we'll find something that spontaneously causes our immune systems to freak out (anaphylaxis). That's unlikely. We almost certainly won't be bothered by viruses and parasites, at least for the first few generations. Viruses only work when the genetic code is identical, and there's no reason to think that will ever be the case. Parasites work by subverting host biology, and while alien organisms will eventually figure out how to do that, they almost certainly won't be a problem at first contact.

The major diseases we will have to worry about are parasites and bacteria. Specifically, these will be decomposers, trying to metabolize our tissues. We're normally bombarded by these all the time anyway. The question on first contact is whether the alien decomposer ecosystem works better than our immune systems do. Best way to test that out is probably with a mouse or something similar.

As for rubisco being a super-antigen, I suspect that's like worrying about whether a monkey-wrench can double up as a improvised explosive. Chemistry constrains how a lot of these systems work. If there are green plants on this world and there's an oxygen atmosphere, I'm willing to bet that they'll be using something that looks a lot like rubisco. Ditto DNA, hemoglobin, and various electron transport chains. Alien DNA codes will almost certainly be radically different than ours, but that's a different story.



*Ayn Rand was a heavy user of amphetamines. In fact, her personality traits have an uncanny resemblance to the list of psychological side effects of amphetamine use..

Wouldn't that also make her like . . . HITLER! I'm not trying to do a Godwin, but it does seem that rather few people know about his amphetamine use (under a doctor's supervision!), and when told of this shake their heads and say, yeah, that would explain a few things. Of course, I'm posting from Missouri, the meth capital of the world (or so they say, by some metric), so there tends to be some familiarity with people going off the rails from this particular type of drug abuse. Note that Phil Dick was also known to use stimulants to boost his writing productivity.


"culture where everyone throws most of their tech away every couple of years, for example, would something like current OLED make sense"

I agree that technology choices should be adapted to the environment, and the fewer people there are, the simpler and more repairable it should be.

A case of where technologies starting going backward was the Viking settlements of Greenland. Metal was very difficult to get and I think they started using other materials for cutting implements as their metal ran out or was reduced to the simplest bog iron.

I’d guess that you’d need 90-95% less people for maintaining the devices, because most things would be done the same way. No 5 cars with most parts different, but rather one type of car, one type of train and one type of train, all using the same parts wherever possible. And the machines creating the parts could mostly be built from these same parts, too.

You mean like all cells having essentially the same types of mitochonria, ribosomes and suchlike? Hmmm . . .



A better frame than space colonization might be: it is the year 2200. The demographic transition has gone too far, and families of 1 or 2 children are "normal", more than 2 being seen as weird -- consequently, Earth's population has dropped back below 4 billion and is still trending downwards. How low can it (safely -- for human civilization) go?

This seems like it's a different question from the original one. You have extra parameters in there like how long of a bounceback is allowable, for example.

You've also got, (again) the problem of the kakistrocary: Say that you've only got a finite amount of fossil fuel (or computer chips or whatever) left and you know that you've got to have a certain percentage of that resource to restart and level out at a certain reasonable, albeit possibly lower level of tech. One way, everyone behaves like an adult, resources are rationally allocated and everything works out fine. The other way, a certain percentage of absolute wankers do everything they can to delay this reallocation, because they don't want anyone else using "their" stuff, and they intend to use it for their own immediate gratification right now, never mind a future fifty or a hundred or two hundred years away.


"kakistocracy" - My new word for the day. Who says you can't learn nuffin' from this intertubes thingy?


I think the premises for your question is wrong. The reason for the complexity and interconnectedness of our production systems is that we have multiple of billions of workers in a semi open global economy. This is the most efficient way of organizing yourself in the current situation. In another situation we would need other ways of organisation. That does not mean that we would not be able to sustain and develop a technological civilization. Not many generations ago we where organised in much smaler semi self sustained units. Design, organization and technology would need to be changed from the bottom up, but I think we could get bay with 100 thousand.
Remained me I reread Days of the Trifids a few days ago. They grapple with this question there also. They put the figure to a few hundreds. So in 50 years we have gone from a need of a few hundreds to 100 millions. What happened?



As for rubisco being a super-antigen, I suspect that's like worrying about whether a monkey-wrench can double up as a improvised explosive. Chemistry constrains how a lot of these systems work. If there are green plants on this world and there's an oxygen atmosphere, I'm willing to bet that they'll be using something that looks a lot like rubisco. Ditto DNA, hemoglobin, and various electron transport chains. Alien DNA codes will almost certainly be radically different than ours, but that's a different story.

Well, we simply don't know. At one extreme any moderately habitable world with an established biosphere looks at Earthly life as ambulatory chocolate eclairs[1]; at the other, like highly evolved automobiles trying to eat human beings for their daily bread. No one knows.

And that's probably my number one reason for doing the whole interstellar exploration thing. If you want to study alien life much beyond the bare-bones fact that it exists, you pretty much have to go out and look at it up close and personal.

[1]And of course, this one's already been done many times in classic sf, Murray Leinster's "Proxima Centauri" published sometime in the mid-1930's for aliens like humans, or C. Smith's "From Gustible's Planet" where predation runs in the opposite direction. More modern treatments get progressively nastier, I've found.



"kakistocracy" - My new word for the day. Who says you can't learn nuffin' from this intertubes thingy?

Appropriate, innit. While a lot of new terminology is contrived, clever-clever, and too precious by half ("Gretchen, stop trying to make 'fetch' happen! It's not going to happen!"), a lot of people know exactly what you're talking about with this one and find it highly apropos for a number of current situations.


Information is a key component that was left out of Charlie's original question here. I see two HUGE questions about information that affect any sort of response to this question very drastically.

(1) Do the people have access to our current body of recorded information?
(2) And if so, is this information segregated into silos along corporate or government lines, or can we suppose truly unrestricted access to *all* of it?

Our civilization has an enormous wealth of information in the form of knowledge, source code, manufacturing setups, etc. For example: It takes thousands of engineers working over decades to create a typical OS. iPhone, Windows, Linux, Symbian, etc. HOWEVER: it generally only takes *one* engineer with the right access to actually build the thing once it's all in place. And honestly I'd bet that one engineer would be enough to manage quite a lot of bug fixes, as long as they aren't required to develop new features. (Of course, finding an engineer who doesn't want to tinker and work on new stuff might be a challenge.)

The other issue is whether the information is chained up and segregated into silos like it is today - source code is hidden, trade secrets are kept locked up, etc - or if you have truly unfettered access to it. With truly unfettered access you can actually have one person maintain a ton of stuff. One or two people might be enough to build and bugfix *all* the operating systems on the planet. (Of course you'd want more than that, for redundancy, training, etc. The point is the number is really quite small.) But if the information is in silos, you'll essentially need completely duplicate teams with the same knowledge for each silo -- much like we have in our civilization today.


""One or two people might be enough to build and bugfix *all* the operating systems on the planet."

So why is Microsoft paying all those people to maintain Windows? Don't you think they might do that if it was possible and divert the rest of technical staff to developing new features and products?


I think it depends on the nature of the technology. One of the things we're seeing is that there is less need for specialization as technology advances. The latest generation of manufacturing technologies, not yet widely deployed, are incredibly simple and inexpensive. There's dependence on information technology, but that is highly standardized, and one way or another, we can make do for everything else. So in 1950, maybe a huge number of people to support a technological infrastructure, in 2050, maybe not so many.


"Alternatively: the Martian Hyper-Scabies epidemic has wiped out 50% of the human population, planet-wide, before burning itself out. After the mass graves are filled, how sustainable is our level of civilization? (Ignoring environmental externalities for the time being, and collateral damage due to things like supply chain collapse/power interruption during the epidemic.)"

To bypass the point completely, I suspect that the answer to that one is simply "it collapses almost immediately", because statistically speaking the chances are that you've lost all members of at least one small-but-vital specialised group. (A world with no automobile designers is going to be ok; a world with no water sanitation engineers or that's completely lacking people trained to repair high-voltage electricity transmission grids is going to be in serious trouble very fast.)

If we assume that the society retains sufficient members of every vital group to avoid that collapse (and train replacements), then things might be different.


Actually Alex, there are a bunch of cases where groups lost technology:
--Arctic: Dorset people lost bows
--Polar eskimos lost bows and kayaks (the relevant crafters died in an Influenza epidemic).
--Several Micronesian groups lost their capacity for making voyaging canoes, notably Kosrae, which is actually a high island, not an atoll.
--Some researchers have argued that the Polynesian ancestors were actually the Bronze-Age Dongson, which would have meant that the Polynesians actually abandoned the use of metal as part of their adaptation to an oceanic existence (not that you can get tin or copper on most Pacific islands), along with pottery, weaving, stone flaking technology, and possibly others.
--There's also Easter Island, although when you lose your trees, it's hard to keep making wooden canoes.

So yes, it's been done. We can also talk about Dark Age Europe, let's not forget. They lost their civil engineering capacity.


In a mass die off situation, who survives is far more important than the numerical total of survivors. Having a billion spare people won't help if all (or most) of your critical path specialists are dead.

If everyone who knows how to design and build integrated circuits is vaporized by alien space bats, the survivors will regress to pre-computing technology levels for a few years (or decades) while the institutional knowledge to build ICs is recreated. If the knowledge base can't be recreated fast enough--or 1970s level technology can't be brought on line fast enough--what's left of civilization will crash even further back.

The safest numerical answer is that maintaining current standards of living will mean keeping everyone who is currently involved in producing substantive goods or services. This excludes subsistence farmers, various forms of social parasites (e.g. pirates and rentiers), retirees, and others who don't participate in the labor force. I suspect the final number will be in the billions.


Charlie wrote: "So no, I don't think we can do without musicians, writers, game designers, toy makers, and probably professional athletes too. Although they may be left out of the planning apparat for a colony, other than by pencilling in some hours a week as "free time" and assuming that some people will fill it with something other than drugs, crime, and rioting."

Can you imagine a world where rioting is a professional sport? Or is that modern-day Europe's destiny, since the US is too pacified by guns and religion no matter how much more it may be wanted here?


Someone just violated Godwin's rule.

For starting a colony and getting it viable, I think you need around 20 mill still. Again Aussie and NZ. NZ grows the food better than Aussie -- we have more rain and the only external in put we need is nitrates, which we classically got from Nauru.

The classic way of doing this is a series of ships -- around 1000 per -- over 20 -- 50 years. And then encourage couples to have 4 to 6 kids each. Charles, I suggest you look up NZ early history -- we colonised (my great-grandfathers grandfather was in the regiment that supported the first wave into Taranaki) against armed opposition.

One problem, however, is breeding. Given a choice, women don't like spending most of two decades pregnant.

After 200 years, Aussie has a mature industrial base. Both countries has post industrial tech. We would need to resurrect heavy industry -- we have Iron smelting (ironsands) coal, oil...

It would depend on communication and social factors. Anti tech parts of teh culture would have to actively discouraged...


Charlie and Other Alex. Let's assume some basic non-stupidity on the part of our colonizers. Before committing billions, (more probably trillions) of dollars to a colonizing expedition, the colonizers will do some basic science first. Checking the biosphere of an earth-like planet very, very carefully, and handing epidemiological issues would definitely be on my list of "stuff to do before planting a colony."

Merely assuming that biotech keeps pace with the rest of the sciences gives us the ability to tailor an immune system against both specific and general threats from an alien biosphere in a hundred years or so. Careful use of randomness when spinning our new genes ensures that some subset of our colony would survive any single threat.

IMHO, the real problems would be more likely to come from some variety of bacteria which we had imported mutating under the strain of a new enviroment, but either problem would kill our colonists just as dead...


Their [Iceland and NZ] colonies almost certainly started as fishermen/warriors establishing a toehold, then expanding it, then bringing families and so on,...

The conventional wisdom on Polynesian in general (and Hawaii and NZ in particular) colonisation is that they were discovered as part of deliberate exploration, then colonised by families. That you assert that an earlier (unsupported) model is 'almost certain' is telling.


My alarm goes off for work in five four hours and I'm still up and playing with these ideas.

I've been approaching this from a different perspective: who already does large chunks of keeping our high-tech civilisation running, how big are they and how much external support from smaller "boutique" suppliers do they need?

For example, Boeing, Intel and General Electric could theoretically, between them, design and manufacture anything from a pocket calculator to a jet fighter to a nuclear power station, given the raw materials. They could do this without very many external suppliers. At the moment, they don't - they rely on others to provide low-end widgets and high-end specialised sub-structures, and they churn units out by the boat-load.

But if the world population is the staff of a handful of generalist megacorps, a couple of hundred more specialist suppliers each a hundredth the megacorp's size, their families, and the medical/educational/cultural/administrative infrastructure they need, then you no longer need or want to be turning out a new 737 every working day - there's no demand. Instead, the same workforce could make one a month, and spend the rest of their time making the bits for it in-house, further reducing demand for people.

I've been doing some Googling/Wiki-ing on the sizes of companies like Siemens or Tata, and poking the Office of National Statistics to get a feel for how a modern western economy is structured (read: playing with numbers in an arbitrary fashion. They have downloadable spreadsheets which make it too tempting to resist). Apply some multipliers for spouses/kids/grandparents, more for admin and for "life-support" sectors like food production and healthcare which are strongly coupled to overall population size, and Bob's your thingumy.

Here's my bottom line: the back of my envelope says a bit over 0.1% of the Earth's current population at the low end, up to a bit under 1% at the top. Hong Kong to South Korea, if you prefer.


Do you know of groups that deliberately reduced their technology?

The Amish come to mind. But I was also thinking that perhaps settlers during the American expansion west did so too. For example, prairie schooner wagons as a much simpler, repairable technology than perhaps city carriages?

For a colony on a new world, assuming that technological complexity is a function of population, then a colony might need to have simpler technologies, or at least more manageable ones, than those of the mother world, until the population and associated expertise and infrastructure develop. Some technologies might be biological, e.g. organisms that produce needed pharmaceuticals and just require simple extraction rather than synthesis.

So why is Microsoft paying all those people to maintain Windows?

Because the OS has to continually be updated to deal with new everything - hardware, software, new models of how information is used, etc - which means it's a *lot* of work just to stay in the same place. Technologically we're in a Red Queen's race right now - things are changing constantly, so you have to run as fast as you can just to stay in the same place.

Eliminate progress and make your requirement into pure maintenance, and the majority of that evaporates.


Bear in mind also that those voyaging canoes represented a huge chunk of the man-power of those communities. For instance, I recently heard than one person can make a metre or two of sennit (coconut fibre) rope in a day. A big canoe has hundreds of metres of sennit rope holding it together. Building the whole thing probably soaked up the entire excess labour pool of an island for a year or more. Then you needed a navigator who'd been trained and drilled from childhood to find his way across oceanic distances with pin-point accuracy.

Not try to stretch the point too far, but I think we need to remember just what a big deal some of this stuff was, historically.

@Charlie, heteromeles, Alex Tolley: you may not be aware that some terrestrial organisms do use D-amino acids in strategic places, specifically because they're so hard to deal with: capsule walls in bacteria, platypus venom, things like that. The aliens are already here, and we only go into anaphylactic shock from them now and again...


Absolutely fascinating stuff.

I've got some thoughts on other inputs. I think you may be doing a lot of farming!

Lack of hydrocarbons might be a huge issue, just for the chemical and plastics you'll need. How far will genetically engineered algae get you? Big greenhouses for sure. And not much paper or post-it notes I guess... What about something as basic as clothes? Either grow cotton or you're back to hydrocarbons again - you have to make your regulation jumpsuit out of something.

Back in the dim and distant past, I hung around the pharma industry. Once you know what you're doing, most drugs are really cheap to make (on the order of 1-5% of selling price). I know that at one point, the most expensive drug on the market ($100k per patient per year) was made in a facility which had about 20 workers. There's lots of specialized knowledge, especially in biologicals, but it's not too difficult.

The other big questions are if you're aiming at North Korea or USA. I'd think NK is probably doable with less than 1 million people, USA probably 10 million. I'd think factories are more general purpose, less mass production, and more serial production type (cars for a few months, a few planes, ships, etc) - more like Ford Model Ts and less like today. One cell phone, one model of car, one or two sorts of plane, and so on. As others have said, one sort of computer processor and many coding options. You can drop the complexity down a lot like that. Obviously you'd probably loose efficiency and designs may be clunkier, but they'd work.

For the minerals - could you simply work round not mining something? Say no copper - use silver? Or no aluminium? That would reduce your extraction problems, but increase design issues.

But then you need to decide what society is for - are you looking for a stable back-up of the human race, or actually pushing forward science, engineering, pharma? If you're pushing forward, you need far more people to support the research of the few.

Cool stuff in any case.


' *Ayn Rand was a heavy user of amphetamines. In fact, her personality traits have an uncanny resemblance to the list of psychological side effects of amphetamine use..'

"Wouldn't that also make her like . . . HITLER! ..."

Ein Volk, Ein Reich, Ayn Rand?


The idea that specialization on the scale prevalent today is necessary is wrong. I'm with Heinlein: specialization is for insects. Any intelligent person with the sort of general education that comes from learning to read (two months education) plus two or three thousand hours of reading a variety of materials, plus a few hundred hours of apprenticeships can do the work of any of dozens of specialties 80% as well as the average specialist today. That kind of education only takes up to the age of ten or twelve, allowing some time to fill in the chinks of what self-directed learning doesn't get to with off-the-shelf canned courses.

Most modern specializations are chasing rather small comparative advantages or are rent seeking, erecting barriers to entry with false requirements. Pharmacy is an excellent example of this kind of specialty. 98% of the economic value of the job is checking the name and dosage of the drug and putting the right number of pills in the bottle. Most of the rest is better done by a computer than by a human. Compounding pharmacists are even rarer than doctors who make house calls. Accountancy is another - most of what separates the job from mere arithmetic is nonsense created by the profession to keep themselves employed and the merely numerate out. There are other specialties that really do require a lot of ability, knowledge and experience - pathology, vascular surgery, neurosurgery, chemical engineering - but the barriers to entry are always higher than the actual requirements.

By having a society of generalists or at least jackleg multi-specialists using a preexisting library of knowledge, the number of people needed will scale very closely to the number of people present from very small numbers (tens of thousands, or even less for up to a few decades of population expansion), so long as there is a sufficient amount of preexisting productive capital (books, databases, tools, seeds, non-perishable materials, frozen embryos and sperm, breeding stock, etc.) to get the society over the hump to a sustainable population level. The fewer people, the more stored capital is needed, and the more specific critical types of capital there will be. With many billions of dollars of very carefully selected and developed capital per capita (including frozen embryos and sperm, and teaching/babysitting machines) the minimum could be as low as a handful of fertile women. With a few hundred thousand $ worth of capital each, one could do with perhaps 25,000 people. With just post-apocalyptic wreckage, a few million people. These would all require a couple hundred years to get back to a fully modern technological level.

If the scenario is a slow population crash with no dark age, and the target level is functional - the useful gadgets, drugs, industries and knowledge, without necessarily having fifty different kinds of toothpaste and five hundred different models of voltage regulator, and working lifetimes and intelligence distributions are similar to today, then somewhere between 250,000 and 25,000,000 people should be enough to stave off a dark age for centuries. (Provided they don't fight each other too hard.) (With an average IQ of 130 and a useful lifespan of 150, then maybe as low as 50,000).


Thoughtful piece, I especially like the part about the interconnectedness between modern technology and the networks of organizations necessary to build, use and maintain that.

Nevertheless, one could take your question and turn it around 90 degrees, coming to something like: "How many years has one to go back in time to find a civilization "do-able" with only 100.000, 1 mio or 10 mio people?"

In a way, you did that with the comparision of older cars with more modern variants.

And what kind of technology and social innovation has to be thrown out of the window?

E.g., is a not smart but digital mobile phone do-able with 1970s technology level? What about 1950s technology level?

What about the PC? The internet? Photovoltaics? Aviation?


With the post sudden apocalypse scenario, there could of of course be immense practical difficulties in collecting (no transport! no comms!) and organising and focussing the group, even assuming that enough had survived with the right skills. So it could be that even though N is the minimal functional group, you need 10N or 20N to survive just because the N will never coalesce. So if you lose move than that, you're still sunk.

(I find this speculation really interesting. There was a BBC series recently, "Survivors" which assumed near wipe-out by disease, but it focussed on relationships not practicalities - people kept going off, getting into trouble foraging and having to be rescued, but it didn't occur to anyone to find the nearest Maplin and grab a few sets of radios so they could keep in contact.)


"The Day of the Triffids" is a work of fiction, not a RAND Corporation report. In particular, John Wyndham had a not-very-deeply-concealed disaste for the English working classes, and may have been writing out a wish fulfilment fantasy along the lines of "why can't those horrible teeming others simply vanish -- I dunno, go blind and be eaten by walking carnivorous plants -- leaving England pure and empty to those of us who know what to do with it?" rather than writing a plausible blueprint for a post-catastrophe future.


I ... disagree. (And I note with interest your comment on the dispensability of pharmacists. Are you trying to wind me up?)

Because the OS has to continually be updated to deal with new everything - hardware, software, new models of how information is used, etc

Actually, it's a bit more complicated than that. You would probably need a very small (in relative size) team to maintain something like XP. More so for Vista, more so for the brand new Windows 7. Maintenance in the software world tend to follow a asymptotically decreasing curve: the longer a software has been in use as it is, the less you need to maintain it.

The reason Microsoft has so many people is what I personally nickname "the Cramer Effect" (after encountering it spelled explicitly in one discussion by John Cramer about large particle accelerators, and how to build them). It's something that was more or less explicitly mentioned by Charlie in his dissertation.

Building modern stuff takes a lot of specialized knowledge and work culture. Merely having the blueprint of an airplane isn't enough; you need also practical experience in making the stuff to avoid the myriad pitfalls in making at the proper tolerances, and bringing it all together in an efficient and economical manner. Cramer mentioned it when discussing the abandoned linear accelerator in Texas. Building it ten years later would cost double - because all the accumulated know-how and trained workforce would have vanished.

Microsoft, among other things, is in the business of making Operating Systems. A modern operating system is the collective work of a thousand (or more) dedicated technical people. These people need to have acquired know-how to be efficient - you don't come into Microsoft and start immediately working on kernel internals. Unless you've been working on some kernel internals somewhere before, that is.

Ok, Microsoft ships one OS. Let's call it Windows XP. What do they do now with all those people? Answer: they put them on working for the next OS. Note that the market didn't want or need a new OS. In fact, the market actively resisted Vista. No one replaced XP with Vista; the only way people got Vista was that their brand new computer came with Vista. Companies aggressively downgraded each and every Vista license into an XP installation. So why make a Vista? Answer: because if you do not make Vista, your OS designers are going to go somewhere else, and work on something else. You can't put them all to work on Office or make games for the XBox, so you'll hemorrage people. And when the market is finally ready for a new OS, you no longer have enough OS specialists. You need to rebuild an OS team. And forget calling back all your old employees - most of them have new work, and they're not going to drop their work to all come back.

So what you do? You make a new OS, every three years. Regardless if it's necessary. And then, you have to spend a lot more money fixing the new OS. But that is the price you have to pay. If you do not, when the market becomes ready for the next OS, you've lost all your advantages. And maybe Ubuntu, or OS X-and-a-half, or picoBSD, or whatever will grab the slice this time.

That's what Charlie alluded. It's not enough to have the knowledge. You have to have a workforce, trained and experienced. If you don't, you lose a lot, and you're a lot less efficient - which means you need a lot more people for the same task. And lot more money. Which makes stuff more expensive, and thus less pervasive.

Of course, if we assume we want to maintain, not improve, a lot of professions can be let go. You don't need Microsoft to design a new OS every three year to maintain our current state of living. You can slash the entire Boeing design bureau - but you need to retain most of their worker. If the population of Earth got slashed by 50 brutally, the temptation to reuse the current plane fleet is great - we'd have enough planes that you could mothball to last us a century before having to make new ones. By then, of course, no one would know how to make new ones; you'd have to rebuild the industry from scratch.

Personally, I think we could probably manage a "kind of" modern, early 21st civilization with as few as 100 million survivors. We'd lose a few bits here and there, but overall, things would be, technically speaking, ok. Socially, they'd get pretty weird of course. And we'd get some kind of dark ages for a while; I suspect our civilization would take at least two or three generations to restart significant progress.


Yup, that's roughly in line with the way I'm thinking. We can hold onto what we've got with many fewer people, but actually starting up progress again would be painful in the extreme.

Here's a piece of the puzzle: why NASA exists, and what it does.

Forget the early days. From 1968 onwards, the glory days were over; even in 1968 NASA began winding down from the Apollo surge. In fact, between 1968 and 1974 the US aviation industry shed about 80% of its R&D jobs.

Some of that shedding was due to the workers who'd driven the peak of aerospace development -- between around 1920 and 1960 -- hitting retirement. But a lot was also due to fundamental physics: if you drive metal through air at >Mach 2.2, it gets very hot and tends to melt. And if you want to accelerate out of atmosphere and do anything much better than circa-1960 ICBM technology, you need nuclear propulsion, which is encumbered with all sorts of political problems. So progress in the area of increasing performance hit a barrier.

Post-1972, NASA exists to maintain an engineering capability in being: the capability to design, build, and launch deep space missions and manned spacecraft. It's horribly inefficient and bureaucratic, but it's there -- and it means there's a pool of talent to draw on should it ever be needed for national strategic purposes.

Private launch companies like SpaceX are a lot leaner and faster-moving than NASA; SpaceX has built something (the Falcon-9) in the same class as the early R-7 Soyuz launcher with just 1200 employees in about six years. But it's unlikely that they'd have been able to do so without (a) the institutional knowledge about space launch technology generated by the earlier big government launch programs, and (b) a pool of skilled employees to cherry-pick from. And NASA, in the glory days of 1958-68, could in turn draw on the gigantic infrastructure that built the US ballistic missile force, going all the way back to Wernher von Braun and the German Rocket Club in the early 1930s.

If we wave a magic wand and kill off all the rocket scientists and satellite engineers, could we restart the commercial space industry and the unmanned exploration program? Or even a manned program? Yes, almost certainly. But it would be a much longer and more painful experience than building a new launch system today, because there's a steep learning curve and a lot of uncodified knowledge -- stuff that doesn't get written down, or (more importantly) written down and published -- would have been lost.


"I ... disagree. (And I note with interest your comment on the dispensability of pharmacists. Are you trying to wind me up?)"

Well... you dispensed with dispensing, so to speak - I guessed you weren't enthralled with the profession, and certainly know the giant gap between the education needed to get the license and the practice itself.

My basic points are:
people can learn and do far more than they are allowed to do in the present system, allowing fewer people to do more things, albeit somewhat less efficiently;

stored knowledge and other capital can substitute for labor,

and the specifics of the initial situation and how you define the level of society you are trying to achieve swamp all other aspects of the calculation so that you can make a case for practically any number you like.


Accountancy may be a bad example. In the UK at least, there is NO statutory restriction on the use of the term "accountant". Anyone can set up as an accountant without any qualifications or experience whatsoever. There are some minimal requirements if you want the Revenue to treat you as a tax agent, but these mostly relate to antimoneylaundering procedures rather than "professional competence" as such. Don't get me started on lawyers, however ...

Although even there, within living memory one did not need any kind of degree to train as a lawyer.


"If we wave a magic wand and kill off all the rocket scientists and satellite engineers, could we restart the commercial space industry and the unmanned exploration program? Or even a manned program? Yes, almost certainly. But it would be a much longer and more painful experience than building a new launch system today, because there's a steep learning curve and a lot of uncodified knowledge -- stuff that doesn't get written down, or (more importantly) written down and published -- would have been lost."

Conversely one might argue that the "genes" for space travel have been fossilized by Nasa's past success and that new variants have had a hard time being generated in it's shadow. It might be a lot better to force "speciation" if the goal is widespread human space exploration.

In the wider context, how many of the useful "knowledge genes" of modern culture are there in relation to the much wider "support structure genes" which could be stripped away in a new culture and reinvented?


Being a pharmacist isn't about dispensing, any more than being an airliner pilot is about getting on the PA system to tell the passengers what the weather at their destination is like. Put it another way: about 10-15% of hospitalizations are physician-induced, in no small part because much medical training focusses on diagnosis and physical treatment much more strongly than pharmaceuticals (and because most people are, generally, clueless about medication). It takes roughly four to five years to give somebody a graduate-level understanding of clinical pharmacy. While some of the educational framework could be eliminated if we were trying to impart that knowledge to medical doctors, it'd still add 2-3 years to their training, which is already close to the maximum feasible duration for a practising profession.

As for your point about stored knowledge ... it takes labour to take stored knowledge and turn it into an active capability; we call this schooling, or academia, or training, and it's not free. We can quantify the opportunity cost of having to learn a speciality from books and by trial-and-error practice and self-training, relative to being able to undertake directed learning (with books to back it up) from an already-knowledgeable teacher, and it's high enough that the teaching profession still exists.


I came across a similar problem in writing up some background for a hard SF game (2300AD). The premise was that a relatively new colony broke away from its founding nation and was still able to survive without simply realigning itself with another colonial power (link here if you're interested).

My attempts at solutions followed some of those already suggested.

  • Use of small scale flexible manufacturing (more rapid prototyping than mass production) with a massive database of slightly older designs.
  • Use of a restricted materials set (loose most light metals and composites for titanium).
  • Use of last generation tech with all the bugs ironed out and an open source version of IP.
  • Use general purpose chips (such as FPGA) to replace specialist chips. These weren't as good as the specialist item they would be good enough.
  • While I didn't posit a command and control economy I did put in some fairly heavy drivers (religion and economics) to get the colony to adhere to the policies.

I didn't work out the minimum population requirement (I assumed that you would be OK with about 3 million) and it was by no means a closed system. I think you could get away with significantly less than our host's proposed 100 million but only if you slimmed down the breadth of your in use technology. You would certainly retain the knowledge of the rest so that it was available in the toolbox and there would be a fairly high risk of reverting to a dark age if you lost something key. The larger the population the wider the sustainable tech base the more resilient it would be.

The answer is therefore dependent upon the scenario. Deliberately planned colony with designed tech base - smaller. Catastrophically damaged Earth (no pre-planning and a disaster to sort out simultaneously) - larger.


You need a certain minimum number of people to knap flint (1), a certain minimum number of people to run an 18thC canal lock (about 6)

One to knap flint, assuming he's in an area with flint. But good stone was traded across enormous distances, and there were well-established flint mines in places like Grimes Graves. Tools and raw materials were produced by specialists even then; iIt wasn't as simple as we now see it…

How many men to build that 18th century lock? Because you included the chaps building the railroad…


Let's turn this around. Charlie can dispense (and possibly compound) drugs, plus he can maintain code and create new code, and probably do a lot of other stuff with computers -repairs, networking, etc. With very little training, he can write documentation, and he does an excellent job of reporting what he sees around him, so he can also function as our colony's journalist/historian.

Going a step further, and given Charlie's pharmacy training, I'm guessing he can be easily trained as a nurse, paramedic, physician's assistant, veterinarian's assistant, or "barefoot doctor." I'd make a similar guess about training in electronics repair, (assuming he doesn't have those skills already.)

Plus I'm guessing he has other skills, whatever those may be, which could also be valuable to the colony. So with Charlie, we start with three specializations, and can train him as high as five.


That's the thesis of Jane Jacob's last book: Dark Age Ahead. Not everything is written down, and once there's a generation raised that doesn't use the knowledge it gets forgotten and must be rediscovered.


followup q to original q;
what advances would go furthest to reducing our modern societal footprint?


Hi Chris,

Yup. My favorite quote on making one (from a Santa Cruz Island chief) is: "if you want to build a canoe, first plant a garden." The point is to make sure everyone's going to get fed while they're working on the boat.

That, for those allergic to management, is why you need skilled administrators.

As for the sennit rope, I haven't seen that number (1-2 m/person/day), but I'm skeptical. For one thing, Hawaiians used to sell meter-wide balls of sennit rope to passing sailing ships (that's a lot of labor, if you think about it), and for another, sailing rope is a multi-person task, just as it was for our ancestors.

It might work out to 1-2 m per day if you count the total time to go from coconut husk to finished multiply thick rope, but I'm doubtful. Anyway, it's slightly irrelevant, as apparently making rope is a common pastime in some places--sit in the canoe shed with the boys, talk, and make rope.



I'm with Charlie, and I know a lot of pharmacists too. It's a vastly under-rated profession. So are accountants. Bureaucrats and lawyers even. All you have to do is suffer at the hands of someone who's incompetent in one of these professions to truly value their expertise.

The reason your idea (read a lot, know enough) doesn't work is amply demonstrated by the massive responsibility we give people trained this way. They're called recent college graduates. At a slightly earlier stage in their development, they're called sophomores: wise morons. Most companies trust them to get the coffee and do what they're told.

There is an important grain of truth in there: 80% of what most people do can be handled by general education. It's that other 20% where the hundreds to thousands of hours experience is vital. This is why anyone dealing with emergencies (or anyone with a competent manager) routinely does training and/or practice, to maintain their skills for dealing with those unlikely events. Just remember: most of a neurosurgeon's time is spent on the computer, not operating on your brain.

Or we can take the example of the space shuttle. A routine launch can be successfully simulated on a smartphone and handled by someone drunk. You push the buttons, and the shuttle takes off. The other two years of training that astronauts get is so that they can correct all the myriad ways a routine launch can get screwed up, because in the normal failure mode for the shuttle is catastrophic breakup.


"Or we can take the example of the space shuttle. A routine launch can be successfully simulated on a smartphone and handled by someone drunk. You push the buttons, and the shuttle takes off. The other two years of training that astronauts get is so that they can correct all the myriad ways a routine launch can get screwed up, because in the normal failure mode for the shuttle is catastrophic breakup."

This also flies in the face of experience. Highly skilled jobs do get obsoleted. Civil aircraft no longer have a flight engineer or radio operator on board, functions that are now done by pilots "pushing buttons" and using computers.

Once upon a time you needed to hire a driver for your automobile too, but even 16 year olds can drive themselves now. And let's not forget that we are all using computers to communicate on this blog, yet when I was young, computers were run by experts behind closed doors. Anyone can probably put together a long list of skills that were once the preserve of designated experts, but are now either obsolete or handled to some level of expertise by the general public.

Getting back to space travel, I fully expect average individuals to be able to "pilot" a spacecraft in interplanetary space, even do landings and takeoffs on some solid bodies too.


Johnson & Johnson has again had to admit they've screwed up. Not prescription meds, but they're putting the wrong meds in boxes and putting some ineffectual meds in boxes.


Eventually an AI will anticipate technological advancement by analysing the human psyche. If it can monitor a large sample group, get an impression of, not what people think they’d prefer, but what as yet unimagined device will be beneficial. For example, who’d have thought the i-pad (in a marketplace saturated with much cheaper laptops, netbooks etc) would be so popular?

Essentially, once the singularity happens, the human creative process will be superseded; the manufacturing side completely automated would be a given. Extracting and processing the base materials – again, no need for any humans.

So a life of leisure or a nightmare world run by machines?

Well, somehow this may not happen, mostly because we humans will always believe we can do it better ... think it better!


Eventually an AI will anticipate technological advancement by analysing the human psyche. ... Essentially, once the singularity happens, the human creative process will be superseded; the manufacturing side completely automated would be a given. Extracting and processing the base materials – again, no need for any humans.

Or, to put it another way: your argument boils down to "if we had a magic wand, we could magic up solutions to all our problems!"

And monkeys will fly out of my arse.


It seems to me that you start out conflating two issues: a desire for simplicity and a desire for "small government." It seems, rather, that the more complex a society is, the less effective a large, bureaucratic, centralized government is going to be in dealing with those complexities. Smaller government, or, more precisely, decentralization, is precisely the right response to increasing complexity.

Now, it may indeed be that most American conservatives mistakenly support small government out of a desire to make the world less complex, a nostalgia for simplicity. In that case, it's just another of their delusions, like creationism. No one ever accused modern American neo-conservatism of coherence.

As to the central question: What size population is necessary to sustain current levels of technology?

That is a separate question from the optimum size of a sovereign state (geography and/or population) and the size/scope of its government, as long as there is trade. With trade between states, you take advantage of decentralization while also allowing ideas/human capital to flow.

Someone else has mentioned Matt Ridley's new book, which deals with just these questions.


Here's a hint, Franklin; I'm not American, and I'm not conservative.

Also? You're mistaking a serious technical question asked by a science fiction writer for an expression of desire for one particular solution to the problem space of autarky in contemporary politics. Here's another hint: my interests probably lie somewhere other than where you might expect.


Say hypothetically that the answer to your question is 800 million. And that the answer to the same question posed in 1900 was 50 million. Does this number always increase with more advanced technology?

At what point in time will we no longer be able to advance technically because we've reached a point where we need every living human to maintain our technology?

Maybe before we reach that point, something happens to reverse that trend. But what?


Does this number always increase with more advanced technology?

That is a very good question.

It's possible that it doesn't. Remember, today only about 0.5% of our population are needed for agricultural work, about 2-5% for industrial production, and probably the same -- or fewer -- for resource extraction. But if you go back to 1945, the UK had around 1M men working down coal mines (3% of the population) just to supply coal -- around 30-45% working in industrial production/wartime material supplies, and a good chunk (10%?) tied up in agriculture. Some low tech solutions are very labour intensive -- mediaeval strip cultivation farming, for example.


Do we get to assume one nation, one culture, and one language? Developed-nation levels of education and infrastructure already in place? Europe-level crime, no need for a military?


Say you've arrived on the pristine surface of Earth 2, which looks a lot like Earth of 5000 years ago save there are no other humans on it, and you want to build an incandescent light bulb. A bulb really only has a few components, and no software, semiconductors, or moving parts. But you need at least silica, soda, lime, steel, copper, tungsten, and noble gases as raw materials. A "raw material" like copper actually requires many steps between the part where you find its ore in the ground and the part where it's ready to be drawn into wire. Tungsten is a relatively rare metal, so either you have to prospect for good ore or expend a lot more time and energy extracting it from a low grade source like granite. You can simplify a bit if you substitute carbon fiber in vacuum for tungsten in noble gas to make an Edisonian bulb. But this bulb will be much less efficient at producing light than a modern design and won't last as long either, plus you still need steel, glass, and copper. There's a long industrial tail before you can even make something as simple as an incandescent light bulb. How much capital, as well as how many people, do you need to take to Earth 2 before you can reproduce industrial civilization?

A common SFnal magic wand to skirt these problems and get on with the adventure is a universal maker machine, a Star Trek replicator or the slightly more respectable nanofactory. A tree is really complex and comes in a tiny self-starting factory, ergo any artifact up to the complexity of a tree can be reproduced in analogous fashion at some point in the nanofuture.

I'm not sure I buy the simplicity of that reasoning and I don't recall any SF that got deep enough into the wiring diagrams to address it to my satisfaction. One problem that comes to mind is that atomically precise manipulation is really slow, because the overseer machine has to account for every atom. You can see an analogous problem with electron beam fabrication of integrated circuits: the machine is much smaller than a full blown chip fab, and it can produce circuits at least as advanced as those of a cutting edge fab, but it would take years to pattern a single 300 mm wafer. There's a tradeoff between capability, complexity, and throughput.

Biology answers this problem by relying on tricks of chemistry and spontaneous self assembly, to take advantage of basic physical processes so that not every part of a system needs active regulation. The tradeoff is that biology is also constrained in what it can manufacture. For example, something like 20% of human-designed pharmaceuticals contain at least one fluorine atom, because substituting fluorine for hydrogen is a great way to tweak polarity and bulk for desirable properties. On the other hand, nature has produced millions of natural chemicals, tens of thousands of which have been studied in detail by humans, and the number of fluorine-containing compounds can be counted on your digits. It's not that fluorine is very rare or a poor addition to the chemical tool chest, but it's difficult to incorporate in vivo given the constraints of biology. Just as earthly biology can't be easily engineered or evolved to produce many fluorinated compounds, so it is not going to produce metallic aluminum, chlorine gas, doped silicon, or any of a thousand other things vital to modern industrial civilization yet incompatible with biosynthesis.

So, back to nanfabrication: if anything like a universal maker machine can be invented at all, I think it's going to a low-throughput machine that is dependent on extremely high purity feedstock chemicals akin to those used in semiconductor fabrication. If there's a role for such machines in the reproduction of industrial civilization, I think it's for the universal atom fabricator to use its precious pure feedstocks making slightly more specialized machines with higher throughput, that in turn make slightly more specialized machines with higher throughput, until at some point you're got lathes, crushers, welders, bulldozers, and all that other macroscopic heavy industry right out of 1950, plus all the newer stuff that's followed. At that point you've completely reproduced industrial civilization and can make more fabricator machines and their feedstock chemicals and place them in another set of deep mineshafts awaiting the next dark age. Even if the end result can be viewed as one universal fabricator making a copy of itself, the life cycle is a lot more expansive. And it suggests that humans need to be on hand to run a lot of the intermediate stages, where you can't go straight from blueprint to finished product, *or* that there needs to be an awful lot of up-front work to formalize every step of the process into machine-intelligible tasks. This is more like the older idea of a von Neumann machine than the straight software-to-substance vision of SF nanofabrication.


Any intelligent person with the sort of general education that comes from learning to read (two months education) plus two or three thousand hours of reading a variety of materials, plus a few hundred hours of apprenticeships can do the work of any of dozens of specialties 80% as well as the average specialist today.

Mr Dunning and Kruger! Paging Dunny and Kruger, call for you on the white courtesy phone!

I am getting seriously tired of reading tripe this, frequently from engineers who ought to bloody well know better.

As I said elsewhere, I don't care that my Doctor knows how to remove my appendix. I want a General Surgeon with dozens of appendectomies under their belt wielding the knife - or if it's a trainee, I want a GS surgeon who has experience on hand.

I don't want a Cardio Thorastic specialist having a go at a spot of brain surgery, just to see how it feels.

I don't want a Micro-Light pilot with a few hundred hours having a go behind the stick of an A380.

If I'm commissioning a massively complicated and expensive chemical plant, I want "Bob", the 50+ year old guy I used to work with who'd done dozens and knew the process inside out over, me, the graduate, with a few hundred hours of experience.

The list is endless. There's a reason why we have to specialize.

And Heinlein wrote a lot of that stuff when it actually was possible to be a lot more generalist than it is now.


1) Yes, people can learn new skills. At 41 I'm trying to learn Golf, I'm paying somebody who actually knows how to play a sum of money to humiliate me for a few hours a week. So far, it's not good and, in all probability won't be. I'm a great cook, but it's a hobby. I'd be deluding myself if I thought that my ability to cook well translated into an ability to cook at a world-class level, not without dropping everything and just focusing on cooking. Even now, I reckon I screw up bread as often as I get it right. It's all about the efficiency and the risk. As others have said, a pharmacist isn't just about dispensing drugs. Likewise there's a damn good reason why you can't buy certain types of drugs to treat yourself after you've self diagnosed on the internet. Basically, until you actually know, what you don't know, you're dangerous. I just want to you be dangerous in ways that can't have externalities which affect me.

2) Based on my experience in the engineering world, knowledge isn't stored anywhere near effectively enough in any form other than the experience of people who've done a lot of it.

I can read a book on Workshop Engineering Theory, so I know how to run a lathe, milling machine, shaper and so on - but, believe, me, it's easier to have somebody with a few decades experience keeping an eye on you for your first hands on. After that, it's practice, practice, practice...


I mean, if we're allowed to reorganize for maximum efficiency, like moving everyone into big cities (except for resource extraction zones), mandating maximum levels of automation in industry, abolishing social security, retirement spending, the 40-hour work week, labor unions, and holidays, eliminating military spending and most spending on the penal system, forcing everyone into a mostly vegetarian diet, and eliminating diversity in consumer choice, we'll end up with a much lower number.

If you want society to resemble the one we have now rather than some sort of nightmare industrial state, then your number is probably reasonable. However, there are plenty of in-between states.


Drexler's ACM-award winning book Nanosystems lays out a potential nanofabricator design in detail using very conservative estimates and proposes many other potential development paths to achieving equivalent functionality. The feedstock problem is dealt with at length, but basically one can design something that takes raw materials apart and sorts and modifies the pieces. The main problem isn't theoretical feasibility, but getting the first system actually built and programmed in a way that is useful for desirable ends without being just as useful for destructive ends.

The other problem is that introducing nanotech into a story makes solving problems too easy, thus making writing interesting fiction difficult. It's like giving all your characters magic wands. Some very interesting stuff has been written despite the difficulties, much of it by our gracious host, but the sub-genre has been almost as thoroughly mined out as that of time travel.

In the case at hand, one could just digitize the whole population and all their artifacts, recreate them wherever there is enough of all the needed elements, and poof! Not a satisfying solution. But not totally implausible, especially compared to such SF standbys as FTL drives.


We have a civilization?


Curious, though that Israel was set up with only a few thousand people. OK the cooperation is going to be a lot more intensive than your regular society, but as long as people can handle being the same for a few decades, no privileges for any occupation, it's going to be fine.

Once upon a time, the warrior caste was the most privileged. Only a a decade ago, the "scientist" caste was seemingly privileged. They all had us going there, didn't they. "They" pretend to be for wars to be "religious" when they never are. "They" pretend for wars to be about ideology when they never are. Sound a little paranoid? Well, when was the last time you weren't afraid of muslim extremists or christian fundamentalist or child pornographers or smokers you can remember?

Have you ever thought about the danger to a civil society being under surveillance all the time. Ever thought about that maybe, with all the scares that are inflicted on us, we are deconstructing the very preconditions that we thought our society is funded upon?

If all it takes for us to "trade liberty for some temporary security, deserves neither liberty nor security", we have already lost.


I've been thinking about this all day, and I think what we're looking at, if we assume colonization rather than the degradation of society, is a twofold problem which goes something like this:

1.) What is the critical path, using this particular biosphere/geology of an earth-like planet, to building a civilization that can take care of it's people as well (at minimum) as late-twentieth century Europe/America within a century or two. Most importantly, what can we leave behind that a society really doesn't need, thus reducing complexity and the critical numbers necessary to build a society? (Vernor Vinge hinted at this in "A Deepness In The Sky.")

2.) The necessary development of a go-to-hell plan. That is, if we run into problems, how do we degrade gracefully to a dynamic 18th century society rather than dying out or experiencing a dark age?

I'm going to handwave a bit, just to keep the focus on Charlie's original question, so I'm assuming a working FTL drive that's not too expensive, an earth-like planet without intelligent life, and biotech that's as good as our current computer tech. (In other words, we can program genes without getting much more complicated that writing assembly language, even if the ASM-equivalent is merely the front-end to some very expensive equipment that actually shapes the genes.)

Imagine that we discover an earthlike planet and dispatch a Stage-One survery. Stage-One discovers that we can breathe the air without major modifications to the human genome, and that there is a minimum of useful building materials, including something like wood, something like clay, something that burns like wood (though it doesn't need to be the same material) and some workable metals of at least the utility of bronze. In other words, we've got what we need for the final stage of our go-to-hell plan, which is a bronze-age society with science, education and medicine based on current levels of enlightenment - sexual equality, good herbal medicine, cleanliness (including sterization where necessary) universal literacy, compound bows, glass, etc. If we can't get to the bronze-age easily, we don't go any further - the planet is marginal and we stay away except for scientific expeditions.

Our Stage-Two survey is essentially a practice colony. The stage-two survey will stay on the ground for at least two years, basing themselves on a large-ish remote island no bigger than New Zealand. This island will, with luck, become the first colony site on the planet, They'll do some minimal terraforming, make sure that earthlike crops will grow in the soil and produce good seeds, and make the necessary genetic adjustments to the earthly biota. In addition, they'll make sure that basic industrial metals, such as iron and copper are available in some kind of quality. What we're looking for at this point is a second-to-last fallback point - something like the technology of the 1870s, overlaid, of course, with a more modern enlightenment.

The Stage-Two survey will also carry the necessary geological (or deep-space) technology to ensure the presence of more complex industrial metals - platinum, palladium, gallium arsenide, lithium, etc on-planet and nearby. If necessary the Stage-Two survey will find an asteroid and impact it on the surface, in appropriate-sized chunks that don't cause too much damage. (If we have really good spacedrives, maybe we'll just land the thing, but a good asteroid impact might be just the thing to prepare the local biosphere for terraforming.) This ensures another possible stop-point if we have to move downwards - a 1940s style tech.

However, the main purpose of the Stage-Two survey will be to design a critical path that gets us from "landing a thousand colonists" to "building a complex society." (Later there will be a Stage-Three survey that will be concerned with final genetic manipulation of the human population, but that's a little off-topic.)

Our development of a critical path starts with what we want; a late-stage technological society that can sustain itself with the smallest possible number of people. We begin by eliminating dead-end paths from our technological tree. Since we want sustainability, AC power isn't necessary - every home (or every small community as the very biggest generating unit) will have solar, wind, hydro, or nuclear generated DC electricity. (Pebble-bed reactors maybe?) Getting rid of AC power eliminates gigantic portions of the tech and training tree, and it enforces local sustainability.

Next we get rid of petroleum power. It pollutes, it cause respiratory problems (which will probably be bad enough on a different planet) and we don't need one more set of chemicals in the mix when we're trying to terraform. Lastly, if we're in for the long haul, we don't want to waste an excellent source of useful chemicals by burning it. (Plus the planet may not have dead dinosaurs underground.)

We eliminate meat from our colonists' diets, at least at the outset - we're not going to be mono-cropping, so we know some source of food will survive, and we can use preserves, canning, drying, nut storage (possibly in the form of nut-butters) and other forms of technology to preserve protein and vitamins against lean times. Once again, meat is a gigantic industry, it has a huge footprint, and it's a source of diseases - an alien virus might adapt to eat terran life by breeding in a meat animal. Further, this helps keep our medical-knowledge tree simple by eliminating entire branches of cardiological medicine, (no heart attacks, anyone?) and we may not need veterinarians, though we may raise sheep for the wool.

We eliminate whole branches of medicine by sending out colonists who don't suffer from diseases caused by bacteria, parasites, or viruses.

We don't need cable TV or broadcast television (we do have fast, powerful networks, right?) and we don't need any form of radio that doesn't use a networking protocol, so forget FM, AM, Ham radio, etc., unless we have problems and need to degrade our tech. We won't have a wood-pulp paper industry. (We want our books and clothes to last, so we'll use hemp, plus it has medical utility.) We don't need most forms of fabric - cotton and wool will do just fine, thanks, so silk, rayon, nylon, etc., hopefully won't be necessary. (We may want some kind of artificial fabric just in case something proves able to eat cotton or wool.)

Since our entire communication system will be packet-based, we won't need POTS technology, though we may keep it around for the sake of "degrading gracefully" if we run into problems.

The stage-two survey develops building standards and weather-appropriate housing technology, so we can standardize on one kind of building/housing (brick houses with sharply peaked roofs for someplace that snows, for example,) so we have the smallest possible tech base for necessary for construction - nobody needs to learn to cut stone or build big glass-and-metal boxes...

In other words, we can leave out enormous portions of our tech tree without ill effect.


Curious, though that Israel was set up with only a few thousand people.

Plus a lot of equipment fueled and maintained with spares coming from outside Israel. Even after 1948, Israel didn't get to where it is now by purely local development.

181: I'm assuming a working FTL drive...

Our gracious host puts that in the magic wand category, so no point even predicating a reply based on that assumption.

When you pull together a number of Charlie's threads - this one, energy for star flight, viable long term social institutions, likelihood of finding habitable worlds, what you get is a sort of Drake's Equation for successful colonization efforts, and the numbers are generally unworkable for technology within the discussion bounds.

This one seems to assume a working population of 100M.
I suspect that if we colonize with the sort of proposed technology available, that the colony is going to be some sort of lower technology colony with some readily achievable high tech and a growth path to a more technologically civilization using a huge library to help bootstrap it.

However I also think that discussions like these are like any earlier age's ideas on what constitutes "magic wand" technology. By the time we can even seriously contemplate actual stellar flight, we will have discovered a number of science and technology game changers that will make these discussions seem quite quaint.


You lead in with a distaste for all things republican and conservative and then abruptly shift the discussion to a question of how many people it takes to maintain today's society at it's present levels...Perhaps the next entry in the Stross diaries could go a little deeper into how the two are related. It is quite a leap of faith to go from republicans/conservatives wanting a smaller government to how many people it would take to maintain today's society.

The number of people required to maintain something once it has been created is far, far less than the number of people that designed/created it in the first place.

Unfortunately, current trends in society have gone from building things that last to things that have a far shorter lifespan and are designed to be replaced rather than repaired. Sorry state of affairs if you ask me...leads to far too much reinventing the wheel at every turn and far too many parts. As well as leading people to research some really stupid nonsense. If you want actual cases, take the research student who stated recently it's ok for complex electronic components such as cpu's to be faulty because they consume less power...(I read this and wanted to smack the stupid out of him, as well as his academic professor adviser, and the person in the US government that approved the grant money.)

However, the need for archivists and people that are actually good with their hands/tools would become the most valued in the static society you are contemplating. Someone needs to maintain the documents the sustainers use to maintain all the equipment used by society.

I just started a new job working at a university nuclear research reactor and as the senior electronics technician it is my job to maintain electronics spanning 50 years of technology. Ever see a voltmeter that uses paper and prints out the value being measured? Or one that has rotary dials to display the value being measured? The problem is getting replacement parts for older equipment that has become technologically obsolete.

Charles static society solves that problem since nothing new needs to be created, I won't have to pay $97 for four electrolytic capacitors in a tin can filled with a paraffin wax since nothing will become obsolete! LOL!


>So why is Microsoft paying all those people to maintain Windows? Don't you think they might do that if it was possible and divert the rest of technical staff to developing new features and products?

Hmmm, because Microsoft is the worst thing that ever happened to the computer industry and software in general? Perhaps because the code is so buggy one could liken it to a toxic cleanup where in order to start over from scratch they would have to scrape off the earth to a depth of one to two feet just to get rid of any long lasting effects? Of course, this would disrupt Microsoft's revenue stream in the meantime and you wouldn't receive your daily patch as they uncover yet another bug/security flaw?

Corporations used XP for years, ignoring Microsoft's steady stream of OS upgrades, and are only just now having to pay yet again as "Support for Windows XP with Service Pack 2 (SP2) ended on July 13, 2010." You did notice an uptick in Microsoft's earnings this past quarter in corporate OS sales? Now you know why, and why many consider Microsoft a monopoly.

As for new features, tell Microsoft I want virtual windows, something I have had for over 18 years on every non-Microsoft OS system I ever used. See if you don't have a clue as to what I am talking about. You'll be glad you did.


Another aspect to consider is internal stress within a society. This is driven on the tribal imperative to know where you stand within a tribe. As tribes evolved into coutries you got tribal chiefs and kings then as the number within society grew you got the three estates of the feudal model and then the move to having craftsmen and yeomenry rising in status as society grew again. This is because each human wants to know how he ranks. Therefore every time society grows the number of social roles has to increase both sideways and in distance from top to bottom. (number in each rank and number of ranks). Where an older system fils to evolve you get social stresses which in part contributed to such events as the French Revolution. One of the reasons capitalism has been so successful (another argument here I'm sure) is that it provides a second way of measuring the individual against those within the society. The larger society is the more technology can be maintained (the arguments for this have already been established). But the larger society grows the greater the distance between those at the bottom of society and the top. This in turn increases the amount of it's energy the society must devote to it's own cohesion. Part of this is the bread and circuses style social model that proved so ruinous to the roman empire and this internal energy expenditure will weaken a society to outside forces.

Therefore when social modelling in this manner you should consider the energy required to maintain it and the inherent need for a complex series of social standings.


I've read Drexler's thesis, which I understand is very close to the Nanosystems book, and it's quite enough to convince me that a lot of molecular manufacturing is possible in principle, but that's far from having a working nanofactory on my desk. In a similar vein, I believe large scale quantum computers are possible in principle but I'm far from ready to abandon public key cryptography.

Drexler's vision of molecular manufacturing, as presented in his thesis at least, remains well ahead of the current state of the art but far short of fairly common SFnal depictions of molecular manufacturing. For example, he confines his detailed examples to "diamondoid" structures mostly composed of carbon with other light nonmetallic elements as minor components. That's enough material diversity to build a lot of things, but it also leaves out a lot, like (say) any sort of battery. He also estimates (page 19 of thesis) that the scale of typical products from molecular manufacturing will be on the order of 1 micron -- so you're not going to get a product too large to accidentally inhale out of such a factory. Finally, on his web site FAQ he explains that he does not envision molecular factories completely reproducing themselves in a closed loop, but remaining dependent on the rest of industrial civilization.

As far as nanotech being a magic wand that makes story writing difficult because it leaves the characters with nothing to do*, I say nonsense. Just because orphans are no longer at acute risk of starving to death in England doesn't mean authors ran out of compelling stories after the time of Charles Dickens. If nobody in the world is deprived of manufactured goods, that still leaves an infinite number of potential conflicts.

*The analogous real life complaint I've heard is that a society where machines do all necessary work, even if achievable, would be disastrous because it would leave people without purpose. I've got to wonder if these naysayers have the world's most interesting jobs or just severely overestimate other people's job satisfaction. Or maybe it's code for "but then rabble would roam public places at all hours."



This is down, dirty and cheating. It also involves (ahem) reading Baen's most right wing author: but Kratman's carnifex series does think about how to make things work with a very small transfer window.

But Kratman is (a) a lawyer (b) ex US military (c) enamoured with Panama (d) somewhat to the right of Heinlen and Genghis Khan. His colonists start with muskets and are at c20 by the end of 400 years.

Think of it as research, since you are from the true atheist kirk of Scottish socialism :-).


@ # 182: "The number of people required to maintain something once it has been created is far, far less than the number of people that designed/created it in the first place."
The mobile, load-hauling steam-locomotive was developed by about 100 people, total, between 1802 (Trevithick) to 1825-30 (the Stephensons) ...
by 1850 there were well over 5000 miles of rail-track in the UK, employing many tens of thousands of people ....

"Simple enough"/good enough - and yet up to reasonable modern stabdards.
The auto industry is a prefect example of how NOT to do it. My car was built in 1996, to an u pdated version of a 1990 design. I can do about 90% of repair and maintenence myself (because it is a Land-Rover). But it can go almost anywhere, travel at the legal road-limit (70mph in the UK) and carry lots.
IF you are prepared to limit the number od=f designs of "stuff" you have, and build-in redundancy, ruggedness and simplicity, you CAN have certainly a very-late-20thC society, PROVIDED you follow those design engineering principles, and make sure the documentation and backups are in place. Note that you STILL NEED specialists to deal with the important 10% of repair/maintenance you can't do yourself.

Education: This is the killer problem.
The UK's population is badly educated, and a huge amount of even the good education is wasted, and I suspect the same applies elsewhere, too, though not as badly as here.
I have an engineering MsC, gained at age 48 - I have NEVER been able to use it professionally, because of the prejudices and blinkered (lack of) vision of UK employers - I'm "too old".
This sort of insanity has to be killed if we are to survive with a more limited population.

The other thing we can do without, of course, is a "holy cause" of any sort, because the moment one of those rears its ugly little head, rationality and forward planning go right out of the window .....


5. How many people can you stuff into a kilo of computronium? Wait till you can get more.

1-4. Are you getting at the small government/big society drive here, Charlie? (Why are "drive" and "drivel" so invitingly close?) As others have mentioned, the size of the government and the complexity of society are not linked in a straightforward way. Our present leaders see an inverse relationship, where the government gets out of the way and the people set up a rich and complex web of services.

Almost all government interventions are framed as some form of simplification, either positive (yes, everyone will have decent healthcare as a right - don't worry about it any more) or negative (no, you can't use guns to defend yourself - we'll provide an army and police for that).

Are you saying that it is foolish for a government to accept some of the increasing complexity in society by relaxing some of these simplifying constraints?


Are you getting at the small government/big society drive here, Charlie?

No; I'm getting into a new answer to the Fermi paradox. Or maybe brainstorming aloud for background to a new space opera 2-4 years down the line. Or contemplating books 7 and 8 of the Merchant Princes series, or 8 and 9 of the Laundry files.

The question is interesting in and of itself -- unlike the annoying ideologically loaded sound-bites in the background.


I'm just wondering, it is probably not necessary to maintain *all* knowledge in a working form *all* the time to its full extent.

Certain technologies may fall into relative disuse, from the construction point of view, but continue to be maintained and made use of.

So, some technologies may need differing amounts of people over time, going through a slumber for decades to be reawakened for a couple of years.

How often do we construct shipping channels these days? Or when was the last time that anyone built a bridge, planning to build houses on it? (Well, some are still being used.)

At some point we will have to replace some of the largest hydro plants ever built, but not any time too soon. But when it happens, there will be a lot of people getting training in this area. And when it's over, it might even be over for centuries, if we really put our minds to it and stopped believing into using a mere 15m of concrete were 10m might suffice and used 25m instead.

Or what about nuclear plants? Very few were built for decades, yet we *could* start building a dozen a year or more, if we thought it was necessary and allow for a run-up of a decade or so.

Also, not even Iceland maintains a squad of lava-fighters, even though they proved their usefulness in the Eldfell eruption of 1973. They simply came into existence when they were needed.


As for those who wonder why we have short-life gizmos, redundant OS, rampant bureaucracy, and any of their particular pet peeves, it is our societies naturally evolved solution to having less jobs than people. That also enables the whole "thinking out of the box" and lateral leaps that have advanced technology the last decades without war pushing it forward.

Otherwise we would be back in the late 19th and early 20th century (and UK and the USA were the kind of privileged places where colonization meant there wasn't much of a job imbalance, so they suffered less than many others, which is also why they weathered less extreme policies).

What that means for our technology level is that you could probably sustain it with less people than those 100 millions (I particularly like the Japan thought experiment), but those people probably would go down further in tech, to the 70s, maybe, because they would live better, and enjoy life more, at that level. And unless you force them, people will prefer an easier life to a hard one. Because that single OS bugchecker's life would be true hell.


I hate arriving late to this sort of thread, especially when it is related to subjects I am interested in. For example, how much in the way of paper pushers, call centre staff, shelf stackers etc do you need in your hypothetical high tech society with all mod cons? Can you get rid of some of them and reduce the numbers by a few million more?

Yet for all that we have more computers, ways of doing things using fewer people and more capital (eg bulldozers not shovels) are we still employing the same percentage of the population we were 30,60, 80 years ago because the system has an increased throughput of energy andraw materials, or because you really do need that many people to have a modern civilisation? Judging by the results of the recession, I say the former - we are keepingmany people employed simply because the system has a high throughput of energy and resources, tearing down and building up, whereas if we just build 300 year lifespan houses, landrovers and so on, bang, thats probably at least 10% of jobs gone just like that.

As soon as you switch society over to maintenance rather than infinite growth, everything changes, including the number of people required.


I haven't read the comments, but I think there's a fundamental flaw here - the thing is, you don't need people to know and do all of this stuff *simultaneously*. No one needs to be designing jet engines all the time. You just need someone competent enough in some core engineering concepts that, if ever you need that someone, you can put him through a crash course and teach him jet engine design in time to fix whatever pressing problem you have. For a computer analogy, you are mixing up hard drive space (and I can buy that it's approximately 100 million people-worth) with RAM/CPU cache, which I'd posit is substantially smaller. Of course there's the frictional cost of retraining constantly, but few problems are urgent enough not to leave enough time to look up some stuff. (And if there's a Matrix-style teaching machine, even better.)

So rather than 100 million, I think you can make do with about the population of an average university, as long as you have The Internet to back it up.


The point that most knowledge is found in books applies not only to high tech but all the way down to furniture and tying knots (in case you think I'm exaggerating about the latter I've just been talking to a retired Scout leader who -if reminded-is agast that the current generation of'trained' Scout leaders two of whom have the wood badge (and are supposed therefore to both possess and be able to teach these skills)! haven't the faintest idea how to make a secure square lashing let alone create a rope bridge to span a river.

We seem to be losing intermediate steps in our technology/survival skill set. So long as civilisation doesn't suffer major trauma who cares if the ability to throw a footbridge across a river using just a few logs and some rope is lost. but without that functioning pervasive society that river becomes a major barrier.

What other unregarded skills would we lose in a slimmed down civilisation/colony?


I haven't read the comments, but I think there's a fundamental flaw here

You should have read the comments, because you're about 48 hours late to this particular idea (which has already been chewed over and rejected).

No one needs to be designing jet engines all the time. You just need someone competent enough in some core engineering concepts that, if ever you need that someone, you can put him through a crash course and teach him jet engine design in time to fix whatever pressing problem you have.

Two problems with this: firstly, you may find that it takes a long time to ramp up various tech speciality. Not all knowledge is written down, and learning how to do something is very different from learning about something through reading about it. In some specialities the learning curve even with skilled tutors to hand is a decade or more -- and you may not have a decade to train up the extra neurosurgeons to tackle that cluster of brain tumours.

And the second problem is that the cost of re-starting some specialities from cold may be much higher than the cumulative cost over time of maintaining a skeleton capability in it, because it requires supporting infrastructure that may be non-obvious.

Put it another way: the UK is probably going to start installing new nuclear reactors in the next decade. They'll be French-built. Why? Because it's 20 years since we last built any of our own, and for most of that time we were training only 4-6 civilian- nuclear engineers a year -- barely enough to cover retirement losses from the maintenance staff of the existing plants. If you want to order up a batch of new nuclear engineers, the lead time is a decade (from high school up through postgraduate training) and the AGRs and first-generation PWRs are due to begin shutting down in eight years' time.


"No; I'm getting into a new answer to the Fermi paradox. Or maybe brainstorming aloud for background to a new space opera 2-4 years down the line. Or contemplating books 7 and 8 of the Merchant Princes series, or 8 and 9 of the Laundry files.

The question is interesting in and of itself -- unlike the annoying ideologically loaded sound-bites in the background."

@Charlie: One additional parameter: legacy vs. rebuild. I suspect you can get two different answers depending on how you get there.

I you're contemplating a legacy system, say, collapse from first-world industrialism into third world quasi-sustainability (e.g. replace all the mechanical recycling separators with legions of direly poor people living off what they scavenge at the dump), then you may need more people than if you're building society from the bottom up. This second case would be where you ship off a starter colony with a huge knowledge base, and they build their system from scratch.


Greg, go a bit further with your analogy of the steam locomotive you presented.

>"was developed by about 100 people."

Granted, I see your point. Let me clarify and go a bit further. That steam locomotive consisted of individual components, not to mention the tracks they ran on. So you have an entire secondary workforce past the development stage to accomplish the actual manufacturing of the components. (tool and die, steel manufacture, molds) Once this secondary work was completed, the steam locomotive could theoretically be produced indefinitely as long as the raw materials kept pouring in...

So what we are left with is:
1) A small development team
2) A much larger manufacturing base for its components
3) A workforce to accomplish the construction of said locomotive and lay the tracks
When all this is said and done we are left with:
4) The maintenance staff (someone has to grease and lubricate, and repair the tracks) I'd say depending on how good the initial design was, the less people required for this particular function.
5) The operations staff (running the locomotive) and selling the service (ticket offices, etc.) Still I believe the numbers for 4 and 5 are fewer than those required for 1, 2, and 3. Perhaps we merely differ in where we draw the dividing line.

Recent correspondence that illustrates our current mentality of consumerism (enjoy!)

Dear Sir,
I am looking for the following part…if you have an American supplier I could go through, all the better. I am looking for the following relay to replace a bad relay found on a circuit board that I am repairing. If you do not have an American supplier, could you give me a price for this component and how I could go about ordering one.

Thank you for your Email.
But it is unfortunate that we have no distributor in USA, it is very expensive to send you a relay. If possible for you to buy an alternative in your local electronics market?

You'd think I was working on something important, but the reality is I'm merely trying to fix a Staples MailMate shredder designed from the viewpoint that if it breaks, one should throw it away and purchase a new one. Manufactured in China, sold globally, and not readily repairable...(User reviews mainly are negative, so you can add my name to that list!)

As for your comments about education:
I agree with you 100%. The United States apparently is exactly the same as the UK. I too now fall under that category of "I'm too old" despite having worked as an electronics engineer for 18 years with a BSEE.
After 500+ resumes and two layoffs in the last two years, I consider myself lucky to have even found the job I just started. However, I am now forced to be that "10% specialist" due to lack of funds which means I improvise a lot just in order to keep ancient technology working.


"You should have read the comments, because you're about 48 hours late to this particular idea (which has already been chewed over and rejected)."

Yeah, sorry. But there comes a stage when the scrollbar at the edge of your screen becomes narrow enough that tldr syndrome kicks in.

"Two problems with this: firstly, you may find that it takes a long time to ramp up various tech speciality. Not all knowledge is written down, and learning how to do something is very different from learning about something through reading about it. In some specialities the learning curve even with skilled tutors to hand is a decade or more -- and you may not have a decade to train up the extra neurosurgeons to tackle that cluster of brain tumours."

But do you need neurosurgeons at all? Arguably not - though brain tumours are distressing and terrible for people with it, they affect a tiny proportion of the population.

I buy the gist of the argument - I don't doubt that you need some skilled people on hand to do some things. But again, we need a detailed idea of what the brainpower bottlenecks are, and how many of them there are. As a rough estimate, we can assume secondary level education + some cross disciplinary (mostly maths? Programming?) can plausibly be taught to everyone. In our society, getting people up to graduate level then takes maybe 2 years, if you rush it and teach very intensively. So our question becomes - what do we need people to do for which either a graduate level education + ample documentation is insufficient, or for which 2 years of training is too long to wait. (Of course, postgrad is sometimes required as well, but postgrad education at present is probably very inefficient as a teaching process.) This number isn't zero, but I don't think it's 100 million people-worth.

As for the reactors thing, surely the core problem is something different - the French are unwilling to share their documentation, unless we bring in the people as well. If the british government is legally allowed to steal their plans, constructing power plants that are replicas of existing ones, as well as operating manuals for technicians, then there would be no problem starting up a power plant from just documentation in a short period of time. It wouldn't be pretty, and probably accident probabilities will go up due to inexperience.

The demand that information be written down and collated in sufficient detail to be easily learnt from is a difficult request in itself. But I think it's easier than pre-educating and then shipping out 100 million people. (And hoping that none of them forget their training or die or anything...)

(Or maybe I'm posting my prejudices, since I am sort of in the education field.)


Thorne: as a point of note, railway maintenance is not cost-free. During regular running, it's apparently necessary to relay about 3-5% of the track network per year -- completely replacing it every 20-30 years. The rolling stock requires maintenance too; at the high end, locomotives can be kept running for a few million kilometres, but that's deceptive -- they can be doing up to 2000Km/day. Realistically, the entire network and rolling stock needs to be replaced every 30 years. And railways aren't a 100% mature technology, even today: deciding in, say, 1850 that we had everything we needed and downsizing the design teams would have been a disastrous mistake.


As a rough estimate, we can assume secondary level education + some cross disciplinary (mostly maths? Programming?) can plausibly be taught to everyone. In our society, getting people up to graduate level then takes maybe 2 years, if you rush it and teach very intensively.

You have a touching faith in "education".

Firstly, teaching isn't merely labour intensive -- it's skilled-labour intensive.

Secondly, many people can't hack it. You want maths or programming skills? Historically -- both in 1960 and in 2000 -- around 50% of entrants to university programming courses have flunked out or passed only by virtue of mastering the cut and paste keys; it turns out that the ascending abstractions of variable substitution, loop constructs, and pointers are fundamentally difficult for a very large subset of the population to master. (The proof is in all the emails cluttering up your spam folder with subject lines like "Dear %RECIPIENT%".) There may be something we're doing wrong that we can fix, and suddenly our education systems will turn into PhD production lines -- but I doubt it. Until very recently there was no evolutionary selection pressure for abstract reasoning; arguably, there still isn't.

Thirdly, we have a technical term for graduates: we call them "interns" and let them fetch the coffee. Actual competence in most disciplines takes around 5000-10,000 hours of hard work to achieve -- 3-5 years of nine-til-five grind.

As for the reactors thing, surely the core problem is something different - the French are unwilling to share their documentation, unless we bring in the people as well.

Wrong. We have reactor engineers and designs; it's just that they're all military -- we built about one nuclear sub every 12-18 months, and the reactors are made in the UK. The problem is, the skill set involved in building a military naval reactor is not directly transferable to building a (much larger, but also much safer-by-design) civilian reactor. The folks who did the latter haven't been in that business since the late 1980s; even the non-retired ones are very rusty indeed.

Knowledge isn't book-learning; very often it's encoded in practice, and if you don't use it you lose it. If you don't believe me, try buying some of the more intricate Belgian laces. They haven't been produced since the 1830s. A glut of cheap machine-woven lace drove the exotic hand-stitches out of the market, and by the time demand revived the lacemakers had all died of old age. And today nobody today knows how they were made.


But the stress we put on systems also scales with the population. If we had a lot fewer people, we wouldn't be running trains 2000km/day either. There's a baseline rate at which things wear down without being used, but that's (usually) lot slower.


Alas, lots of stuff wears out even if you don't use it.

Take railways: you don't need to run lots of trains on the line to have to deal with landslides caused by rain. Or rust. (Or human depradations: idiots stealing the copper wire you use for signaling because it's valuable.)

Take locomotives: rubber and plastic become brittle and porous when exposed to sunlight (UV radiation). Steel bodies corrode. If you don't run them, you've got to mothball them -- and de-mothballing them takes time.

Take medications: most pharmaceuticals are unstable organic chemicals that decay over time. (Ever opened a ten year old jar of aspirin? The smell of vinegar will knock you over. That's because aspirin -- acetyl salicylate -- decays to salicylic acid and acetic acid in the presence of moisture, even in the dark.) The shelf-life on your meds isn't there because of built-in obsolescence, it's there because they have a half-life.

Take lithium-ion batteries: ... no, on second thoughts, don't. An elderly liion cell is useless as just about anything other than a fire-starter (or recycling input for new batteries).

If you want long-lived systems you've got to design them for longevity. But in an era of rapid technological change, longevity means building in inefficiency. So we don't do that, much.

My apartment (where I'm typing this) is long-lived; it dates to 1823 or thereabouts, and will probably be standing in 2187 if it's looked after. It pre-dates the idea of rapid technological change. But most of our cultural artefacts don't.


"You have a touching faith in "education"."

Heh, yes. And students test this faith daily, but I persist in believing that a society of smart people is attainable, if provided with sufficient motivating pressures.

But it would be an interesting experiment to do, to assign some random people to a deserted island, give them arbitary access to research material and a set of starting tools, and task them with recreating copies of their original tools before the original ones break.

(Rough Science was sorta like this, but didn't permit the scientists to do research. And of course the tasks were substantially easier.)


But do you need neurosurgeons at all?

Then pick another medical discipline. Funnily enough I had dinner last night with a Physician's Assistant (specialist in theatre work) and an Anesthetist and we discussed this topic. As she pointed out, it took her a solid 5 years to get good at putting people under safely, and recalled an incident with a senior registrar, about to finish their training rotation who get into serious trouble with a seemingly "easy" procedure.

...the French are unwilling to share their documentation, unless we bring in the people as well. If the british government is legally allowed to steal their plans, constructing power plants that are replicas of existing ones...

Seriously, do you think you take the plans for anything, anything at all and just build it and not suffer some massive penalties for doing so - either in time or completely screwing it up. You can't even do that with a "Billy" bookcase, so I'm struggling to think why people would think a Nuclear power station is easier.

Once you have the plans, even if they include the full Bill of Materials and Critical Path Analysis - you still need to actually get a team of Project Managers and Project Engineers to turn that into work packages and work how each part of the sequence will work.

If you don't have people who've done that before and you're working from first principles, it's going to be a complete nightmare. Not impossible, I'm not saying that, but hugely difficult. And commissioning any type of equipment from the manuals without somebody who's done it before is going to be a horrific experience. Honestly, it's bad enough with a team who know what they're doing, I can't imagine trying with neophytes working from documentation.



Your comment at 199:

Stress on rail roads depends a lot on the speed of the trains running on it - stress increases 16-fold when you double the speed (that's right, speed to the power of 4!). If your use of the tracks is disciplined, you can squeeze a lot more years out of them. Unfortunately, disciplined isn't exactly the way they are being used.

When a train runs late - for example because he had to go slower on a track in disrepair - conductors try to make up the delay by going faster. But from the standpoint of longevity of the track, that's pretty much idiotic. It ensures much faster decay of the track.

Also, if someone had decided to downsize design teams in 1850 it would have been a mistake - a mistake that would take a decade to correct, but certainly not quite as much of a disaster in grand scope.

As for 202:

> But in an era of rapid technological change, longevity means building in inefficiency. So we don't do that, much.

But since your premise was independent of the speed of technological change, shouldn't that imply that such inefficiency would vanish and longevity as a concept would prevail? Wouldn't that save much more of the necessary manpower than your first estimate?


I have some experience of the problem of re-creating older technology from a standing start, albeit with the advantage of modern materials science knowledge. On and off for the last 3years I've been trying to re-create medieval alchemical chemistry and medieval pewter and bronze casting techniques.
Firstly, you need a week or three of reading of modern research reports to get an idea of what works.
Secondly you then need to spend at least 3 or 4 weeks actually trying things.
By that stage you probably have a good undersatnding of what works and what doesn't work.

Give it another 2 or 3 months practise and you will probably be able to get it right 3/4 of the time.
Perhaps a year or two's practise, part time, will see you as an expert who can actually do what you want to do when and how you want to do it. It probably doesn't take 9-5 for 3-5 years, as it can do for modern skills, but it is a lot, lot harder and time consuming than many people think.

Just getting the skill to carve the kind of stone moulds they made in the medieval period would require an apprenticeship as a stone mason and several years practise.

I was going to comment on the necessity or otherwise of banks, insurance and suchlike who many people see as unnecessary but are required if you have a modern complex market based society, but it seems that has already been dealt with.


A number of us have assumed in the comment thread that we could perhaps drop back a level in technology, e.g. the 1970's.

However, as Charlie has mentioned in another thread, and is implying here too, we can't just do that. His other thread was about the difficulty of repairing/rebuilding obsolete jet aircraft, unlike their piston engine predecessors. This is bound up with education, experience and the specialized way we need to manufacture things.

My point is, suppose we do try to colonize a new world with relatively few people, what is the highest level of technology that they could even hope to maintain? For us, it might be early C20th. For a colonist a few hundred years hence, it might still be early C20th, something that would be like us living in the dark ages.


I hope I'm not pissing people off with this angle of argument, but anyway.

Well, my point is that you can dispense with certain specialist medical officers. General medics, yeah, you need, because you can't expect to train new ones before the patient dies. But I imagine the number of medics you need to treat common problems would be max(a baseline constant number, a proportion of the population). As long as the baseline constant number is few in number, this doesn't pose a big problem to our minimal society population.

With the nuclear power plant, this will depend on how good your documentation is. Flatpack furniture documentation is horrible. The manifacturers want to fit in as many languages as possible on to the same single A4 side, and this makes for a very unreadable piece of junk.

But what about this?


I'm not sure where to start... actually, let's take the video (dear ghods, a video by that tit from Big Brother!) - first step - lay out all your pieces and make sure you've got them.

Now, think about this in terms of a 5 year build process and get back to me.

You see, if you think that a piece of furniture has bad documentation, imagine working on something where there is no documentation. Most engineering construction projects are working with teams with experience of doing it because you can't possibly document the work on a 5 year build and keep up with it - you'd need more people documenting things than doing them, and there's nothing that engineering fitters hate more than graduates asking bloody stupid questions.

There are operations manuals, and safety guides and start up guides, but these are designed for when everything is up and running.

True story: when they were building the Sellafield Fuel reprocessing facility, they built a complete 1:1 scale working model first, to figure out how to build it. Then they tore it down and rebuilt the production one. Just to figure out the stuff they needed to understand. Even then, there was a joke among the engineering team on site that British Nuclear Fuel Industries (as it was) stood for Better Next F'ing Issue, because new assembly drawings were being issued practically daily as they tried to fix problems.

Now that was pre-CAD, so things are somewhat easier now. But I can pretty much guarantee that there isn't a manual anywhere on how to position, level and align a reactor core block. You just do it, and the more you do the better you are at getting it right.

Same goes across the board in all sorts of engineering fields. Even hiring really bright people and putting them onto something new to "break" into a new market can screw up. Ask Apple.


You are pissing people off a bit, but whatever.

The two fundamental issues you need to deal with in your "simplify medicine," scenario are:

1. Who are you willing to let die or become needlessly disabled to the point that they are a net drain on your hypothetical society?

Case in point: I was reading Sacks' Island of the Colorblind, and he mentioned several times how many perfectly treatable maladies (achromatopsia, cataracts) were left untreated while the people focused their very limited health-care resources on more critical issues: obstetrics (at the midwifery level) and infectious disease. Cataract surgery isn't expensive. Achromatopsia can be "treated" with sunglasses and magnifying glasses. Plus diagnosis.

When you start cutting medical specialties, you waste lives, both those of victims, and those who have to care for them when they could be doing something else. Even hospice care (treatment of the dying) adds meaning, reduces suffering, and massively reduces cost of death, compared to hooking someone up to a bunch of machines and bankrupting their family while they die in pain.

2. Medical priorities change over time. When we get viral epidemics (such as AIDS) we suddenly need a lot of virologists. Ditto with parasites like malaria, or trauma surgeons during disasters or war. Specialists carry the knowledge to train others, and if you let them disappear, you're inhibiting society's ability to respond to crises. Imagine dealing with a malaria epidemic when we're squeamish about funding research into such an icky, third-world disease. Oh yeah, that's where we are right now.

So here's the question that pisses people off: What are you going to cut? Or to put it bluntly, who are you going to cut?


perhaps drop back a level in technology, e.g. the 1970's.

On LJ we've been playing the same game with this and my gut still tells me the 1950s with better medicine and a few other things would be the regression point you'd need to be stable at before you could grow.

Electronics and ultra-high precision manufacturing would have to wait in line behind essential stuff like mining, chemicals, paper and clothing.

Medicine would be easier, at least to a 1980s level because we could bypass a lot of boring trial and error research in statins, BP medication, diabetes meds and stuff that makes a huge difference easily. You could also have CAT Scans and proper ECGs and the like. MRIs would probably have to wait a bit.

The only way I could imagine that you'd avoid a generation or two of pure hell is if you could wave a wand and send machines out to build everything and set it up for you... hmmm... I'm sure I've read that in a book ;)


"My point is, suppose we do try to colonize a new world with relatively few people, what is the highest level of technology that they could even hope to maintain? For us, it might be early C20th. For a colonist a few hundred years hence, it might still be early C20th, something that would be like us living in the dark ages."

I think we can do a bit better. The main reason is that in the 20th century, technological advancement moved in such a way that I think technology, at least for a time, became much more idiot friendly than before. The key is mass production - the basis of mass production is modularising production in such a way that each individual worker only needs to do one small part that requires no understanding. In contrast to the single artisan, single application based model before it, it's relatively trivial to take one person who has never even seen a car before, teach him in a few hours to slot component A into receptical B, do the same with a bunch of other workers, and so start up a production line.

Ferdinand Verbiest was aged 50 when he was able to take on the task of building a model steam car. By the 1940s, the warring nations were creating tank builders out of totally inexperienced and uneducated (sometimes even unwilling) women, slaves, and boys in a war situation in a matter of months, whilst constantly revising production methods. That's the power of good educational processes and modularisable designs.


Trauma surgery is a good example, because trauma treatment manuals don't get re-written all that often and apparently we're in one of the largest changes in trauma medicine since the 1970s... I'll let people figure out for themselves what is causing that.

A good example of how general disciplines have morphed is in nursing. Nursing has generalists, but you want specialist OR nurses and ICU nurses and ER/A&E nurses who just do that.

As a friend said last night, you have to get yourself to the stage where if something goes wrong, you don't even have to think about what to do, you just start cutting and fixing and she reckoned that it took her 7 years of school and another 5-10 years post graduate to be at that point.

Stuff is just a lot more complicated than it seems. As I often say to really really wind up my CTO - "it's software, ones and zeros, how hard can it be???"


The key is mass production - the basis of mass production is modularising production in such a way that each individual worker only needs to do one small part that requires no understanding.

So while I harp on and on about specialization being important, I think this needless denigrates production line workers, especially in complicated modern manufacturing. Skill, training and experience really do matter in the end quality of what you're making, unless you have the shear human and resource availability to make stuff with a high failure and wear out rate.

In the scenario, I think you're going to need to make a LOT more vehicles like a Mercedes, where a team of highly skilled German production workers are building them so they'll last decades, over something that will be falling apart in a few years. There's a reason why Germany has a massive manufacturing sector even with a highly paid and skilled workforce - and I think that your colony won't be able to afford to have a "China" approach to goods.

The massive change in quality manufacturing that Deming wrought in the 60s, really increased the skill and worth of production workers so they could not only do multiple roles in a plant, but also spot and rectify quality issues if they saw them.

And all of this ignores the vast complexity of designing, maintaining and managing a manufacturing system in the first place, and the supply chains to keep them operating.

50 years ago, a factory would carry the stock materials to run for a set period of production, now they carry a few days, or, in extreme cases in the car industry, only what they need to meet the orders that they have in process. So the final assembly is at the end of hundreds of suppliers each with their own supply chains.

That's yet another reason why GM and Chrysler going under would have been such a catastrophe.


"You see, if you think that a piece of furniture has bad documentation, imagine working on something where there is no documentation. Most engineering construction projects are working with teams with experience of doing it because you can't possibly document the work on a 5 year build and keep up with it - you'd need more people documenting things than doing them, and there's nothing that engineering fitters hate more than graduates asking bloody stupid questions."

Well, there's the question. Are we talking about everyone but 100 million people dying off tomorrow? Or do we have some preparation time beforehand? If we don't have some preparation time, then we are in for a world of hell even at 1 bn people, if they are randomly selected. If we do have preparation time, then I think we can do much better than 100 million.

"So here's the question that pisses people off: What are you going to cut? Or to put it bluntly, who are you going to cut?"

Well, if this is a hypothetical moral question, then Tough Choices Will Have To Be Made. I imagine the baseline number of specialists you'd need to have 'a civilisation that is similar to ours' to be approximately the size of the medically trained staff of a medium sized teaching hospital. (I have no idea what that would be. A few hundred?) Or the UK proportion of doctors to the total population, whichever is higher. This would probably be enough to treat most common urgent conditions that don't require the patient to be moved to a specialist facility - I assume the latter category is not large as a proportion of the population, but correct me if I'm wrong.

The people who will be 'cut' under this regime would be those with pressing conditions that require a lot of training to be treated.

Now, cutting population numbers will lose you redundancy, and there's no getting around that. But it will also lose you some risks. Increased population numbers interacting with each other helps fuel the current rise in virus infections, and the rate at which we're running out of antibiotics bacteria aren't resistant to etc.

I'm not arguing this is desirable.


I think the terms of the original thought experiment are clear on this. Even if you prepare, what is the minimum number to operate modern technological civilisation?

I think the preparation issue is a bit of a red herring because there's a limit to the number of specialties that an individual can usefully be skilled at before you just get bottlenecks everywhere.

Or, as my old Just In Time lecturer used to put it, 9 men can't necessarily dig a hole 9 times faster than 1 man, and 9 women can't make a baby in a month.


CAT scans require -- absolutely, I'm pretty certain there's no way around it -- computers with a 32 bit wide address space. I've seen the first ever prototype CAT scanner in a museum; most of its control panel is occupied by a 19" rack labelled "Data General Eagle". It's 1970s tech, but it's basically all about 3D positioning and imaging (using X-rays rather than optical wavelengths and CCDs rather than photographic plates).

The requirement for MRI on top of CAT is the ability to handle liquid helium and make high-field superconducting electromagnets; they actually came along at roughly the same time.

Without CAT or MRI we're basically back to ultrasound imaging -- maybe -- for soft tissues, and 2D X-rays for bone. At which point surgery goes right back to the 1950s and a lot of stuff we take for granted becomes impossible -- hip joint replacements, for example, are fitted to the geometry of a specific person's hip, and CAT scans make it a lot easier and cheaper to get them right.


"Engineers and manufacturers don't want to record what they did in such excruciating detail that even a new graduate could retrace their steps without hands-on assistance" explains why specialized production lines that have been stopped for years are very difficult and expensive to restart. It offers a warning about the consequences of a rapid, drastic drop in population, even one that leaves infrastructure standing. It doesn't seem to be an insurmountable barrier to a deliberate, planned project to formalize a "snapshot" image of the praxis of industrial civilization in preparation for industrializing somewhere far away or somewhen after a collapse in our future.

I think that the formalization of mathematical proof, so that it can be followed by a computer (or green student) which lacks the slightest shred of "mathematical intuition," is an interesting analog to the hypothetical formalization of industrial blueprints or processes. A formal proof is usually much longer than the less formal sort normally written and verified by mathematicians. It also takes a long time to construct: a one-page proof that undergraduates are exposed to may take a couple of days for an experienced mathematician to formalize. However, apart from the advantage of being completely machine-intelligible after formalization, it also can offer new insights by forcing every little step to be considered explicitly rather than glossed over with mathematical intuition.

Likewise, I would expect that a "formal blueprint" for making a light bulb, one that can be followed by the greenest of engineers or even a computer, would be much more voluminous and slow to develop than normal engineering diagrams. Yet it should also eliminate the problem of important steps left undocumented. The other tricky thing, mentioned in an earlier comment, is that if you want a formal blueprint of a light bulb that goes back to "scratch," you also need recursive formal blueprints for everything that goes into its production, all the way back to basic mineral extraction industries.


To paraphrase Gibson, "the technology of the future is here, it is just uneven".

I tend to think that you might be about right about the 1950's. But I'm biased, I was born then, so this is my personal baseline. I also agree that medicine could be updatable, especially as a lot of it is knowledge and technique, and pharmaceuticals are similar.

I guess my real question is is this some sort of "technological singularity" which is hard to surpass until the population, infrastructure and knowledge reach a much larger size, and is a relatively invariant technology level going forwards too, unless we can completely change our manufacturing methods and underlying technology. In other words, is the 1950's still going to be the best target level our colony can start off with (more or less) even in 2200?


The tanks pumped out of the Soviet tank factories weren't unreliable though. Anything but. One just got pulled out of a swamp recently, and still works. The same goes for a range of items manufactured in that period.

I do not intend to denigrate the skills of manufacturing staff. But I have to point out that WWII was maybe the one example I can think of where on a vast scale, a whole slew of different societies found, on very short notice, that they had to re-organise their economies to run off untrained new staff with only a tiny number of remaining trained workers to teach them, under a very tight deadline, resource shortages, and indeed people shooting at them. And on the whole, most of these societies succeeded, and technology did not devolve but went forward at this time. It's just one data point, but still...


"The key is mass production - the basis of mass production is modularising production in such a way that each individual worker only needs to do one small part that requires no understanding. In contrast to the single artisan, single application based model before it, it's relatively trivial to take one person who has never even seen a car before, teach him in a few hours to slot component A into receptical B, do the same with a bunch of other workers, and so start up a production line."

I guess the Germans managed to make V2's with slave labor too. But you forget that the production line is a technological web too, which embodies knowledge and expertise. So no, mass production does not solve the problem at all, it solves the problem it was designed for, to make production as rapid and cheap as possible using standard parts and work processes.


CAT Scans I'll give you, but we've had CT scans since the early 70s and other X-Ray Tomographic techniques since the 1900s, so I think we'd be able to fudge something better than just X-Rays with 1950s/60s tech if we needed to.

You are correct that we don't have nice 3D rotating images though, although I reckon you could, as we did, do that with 16 bit stuff... might be wrong there.

I wouldn't underestimate ultrasound either - very handy technology that I'm told has radically changed a lot of diagnostic stuff.


"That's yet another reason why GM and Chrysler going under would have been such a catastrophe."

I think not. There is a vast pool of car expertise globally, arguably even better than the US in some respects. GM and Chrysler should have been allowed to die, because they simply took up the that could be occupied by better manufacturers. You may even recall that GM made most of it's money in finance, through GMAC, until the 2008 bust blew much of it up.


There's a comparison point in here somewhere. Whether a society is technologically viable depends on many things other than the number of people in it. We're actually being asked to imagine all societies that are basically the same as ours apart from the removal of people, and find the smallest one with a technology that's viably the same as ours.

Which could involve the question "how much preparation can be allowed for", with a parallel point "if preparation can improve things, the number is likely to fall with time".

As another question, how much really matters? Is, say, landscape gardening a technology? How about the screens on the back of the seat in front of you on aircraft that show films? If it's entirely the same, including having leaf blowers instead of rakes and so on, then it's just a question of shrinking the population until we run up against the limits of skilling people. Which is going to involve some rather odd situations where advanced specialists work on their specialties on monday and teusday, but then go around doing gardening wednesday to friday. Furthermore, would there be a larger educational time then present? Would there be subsidised university education more than once in a lifespan?


@Charlie At which point surgery goes right back to the 1950s and a lot of stuff we take for granted becomes impossible -- hip joint replacements, for example, are fitted to the geometry of a specific person's hip, and CAT scans make it a lot easier and cheaper to get them right.

Hip replacement was being done reasonably well in the late 1970's, early 1980's without CAT scans. CAT scans and robots have certainly made the fittings much better, more like the transition from artisan to modern manufacturing.


So how would this get handled in the real world? Let's imagine that we drop a medical clinic and a bunch of healthy, experienced farmers and miners on our colony world, and these folks thrive and produce children. When the first kid turns three do we start building a school and unfreezing the teachers, or shipping them in from Earth? Do some people stop mining and start teaching? Plenty of room for conflict there...

When we have a small pile of ore and we know there's more coming, we've got to build a smelter. Do we drop one from orbit? Build one of bricks and clay? Bring in a group of smelter techs? Is it worthwhile to have a trained blacksmith who can take things from an iron ingot (or even a pile of ore) to braces, screws, and tools, or do we do better with a whole chain of specialists? Or do we need both so we can drop down to a technologically easier level if necessary?

It isn't just a matter of "how many people do we need for a working colony?" It's a matter of bringing in the right people at the right time, and having them available (or on hand in cold sleep) and ready to go, along with the technology they need, plus smart generalists and raw materials for the inevitable "Things Aren't Going The Way We Planned" moments.

Things get interesting, of course, when the colony has to leave their critical path due to X, Y, or Z, or because someone did something unexpected with the technology - using newly built motors for tanks instead of tractors, maybe.

It's entirely possible that the national holiday of a particular planet might be the day they built their first homegrown electric motor, or their first solar panel...

I'm sure someone reading this knows more about the subject I'm discussing than I do. Please enlighten me on how this works.


Yes, but WW2 wasn't remotely sustainable. As Charlie pointed out elsewhere in this thread - by the end of it, Britain was falling apart.

Plus manufacturing industry in the 1930s was/is nothing like it in 2010 - Deming changed pretty much everything about the way we build stuff.

I think we're talking about two things here. The bare minimum to run a society that will eventually collapse, or running something people recognize as a society.


I think not. There is a vast pool of car expertise globally

Oh sorry, not a catastrophe for the car industry, I don't think that would have even noticed globally.

I meant that JIT methods and the nature of the supply chains are why their collapse would have been a catastrophe for the US regional economy.


...hip joint replacements, for example, are fitted to the geometry of a specific person's hip, and CAT scans make it a lot easier and cheaper to get them right.

That's certainly what Ortho surgeons and the manufacturers like to say :)

We did a module on this when I did my engineering degree and it was an interesting example of an amazingly precise piece of engineering, fitted in place by a couple of guys doing a fair impression of a chippy hanging a new door.

There's a lot of work to be done on Hip replacements and I hope to god it gets done before I need one.

My brother had his first a couple of years ago at 50, and the recovery wasn't pretty.


"I guess the Germans managed to make V2's with slave labor too. But you forget that the production line is a technological web too, which embodies knowledge and expertise. So no, mass production does not solve the problem at all, it solves the problem it was designed for, to make production as rapid and cheap as possible using standard parts and work processes."

Yeah, but a production line can be written down, thus dodging the 'some things require years of real experience to acquire issue'. My point is that pre-mass manufacture, Von Braun would have to assemble and train a team of skilled rocket artisans for the best part of a lifetime, each of them individually making a rocket, capable of feeling the right curvature of the pointy bit by hand etc etc.

Whereas by WWII, Von Braun could be shot in the head, all the V2 factories bombed into vapour, but as long as an engineer lived in Germany who knew how to read the plans (a transferable, generalist skill), the Nazis could restart V2 manufacture within months. The point is that industry had progressed to the stage that standard parts and work processes included the workers, and the unique designer only has a role in producing the original.


I hope I did not imply that the maintenance came free.
It was a lot to try to cram into a single post about trying to draw a dividing line where design and manufacture end and maintenance and operations begin in terms of how many people it takes to accomplish each function.

And yet the one thing that you have to appreciate is there is a design that has truly withstood the test of time when it comes to railroads...the coupling mechanism for the rolling stock...The knuckle coupler or Janney coupler that was invented by Eli H. Janney and patented back in 1873. Not much has changed except the name. Oddly enough, this is something the US appears to have gotten right, as once Congress passed the Safety Appliance Act mandating conversion from the link and pin coupler to the Janney coupler, railroads in the United States had only a few years to implement the change. The railroads in North America, except for mass transit, form one unitary system, and uniformity of couplers is important for smooth interchange of rolling stock.

and wikipedia

from your 202 comment:
An elderly liion cell is useless as just about anything other than a fire-starter (or recycling input for new batteries).

If it is large enough does it fall under the category of yet another "boat anchor"?


>>"That's yet another reason why GM and Chrysler going under would have been such a catastrophe."
>I think not.

Have to agree on this point. Letting GM and Chrysler go under might have allowed other manufacturers to step in and take over. Got into a conversation with a Ford salesperson on a flight out to California. Turns out he was in charge of all midwest sales, so we got into a discussion about the longevity of vehicles. I explained that I bought a BMW back in 2000 and it was still running perfectly after 170,000 miles (routine maintenance aside). I wanted to know if Ford came close, and sadly he said no. About 100,000 miles on most Fords. So do automobile manufacturers outside the US make better cars. Absolutely. And keep in mind Ford was the real survivor of the big three over the last two years.


On the flip side, Britain was fighting a war in WWII. I don't see the reasons Britain was collapsing Stross outlined to be relevant to our scenario. Services were undermaintained, because production capacity was allocated to military production.

If you look at GDP figures, you can see that overall GDP increased or remained stable during the war, while the proportion of GDP spent on the military exploded to about 50% for the US and the UK, and 75% (!!!) for the USSR and Germany. If we are talking about WWII style economic shifts without a war draining resources, Britain would be far from collapse by the end of the war. It might be better off than at the start of it.


... fitted in place by a couple of guys doing a fair impression of a chippy hanging a new door.

That's how you know they were good, right?

Because if you miss by a millimetre with a single screw when you're installing a door, the door-frame doesn't bleed to death all over you.


Whereas by WWII, Von Braun could be shot in the head, all the V2 factories bombed into vapour, but as long as an engineer lived in Germany who knew how to read the plans (a transferable, generalist skill), the Nazis could restart V2 manufacture within months. The point is that industry had progressed to the stage that standard parts and work processes included the workers, and the unique designer only has a role in producing the original.

Nope. Read your history. The allies captured only a fraction of the V2's. The real prize was von Braun and his associates. The Russians in contrast had the factories, but not the expertise. It was Korolev who kept them on track until his death.

Rocket technology is an interesting issue to illustrate Charlie's point. The Russian space program is incredibly fossilized, cheap by our standards, but using 50 year old technology. As Charlie has suggested in another thread, the Falcon 9 is basically a private version of the old R-7 technology, spiffed up a bit. The US in contrast has created quite a wide variety of rocket engines, including scaling up technology of kerosine/LOX to the Saturn V engines. It is also an aside about un-codified knowledge that supposedly we could not recreate those F-1 engines as the plans are no longer available and no one knows exactly how the chamber pressure instabilities were solved with the baffles. In effect, we cannot rebuild the Apollo ships, and arguably Nasa has regressed, unable to effectively build a replacement in the Ares.


I think that arguments about mass production lines solving everything are a low level form of "magic wand" for the colonists.

Another magic wand might be "what if we could maintain our technology with self repairing robots"? These devices would not need to be trained, presumably could replicate and simply get their expertise from the ship's library, or possibly updates via transmission from earth,


If it is large enough does it fall under the category of yet another "boat anchor"?


If you ever feel like performing that experiment with a liion cell, please warn me so I can arrange to be a long way away at the time. (Preferably with a webcam so I can watch the pretty fireworks safely.)


Some things will wear out if not used enough. Collectors of pinball machines are constantly looking for people to come around and play the machines because they break down if not used frequently.


Well, reading my history, Von Braun had very little to do with actual manufacture of the rockets - little enough that he got away with claiming not to know the conditions his workers worked under. The allies needed Von Braun because they wanted to make new rockets that aren't V2s. You need a designer who understands the ideas behind the design to do that.

The Russians didn't even have blueprints, being forced to reverse engineer V2s parts to reproduce them. Still, it only took them two years to do that and make eleven new V2s. Due to budgetary differences, the Soviets actually exceeded the progress of Von Braun's team during this time.

The question we have here is not what we need to advance rocket technology, but what we need to maintain it. And Von Braun is surplus to these requirements. I really don't get what you are trying to argue here.


People have been thinking about this a lot. Read "I, Pencil" or The Toaster Project.

Our civilization expects a certain amount of improvement. Cellphone were pretty much stagnant until the iPhone came out. Then look what's happened in the past 3 years.

A civilization that has gone into maintenance mode is at best in a dark ages and at worst in a death spiral.
Maybe it entered into it involuntarily because of a megadisaster.
If chosen, then they have snuffed out innovation and creativity and even the very question, "There has got to be a better way?"

They won't have scientists and engineers or much in the way of entrepreneurs. Just technicians who now how to replace a broken part, but not how to create it in the first place. They are one disaster from losing a particular capability forever. If a fire takes out the factory they couldn't replace even if they wanted too.

It would be like living in East Germany all over again. No thanks.

Now a space colony is entirely different.
The colonists would do a cost benefit analysis that has completely different variables then the stay-at-homes.

At some point it would make sense to take a semiconductor fab rather then a bunch of computers and radios and such, and have it make the parts for computer, radio and such on and as needed basis and a carry bunch of different programs for different designs.

Add some 3D printers, repraps and the like and you can make everything you need when you get there.

A space colony would have to know what feedstock exists and those are more variables to consider.


Because if you miss by a millimetre with a single screw when you're installing a door, the door-frame doesn't bleed to death all over you.

Yeap. But all the tech in the world doesn't improve on surgeons who have a LOT of surgeries under their belt and an eye for what's right and what's not.


The Russians didn't even have blueprints, being forced to reverse engineer V2s parts to reproduce them. Still, it only took them two years to do that and make eleven new V2s. Due to budgetary differences, the Soviets actually exceeded the progress of Von Braun's team during this time.

This depends entirely on the level of time pressure involved. If we're discussing a Laundry novel, where unexpected developments must be handled quickly, then you need the existing production line, and experienced engineers and experienced line workers, and an experienced Von Braun.

On the other hand, if we're exploring how a colony might develop during the back story, we don't really care if it takes an extra ten years to reinvent the disruptor beam - nobody cares in that case except Charlie will state the number of years it took to reinvent the industry, and he wants the number to read as if the author has done his thinking. This makes for a better book, but it's not otherwise a Big Deal.

However -


I suspect it matters.


Cellphone were pretty much stagnant until the iPhone came out.

What exactly was "stagnant" in the cellphone industry before iPhone? What the last 3 years have shown is that interest in cellphones in the US has caught up with where it already was in the rest of the world 3 years before the iPhone came out.

So the US carriers are finally scrabbling to get networks out there that deliver the sort of service quality everybody in Europe (for example) doesn't even question. "Fewest dropped calls?" Eh?

The only time I'd experience dropped calls in London would be between 1pm and 5pm on a Friday afternoon while 6m people tried to find out which pub they were in. The rest of the time? Dropped call? What's that?

The iPhone has birthed a new UI paradigm, and by having the solid walled garden of iTunes in place, they were able to spawn an app market - although you will recall that Apple were highly resistive to that to start with because they knew that up until then, nobody had made a penny from apps, even though app stores have been around since the late 1990s.

Even then, there were lots of features that iPhone didn't have that have been standard on other phones since the mid-00s.

There are now 5 BILLION mobile phones in use on Planet Earth by 2015 it'll probably be close to 100% penetration globally. At the moment a fraction of those will be Smartphones, but to say that the cellphone industry was stagnant before iPhone is laughable.


there are whole academic disciplines out there exploring cultural evolution - mostly in the past because that's were the admittedly ropey data sets are to be found.

The Tasmanian case is fascinating; 'exile' from the Australian at the end of the last ice age and loss of any marine technologies leaves you with the people the Dutch 'found' in the in the 18th century. Here's a start (there's pdf for free somewhere on the interwebs )

J Henrich - American Antiquity, 2004
"Demography and cultural evolution: how adaptive cultural processes can produce maladaptive losses: the Tasmanian case"

No social or technological complex is stable. Not only is it subject to initial conditions, but niche construction influences every generation.
Even if you could 'bottle' early 21st century civilisation, it wont survive implementation. Will head office mind? How much does it matter if its reinterpreted
This brings to mind U Le Guin's "The word for the world is forest" couldn't bear to read all of it though.


My issue was about the economic fall out of not acting on the US car companies, not whether or not it might have been preferable to let them die.

IMO it probably would. Ford were saved by having a massive global market of decent cars, I've owned several in the UK, I have no idea why, when they had the Mondeo in Europe, they'd even bother making and selling the execrable Taurus in the US.

The issue isn't if other companies would step in, but at suddenly removing the end of a very long supply chain from the economy and watching everything fall apart in a matter of weeks. Modern car manufacturing doesn't carry stocks, so you have 1000s of suppliers making everything from windscreen wiper blades through to motors. While some of these could find new markets for their products, a lot would just shut up shop, and then their suppliers would shut and...

The cost to the US economy would have been terrifying.

The British proved in the 1980s that shutting down large industries without a Plan B is a really bad idea, even if those industries need to be allowed to shuffle away.


Just my two-penneth on the training/learning by experience debate.

I'm currently doing an MSc part-time funded by my employer in a field directly related to what I already do.

It's great at teaching me WHY I've always done what I've been doing(and seen being done). I can combine it with what I've learnt on the job and do new things I couldn't have done before. But just doing it on its own wouldn't teach me the job. Absolutely not. Not even close.

There's two layers. Technical concepts sometimes need to be actively taught. You need showing. You don't get from just reading it.

And the there's how you apply this stuff, that you get by doing. Ideally starting small with some guidance on the job.

This doesn't only apply to medicine and nuclear power stations, but all those other professions and specialisms (how many?... 100,000?)

You can learn from scratch and muddle through with no training but it will take ages and the consequences of learning by getting it wrong had better not be too severe....


Are you really getting into a new answer to the Fermi paradox? The answer you seem to be implying seems very similar to the one you already wrote about in Accelerando, just expressed in slightly different terms.


>if chosen, then they have snuffed out innovation and creativity and even the very question, "There has got to be a better way?"

Too late...(or so at least Newsweek is telling me)


This may not happen over there, but we have rails get permanently misshaped by heat. Then they have to be replaced.



> Rocket technology is an interesting issue to illustrate Charlie's point. The Russian space program is incredibly fossilized, cheap by our standards, but using 50 year old technology.

Absolutely not.

While it is true, that 90% of the Soyuz rocket is more than 50 year old technology, it is not the whole of it. For starters, there is the Proton Rocket that was developed in the early 60ies to launch 20-50MT nukes to the US (never deployed as such, since they got a lot better at aiming after that).

Then, there is the N-1 Hercules, that was supposed to be the Soviet Moon rocket. It was built in the late 60ies. It used copious amounts of NK-33 engines that are still in existence and will be used for the American Taurus II rocket (build by OSC).

Then, there is the Buran. While the Energia rocket that carried it has flown only twice, the booster rockets could fly all by themselves. In 1985, the SU introduced it as the Zenit rocket that has seen use by both the Russians and the American corporation SeaLaunch.

The Zenit used a rather powerful engine RD-170 that has been introduced into the US in a downscaled version called RD-180, used by the Atlas V.

The only modern piece of equipment that the US has to offer is the Delta IV RS-68 Engine, which was introduced at about the same time as the RD-180. (And don't get me started on the Merlin 1C - it is new, but it is not modern.)


I have seven doctors; all but one are specialists. I'd prefer not to die, thank you.


Robert has the right idea in pointing you to Jane Jacobs: I think that she's already done a lot of the thinking you need for this problem. "Cities and the Wealth of Nations" would be my recommendation. Now, she was more concerned with quality of life and community than of technology and civilization per se, but the thinking is similar.


In determining the minimum number of people, are we going to consider just the number required to do the work, or also the number required to create demand (and pay) for the work?


@225 they do that in Cuba!
our tour guide was a history professor for part of the year.!
seems thats common there, doctors that work as bar staff for some of the time


>>>And at the end of that seven or eight year period you have, at best, produced an illiterate juvenile hunter-gatherer with minimal language and no social skills: little more than an animal.

Charlie, this is the first time ever you've posted something I have to take severe exception to. I know it's a slip of the keyboard rather than a serious, firmly-held belief of yours, but believe me, hunter-gatherers, even the juvenile ones are certainly not 'little more than animals'.

Hunter-gatherers survive (especially in the harsh environments in which the remnant h-g populations of today live in) precisely because they have the numbers to support a certain base of knowledge of their environments and how to exploit them. Even where this knowledge is connected to systems of mythical and religous thinking, there is a cold hard, human, rationality at the heart of it; survival demands that rationality.

I'm frankly surprised at you, to be honest. This is the kind of crass error the yankee 'libertarians' you regularly roast on this site would make.


Really? Why?
It sounds so counter-intuitive.


There's a lot of mechanical systems that don't like to be left alone for a long time. Most typically, in small electro-mechanical stuff you'll get dust adhering to lubricate which basically causes poor circuits and stuff to stop moving properly.

In cars, apart from batteries slowly dying, you can have too much "crud" collecting in the oil sump and then having a tour of the engine when you start it up after a long period of time.

That's without looking at rubbers and plastics that get brittle if left cold for long periods of time so that as soon as you put them back under a load they'll break or start leaking.


Whoops; bad phrasing on my part. Yes, you're right about hunter-gatherers being well-adapted. What I'm saying is, raising children is a community activity and very labour-intensive, and in the absence of those inputs you get something that at best can forage for food and has some language. Not remotely as well-adapted as a real hunter-gatherer.


Regarding those T-34s, a few years ago, a Brewster Buffalo was pulled out of a lake in Finland. It was so well preserved that the tyres were still inflated. Who made the tyres?

Nokia, and I bet they didn't have any great base of mystical expertise in aircraft tyres when they lost their only source for natural rubber in 1940. They did, however, have a culture of good engineering.

That still doesn't explain why the Finns could make something of a flying coffin like the Buffalo though.


'What is the minimum number of people you need in order to maintain (not necessarily to extend) our current level of technological civilization?

My personal observation on this - our current 'level of technological civilization' is not a state - it's a vector.

The technolgical infrastructure of our society is fundametally built to develop and improve our technology.

Take for example the 'Apple' IPhone. It says on the back 'Designed by Apple in California Assembled in China'. To strictly mantain our level of technology we simply do not require Apple. We do need Foxcon, because our IPhones will wear out and its them that build the new ones. It's built from chips largely sold to apple by fabless electronics companies - we don't need them either, or the majority of their suppliers - but we do need TSMC and ASE and we do need their supply chain such as Applied Materials - but only the portions of those companies that produce, not those that innovate.

Silicon chips are typically designed to last for around 10 years of continuous operation at a reasonable temperature (the limiting factor is electromigration, metal atoms being knocked out of position by passing electrons in critically sized locations so the interconnect wires flow away). So to maintain our current level you would need to reconfigure this supply chain, post event, into 'maintenance mode' within that time period. (Which is ALOT of work - it's not designed to last because it's expected to be superceeded by a new generation within 2 years and be hopelessly outdated after a few more).

I guess you could argue that there's no need for that final redesign, but your resulting society would be hugely prone to further degredation if you did not.

In terms of numbers - i think you first need to decide if it's state or vector you wish to maintain - becuase there is a very big difference in headcount.

And if you did choose 'State' remember you've not chosen 'limited new toys' you chosen 'replacement only'...


One thing that's been largely ignored so far is the issue of time. The case of the instant die-off of half of the population is quite different from that of a slow 0.7% annual population decline for a century, though the result is about the same. Businesses will plan, migrations will happen, and plans will be made. Some difficult-to-propagate knowledge may even get codified somehow.

If progress isn't happening any more, some industries could be simplified. You don't need 200 different varieties of cell phones when there is no "new hotness" on the market. I suspect there would only be a few models representing the different expense ranges. For a situation like this where the designs are set in stone (or perhaps re-engineered every 50-100 years) you might not even need to have a full technical knowledge of how the thing actually works in every generation; just the knowledge of how to follow the instructions to build the thing successfully (still substantial).

That brings up the idea that instead of having people who know how to do everything alive all the time, perhaps you could "skip a generation". This would be very difficult, as Charlie has brought up, but perhaps not quite impossible. I imagine one generation of workers building (and mostly mothballing) the flash chips that will be used for the next 100 years. They're trained together (some going first, more slowly, and teaching the others), work for the same few decades, and retire together. When they're done, perhaps the capital equipment (if it still works) is used for fabbing low-power CPUs or something similar. The chips they made will be used up over time, and when it looks like they will run out in a few decades, the next generation of flash chip makers will be trained.

How could this be possible? It would require massive forethought and investments spanning centuries. I'm not convinced that people have the patience for such plans, but maybe... The existing training methods would be used, of course, but other techniques would have to be invented too. An equivalent of on-the-job training would have to be stored for the next generation. One advantage is that the natural tendency of people to keep their little tricks secret (to prevent competition) would be reduced since the competition won't be born for many decades. Perhaps a seasoned veteran at (job) would wear a webcam pointing over his/her shoulder for a year, recording the output. The future replacement would watch it, piece at a time, (not all at once; that would be sickening!) during training.

If something like my vision is possible, you may be able to get far below the 100M population minimum, though it would obviously be brittle to losses of knowledge, failures to re-start industries or unexpected needs for particular expertise.


I thought it was obvious that Charlie was referring to the outcome of a child brought up by one or two parents with no input from any sort of culture. Of course, any real hunter gatherer culture we know of is an excellent example of Charlie's thesis that we need a large number of people to maintain any technology.

The technology of the hunter gatherer lifestyle can be immensely complex and requires much education and practical training in plant identification (which yams can be eaten just cooked and which require more treatment to detoxify them, for example), seasonal variations in production, geography, animal behaviour, tool production etc.

The hypothetical two-person input hunter gatherer might be close to an animal, or even lower, as granny elephants have been shown to pass on their knowledge to future generations. The product of a multi-person "primitive" hunter gatherer culture will have intricate language and social skills, much leisure time to produce art and criticism and may learn specialisms they will be required to pass on to the next generation.

Having to re-learn which things that look like food are actually poison would be as fun as learning to do brain surgery from Grey's Anatomy (the TV series not the book).


Charlie, you misrepresent conservatism. As an American conservative/Republican and -- as you may remember from the MacPerl days -- computer guy, I understand technology pretty well. Conservatism is small-government, yes, but it's not about reducing the complexity of society, but about giving people more freedom and more choices.

Certainly some conservatives want a simpler world, as do many liberals: go to any community college and witness the "progressives" wearing their homemade hemp clothes telling each other that life is all about love, man. This obviously is not unique -- and in my experience, significantly more prevalent -- on any part of the ideological spectrum.

Your thesis seems to be primarily "conservatives are fundamentally wrong because they want a simpler society that would make much of what we require today impossible," but that thesis is fundamentally wrong because conservatives don't believe the view you attribute to them.

Hell yes, I -- as a conservative -- think government should be much smaller, especially the federal government. But that doesn't mean I want less people working on cars and spaceships.

Framing this in political terms is a failure.


I defer to your greater knowledge on this.

I do agree that the US technology isn't that new, but I'm surprised that you say that Russian rockets are new[er].

I will need to read my books on this more carefully before putting foot-in-mouth.


Sorry, Charlie: I was trying to understand the "huge political implications". I leave the science fiction to better minds than mine.

Western society is amongst many other things a machine for filling every conceivable niche in every conceivable market. Or, put another way, a machine for building complexity. So of course it supports a lot of specialisms.

But a community under threat can be organised differently. Then the question becomes, "which specialisms do we really need to support a viable society?".

There is a big question as to whether we need to afford the sort of internal competition that leads to explosive improvements. The mindset can be changed to need (rather than want) and the greater good in a heroic enterprise. If the external threat is clear and present, that will in itself spur rapid advances.

Perhaps the notions of acceptable waste of life change: if you get certain diseases you die (because we can't afford those specialisms) but no-one has the right to put their life at risk unnecessarily by riding a motorcycle. Or something.

"Leisure? Please explain the concept?"

"Art? Is that another word for design?"

I am not convinced that our current system of trying to give our children something to think about while they master reading, writing and rational thought is either effective or efficient. And class sizes of 10??

In a minimally sized society, they could presumably start their real apprenticeships much earlier with some arrangement for sorting out square pegs and round holes a little down the line.

Passing on the know-how and cultural coherence does become a matter of life and death for the small society, though. Even if the methods change, this is clearly a major activity.


Bear in mind also that some of the roughness in a hip replacement operation is deliberate: they need to beat up the bone so it bleeds and bonds properly to the implant. No, it's not nice. But you'll be pleased to hear that the recovery has improved a LOT in recent years (I have an uncle who wore out an artificial hip in about 10 years, he was much happier after the second operation).


it would seem that this question is hinged upon so many socioeconomic and technological unknowns the range of error for prediction may get awfully close to the ranges offered up. As of right now so many of our systems are redundant and unneccessary, ie( most trinkets and gadgets) and shipped halfway around the world. It would seem most of our production and output is put towards some strange mutant mockery of sexual selection displayed by both sexes including high speed automotive penises and bacterial toxin lips. If theses things qualify as neccesarry to our modern comforts then we may need a very large percent of us indeed. However if we imagine cases where radical social and technological changes the numbers may have less to do with keepng up production and more to do with maintaining genetic diversity. As information mining and collecting become more abundant humans may be able build and create and troubleshoot in ways few of us now are able. I am reminded of a show about oragami which featured a young man who was homeschooled by his father and became the youngest professor at MIT, anyways his father mostly just taught him to do as many puzzles as possible from a young age. No doubt this kids a genetic genius as well and is not typical but if education radically shifted from fact based cirriculum to problem solving and data analysis perhaps the technological achievment would be fast. However this has always been the techoanarchists dream, but the reality is humans do not have a good record of spreading technology without using it to exploit others. In order for a big social change some giant unforseen memetic and possibly genetic selection force would need to be present. Who knows could early humans have predicted us.


Collectivism and libertarianism remind me of the old joke about women. There are two theories on how to deal with them. Neither one works.


I’m leaving it late to reply, however...

Just to clarify an earlier point: the distinction between what is possible and probable. It seems to me feasible that (if you believe that humans are essentially biological machines, as many sf writers do) every human-based industrial and manufacturing process can be artificially replicated – eventually. There could be a tipping point beyond which the machines (AIs etc) could assume creative superiority, or rather decide they know what’s best for us … until they evolve beyond their programming sufficiently to decide what’s best for THEM. So, if it can be analysed it can be replicated, given enough time and the right initial programming.

Probably, though, psychological factors will not allow this: people becoming redundant, lazy, dis-empowered. And moral judgements. Also the philosophical considerations of what it means to be human. Perhaps seeing the tipping point as the much mooted singularity – when the machine argues there is no evidential difference in its perception – is the time to pull the plug, so to speak.

Some fascinating and complex arguments about contemporary civilisation, btw.


You make a wonderful point about the benefits of specialisation: it allows specialists to work to tighter tolerances. IE your general surgeon will kill a lot less patients than a GP conducting his first (un-supervised) appendectomy.

To you that makes a huge difference. To a colonial society that increased risk may be worth the benefit of not having to maintain that speciality.


@255 In Cuba, professors work as tour guides and doctors work at bars because they can get tips in dollars; those tips are worth significantly more than their annual salary. Highly-trained experts working in the service industry, or being forced to labor in the cane fields (as also used to occur), is not useful to winnow the necessary population to maintain a tech base. *News from Nowhere* uses this model, and it's not a coincidence that technology has been mostly abandoned in that novel.


Very interesting questions. Clearly education and training are key issues. Also, related, is adaptability eg how to apply learned information in different ways.

Information technology is allowing some(to-date albeit pseudo) enhancement to learn and adapt. Per Kurzweil et al, IT is on exponential curve. Efficiency results from increased learning and adaptability eg using new basic science and multiplying benefit across industries. Charlie noted benefits of this:

"It's possible that it doesn't. Remember, today only about 0.5% of our population are needed for agricultural work, about 2-5% for industrial production, and probably the same -- or fewer -- for resource extraction. "

Perhaps what we haven't really seen is the emergence of, at least en masse, a "real hunter-gatherer" that marries enhanced education due to technology to adaptability.


@heteromeles, if you read this far: the overlap in our interests is intriguing. Can you send me an email? You know where to find my address, and I'll respect your privacy.


Most of the Buffalo use was in the Pacific, against fighters such as the Mitsubishi Zero (which came into service in summer 1941). And the pilot training, experience, and tactical doctrine was heavily in favour of the Japanese.

The Allied Air Forces had to develop tactics which took advantage of the differenct characteristics of their 'planes. Don't get into a turning battle with a Zero. And they had better 'planes than the Buffalo coming from the manufacturers.

Finland didn't have the replacements, they likely had the pilot quality advantage, and they had a breathing space to work out how to use their 'planes. There's plenty of reason why they got more out of the Buffalo.


Oh well, while the foot-in-mouth-disease can't be cured, at least there's no culling. (And I should know ;) )

I think that Russian technology was starved for money, while American technology was drowned in it. The result isn't that pretty on either side.

The real problem, though, seems to be a lack of actual use. The number of space ships and satellites launched in recent years was a lot lower than in the 90ies - mostly due to ubiquitous glass-fibre connections on the ground limiting the need for communication satellites.

We'll see better days some time.


Actually Foot and Mouth disease can be cured, more importantly it can be innoculated against. The reason it isn't is purely economic...


That may be so.

But it's a different illness from the foot-*in*-mouth disease. And trust me as one of its sufferers, it *can't* be cured.

(Occasionally though, you may be able to hide the symptoms.)


Terminology alert:

Foot and mouth disease -- contagious viral disease of cloven-hoofed cattle, aka FMD.

AIUI, FMD is known as "foot and mouth" in the UK and elsewhere but as "hoof and mouth" disease in the USA, where "foot and mouth" disease is sometimes used for anthrax, just to confuse the hell out of foreigners.

Hand, foot and mouth disease -- human syndrome resulting from infection with intestinal viruses.


One of the Buffalo's biggest faults was that it was prone to overheating, so that made it worse in the hot Far Eastern theatre and better in the cold Finnish theatre. The Finnish pilots were very well motivated too - they were literally fighting for their country's survival.


Just as a point on landing on other oxygen atmosphere planets, for quite a while, the earth had an oxygen atmosphere, but no land based life that we know of. Such an environment might well be conducive to colonisation even without compatable biospheres, unless their life really can feed on ours without being noticable to our immune systems. I have this notion that we contain many, many chemicals, and I'd bet that some of them would be poisonous to a non parallely devloped life system.


Foot and Mouth disease can indeed be inoculated against. The problem is that an inoculated animal would test positive for the disease as it would be carrying the antibodies for the disease in its bloodstream. Only if a better test became available that would not flag an inoculated animal as suffering from the full-blown disease would it be possible to inoculate existing herds to protect them.


No worries, Charlie, I didn't think that was what you meant.

By the way, I saw a lad reading The Fuller Memorandum on the street here in Dublin yesterday. He had chest-length hair, a beard, black clothes and a studded leather belt. Your key demographic? ;-)


Wow, Charles I also asked myself exactly the same question :-)

As a scientist I would extend the problem. Not only would you need all the people to build and develop stuff. But if the ultimate goal is to have a vibrant society that can develop new concepts in science, fiction and art, then you need to have enough people who can sustain these fields.

As a scientist I would say that there is a minimum of experts in any one science field necessary to develop anything new. Having just one person who is an expert is not science as science only works in a competitive environment where new ideas are cross checked for their validity. (This does not only encompass peer review.)

Still 100 million sounds like a reasonable lower figure.

In the case of mars, I would say that you could get away with much less. As mars is close enough for a fairly rapid communication. Hence, a lot of stuff developed on earth can be communicated to mars and you don't need a lot of people to just build stuff.

Let's say humanity finds a nearly magical way to reach for the stars and build colonies in other star system. As long as communication between earth and these system exists a high technological level would be sustainable. If communication would break down, for what ever reason, those colonies would stagnate and the technological level would recess. If a level of technology is necessary to sustain a living habitat then the recess could lead to a local extinction.


An idea I first came across in an SF RPG setting with one-way (cannon style) JAFL colonization:

You probably cannot make a 20C equivalent society out of a few hundred folks and a few hundred tonnes of equipment. (Certain kinds of very optimistic tech might allow this, but this was conservative tech SF.)

However, you could send a starter set of equipment, including a very large ruggedized library (compact even with current tech), a bunch of other very useful highish tech stuff that had been ruggedized out the wazoo (frex: a rugged nuke with minimal moving parts and a 300 year fuel supply), various backups for critical path stuff, and a very complete plan/program that distilled the work of thousands of specialists in colonial development into a single compact package. "You appear to have taken over the fledgling colony is a military coup, killing critical specialists. Would you like help? Y/N/Die COLClippy Die!"

It might be worthwhile to make something similar in order to deal with a potential apocalypse-like-event on the Earth, but it is the sort of thing that won't make you rich or make anyone vote for you, so I'm nut sure how it might get funded.


To a colonial society that increased risk may be worth the benefit of not having to maintain that speciality.

There are two things here to consider. Are we talking about a colony living a "try and survive at all costs" life style. OR, as Charlie's original point was, an advanced technological society.

In the former, I don't think you'd want to risk losing useful members of society, in the later, I don't think you'd meet the criteria if your medical services we so constrained.


I suspect a lot of soil algae, fungi, and cyanobacteria would disagree with you on the "no land life to speak of" prior to the Devonian (heck, I'm guessing prior to the Cambrian). The only reason it looks like there was more life in the sea was that subaquatic environments do a better job on fossilization, especially of microscopic organisms.


"I suspect a lot of soil algae, fungi, and cyanobacteria would disagree with you on the "no land life to speak of" prior to the Devonian "

Well at least least the worry of contamination by micro organisms can be solved by wiping your feet on the doormat as you enter your habitat. ;)


To a colonial society that increased risk may be worth the benefit of not having to maintain that speciality.

The problem is that as we continue to advance, what is acceptable to let go becomes harder and harder. Imagine being asked to today to give up all surgical techniques since 1950. Not easy. Now ask whether people in the C23rd could live with surgical techniques of 1950...

As McCoy says "Unbelievable. It sounds like the Spanish Inquisition to me".


I suspect that if you offered them the chance to go to Space, a LOT of people would accept a severe reduction in the number of specialisms.

Do you get a significant saving in the number of people if you take out, say, 75% of the specialisms at the top of the tree but keep the remaining 25% at the same level? Are there whole groups of specialisms which pretty much go together and could be sacrificed (or downgraded) en bloc?


There are, as far as I understand, no fossils of macro-scale life (which could probably mean almost anything, granted) from before a certain era.

From wikipedia's article on the Cambrian era.
While life prospered in the oceans, the land was barren — with nothing more than a microbial 'crud' known as soil crust gracing the soils. Apart from tentative evidence suggesting that some animals floundered around on land, most of the continents resembled deserts spanning from horizon to horizon.

It's also possible that crude life is less capable of competing with transplanted life from earth than more advanced life would be - it may be that if a robotic probe finds a world with an oxygren rich atmosphere it would be reasonable to drop onto one of the larger continents and distribute grass seeds.


I'm just as interested in the implications of a society that might overcome these sorts of issues - say if you lived in a future world where you could reliably become your own temporary expert on nearly anything with the aid of computers, and manufacture anything, given enough time, within the comfort of your own home somehow ala Star Trek.

I mean, clearly then you could lock the front door and have a technical society of one - but we're not built that way, and in any case a society of one isn't very redundant unless there's a way of replacing your corpse in case of accidents too. A billion sociopathic nations with cabin fever?


I think tp1024 was being sarcastic -- that a lot of people say things they shouldn't.


I'm not sure, and I'm certainly not Charlie, but I suspect our host's key demographic is literate people with money.


Actually, there's another interesting point there.

One thing to need to keep a civilisation going is enough people and the right skills.

But (as your Ford supply-chain example makes clear) you also need to have momentum. You need all those people in the right places, doing the right things, to make the inputs that other people are going to need to get their job done.

Right now, that momentum and organisation exists, but it has grown organically. It's complicated and finely tuned and essentially undocumented, and the whole assembly only works because the flows of goods and services are moving through all of it all of the time.

If you wanted to start up another civilisation that did all of the same things, you need to get all those flows going, and I don't think that is an easy thing. It is beyond the scope of a planned economy, and it would take a very long time to re-grow organically.


@271 , probably true about the slary thing, but theyre basic needs are in a different currency.
convertible pesos are for luxuries. thats what the tips are for.
and the Dollar, they dont want any of it.


Yeah, that was the intention.

I thought that replying to a post that explicitly mentioned something along the lines of "I should think before putting foot in mouth" and replacing the "and" by "in" was sufficient to make that much clear.

Well, it wasn't. And it seems like I wasn't the first to invent that pun ...


I've often thought the same thing about cities and road management.

Designing a city to handle large traffic loads from the get go and getting things like traffic light phasing correctly for standard loads isn't something that's all that easy to go out of the box, as most of the systems we have a evolved.

That's also another problem in that they're probably less than optimal too.

But a lot of stuff around us isn't all that well designed and planned.


"The problem is that as we continue to advance, what is acceptable to let go becomes harder and harder. Imagine being asked to today to give up all surgical techniques since 1950. Not easy. Now ask whether people in the C23rd could live with surgical techniques of 1950..."

Well, we're living in a world where the wealthiest and most powerful country in the world finds it acceptable that a significant portion of its population doesn't any access to healthcare...


Well done Mister Stross. Very good question....

It depends upon the amount of written knowledge that the people have access to, in that case, pick a number at random, and have them surprise us, or better yet, for a better answer, find a population that is white, geographically isolated and uses population statistics from the 1950's - When most of the technologies of the modern world existed but before mass air travel started to affect the situation by permitting mass population movement - That is the source of your inability to resolve this situation.

The answer is 2 million people.
That is the population of New Zealand in that time period.

Predominantly white, agriculturally self sufficient, energy independent via geothermal power (A form of cheating but we'll mention it as you haven't restricted that.) and sufficiently geographically isolated that they essentially had to repair everything as importation of new-build technology was expensive.

Could they compete with modern technologies? - I can only cite the motorcycle engineer John Britten, (Designer of the Britten V1000) who is worth reading about, and thinking upon, as real proof of what you can do if you are far from anywhere and sufficiently motivated, so if they wanted to, probably.




I'm aware it's a technical question, my point is it has more than one answer. The fewer impediments there are between people trading ideas, the smaller your base population for maintaining the status quo can be.

Best case scenario to replicate a U.S. technology standard, for example, I'd say is about 300 million, i.e. roughly the population of the U.S. plus enough extra people to proxy for its imported wealth. You could get a similar back-of-the-envelope calculation for the EU by taking the difference between the EU's GNP and GDP and estimating how many more people you'd need to make up the difference from imports. Wealth isn't a perfect proxy for technology, but it's probably the best we have.


How many people, and what sort of resource commitment would it have taken to build a Shuttle in 1842? I don't know the answer to that one either...but at least that question is somewhat answerable in principle. for the one sparking the debate? Hmm.

First, start with a society in which people can meaningfully choose to move to Mars. Examine that society's technology & communications. Define "self-supporting" in a context they'd agree with, or at least be willing to entertain as a topic for conversation.

If you're looking for a useful answer with current tech levels and resources and political realities assumed to remain constant, and a planned society created and implemented by, well, fiat, in such a way that results actually follow specifications...uh, we're done here, and twice over. Won't work, and won't work.

This isn't even a question--it's an obfuscation of assumptions, leading nowhere.

Though the discussion is interesting, I admit.



Ghod knows how many it would take to maintain a current world, but I can give a historical example.

About 4000 Tasmanians were isolated for about 10k years when the water came up at the end of the last ice age.

About 700 were isolated on a smaller island between Tasmania and the Australian mainland. They died out. That brackets what it takes for stone age technology.

Current technology would take a heck of a lot higher number. Post nanotech probably a lot fewer.

Doubt this helps much.

Keith Henson


First time commenter, long time fan

Space colonization?

Purely for the sake of playing Devil's Advocate, I would suggest that we've missed the boat on that one.

It's the only answer to Fermi's Paradox that seems to make sense. Squishy life forms get one shot at it. If they haven't made it off world by the time they've made the kind of complete hash of their planet we (and, I suspect, pretty much any other life form in the Universe) have managed to do, before population pressure and competition for dwindling resources takes precedence over the level of technological effort required ... they ain't going to make it.

Maybe next time round.


"is the ability to handle liquid helium"

And actually, this points out another, much overlooked, aspect of high tech: Many of modern societies "trace elements" are by products of mass production.

If you scale down or eliminate the mass production, you have to start paying full price to mine those trace elements.

For instance, most of the Helium we use today was extracted from natural gas wells during the cold war, under a fat DoE grant.

Likewise, some very metalurgically important trace metals for alloys would have to be mined from natural ores which have much less than 10PPM content. Today they are "mined" in China, from flyash from coal burning power plants.

Therefore anybody contemplating building their own atomic reactors, had better get used to idea of the Chernobyl model, because they sure as hell will not have the metalurgy to build the safe western designs.

Ohh, and isn't there something about there only being one japanese steel-forge that can produce the reactor vessels, there were a nice story about them in one of the major US papers some years back.

And don't even get me started on what it takes to make sure your concrete is sound, ask about that in Finland instead...



It just occurred to me that everyone here assumes that a colony would possess less advanced technology than the mother-civilisation.

I'm not entirely sure that this is correct.

The Law of the handicap of the head start predicts that the colony will actually deploy more advanced technology than the mother-civilisation.

Due to this, the colony will also be more efficient, requiring less people to run a given city or factory in the colonial civilisation than a new factory or city founded in the mother civilisation.


This is not a discussion of space colonization.

Where did you get that idea from?

It's an attempt to look at the minimum level of complexity necessary to sustain and extend contemporary technological civilization.

(Where are all the space cadets leaking in from?)


While I'm not sure what the Kiwis being 'white' has got to do with it (in fact, they're not all Pakeha, some are Maori, and there's been a lot of intermarriage between the two since the 19th century), Mr. Mooney raises an interesting point.

What he forgets to mention is that the Aotearoa/New Zealand of the 1950s was heavily dependent on the 'Mother Country', i.e. Britain, for both its export markets (their lamb really is bloody good) and for technological imports (by the 1960s, the sheepfarming boom produced a situation where even small country towns had car showrooms offering the latest Jaguar models). That dependence was highlighted by what happened after the UK joined the EEC in the 1970s, locking Kiwistan out of its old markets, and sparking prolonged crisis.

In other words, the two million inhabitants of Aotearoa in the 1950s were part of a much larger community numbered in the tens of millions - hence their particular forms of tech in that era.


The space cadets come here because you're known as a science fiction writer and readers of science fiction often are at least part-time space cadets. Can't help that.

The other point is, it is easier to think of something growing up to something complex rather than thinking of something shrinking from complex to less complex, or equally complex but with fewer members - since it involves reducing entropy, which is a class of processes that require a) somewhat special circumstances to happen and b) an observer paying attention to them. (And the less special the circumstances required, the less attention-grabbing they are.)

What your question could be translated to is, which state that corresponds to the technological level of our society has the least entropy.

Question: What *exactly* do you mean by technological level?

If a definition can be found (which I started to doubt), we may find the answer by asking how many people we could make perfectly unemployed in the current system by means of any process that will end up at the same technological level that we started from (the level doesn't need to be constant!). Now, the system will have a lower level of entropy. After that, we can reduce entropy further, by reducing the population space through age/illness/genocide.
(A 1TB harddisk with 500GB of data has a higher entropy than a 1TB harddisk with 100GB of the same data in compressed form, which in turn has a higher entropy than a 100GB harddisk full of the same compressed data.)

Unlike a harddisk, in a society you'd have to repeat the process and we may find reasonable estimates for an upper bound.

But again, what is your definition of technological level and how do you define associated terms?


The more I read this thread, the more I think we may have missed the elephant in the room.

How do we define technological level? Is it what a society/country/civilization does, or is it what it knows how to do? Better yet, is it what it possesses the infrastructure to support?

For example, Canada does not possess nuclear weapons: but I very much doubt anyone believes that Canada couldn't build them if it wanted to. The same holds true for Germany and Japan. It might require some engineering research, but any one of those three countries - and a number of others - could do it without breaking a technological sweat.

Does that put nuclear weapons within their technological level or without?

That's ignoring regional disparities in tech level, too. Can a rural area limited to internet on dial-up really be considered at the same level as an urban area with not only high-speed broadband, but also high density 3G and Wi-Fi?

For example, I have an Android smartphone. If I'm in Washington DC, I have access to turn-by-turn GPS, augmented reality overlays, email, voice and text communication, and a whole host of other abilities. If I go about a hundred miles to Berkeley Springs West Virginia, I have no data service on my phone and all I have is mobile telephony. Has the technological level just dropped? Just to add to the confusion, someone on a different carrier may have all those things I lose when I go to Berkeley Springs.

I won't even mention the differences between Canada and the US.

Another question that leaps to mind is whether maintaining but not extending our current level of technology is even possible? We've already had the examples of Microsoft and NASA maintaining institutional knowledge. Taking things a step further, what if maintaining the infrastructure and institutional knowledge to sustain our current technological level requires continual innovation? If you can't have one without the other, then the numbers may have just crept up.

I know in John Dalmas' novels, his "standard technology," which sounds to be pretty much along the same idea as maintain but not extend, was only possible due to some very serious mental conditioning.

I'm beginning to think that the question may not be answerable as framed.



I fear the space cadets were here already. "Fifthly: space colonization? Get back to me when you've tracked down how many people it takes to design and build a space suit." was always going to lead them in one direction.


Nice approach.

Was New Zealand in the 50s on a knife edge, requiring every immigrant it could lay it's hands on to maintain the status quo? Or could it have managed with fewer people?

Were all the technologies it had necessary to the viability of the nation?


As has been mentioned elsewhere, you need to think about foreign trade. The question is, how many more people do you need to eat those lovely little lambikins, and how many more to support them, and...

Can we demonstrate that this isn't significant? Assume an average of one lamb dinner per person per week: how many lamb portions were exported at this time? Ignore if the number is less than a million pa (20,000 direct consumers) if we're looking for order of magnitude.

(What happened to the wool?)

If that shows that the number was in the 1m-10m range, all you need to do now is factor in the advances of the last fifty years and you're there. I leave this simple task to Charlie's other intelligent readers.


Here's a sub-question - how quickly do we want this civilisation to make new progress ?

Take, for example, the mobile phone industry, and all its penumbra of engineers and developers. Right now, it generates hundreds of new phones and many thousands of new applications every year.

If you cut the industry back to 5 or 10% of the people, they'd be able to maintain the current level of sophistication very nicely, but the rate of making of new stuff would drop like a stone. I assume the same is true in a lot of other high-tech industries.

Would that count as a success ?


I don't have anything of my own to add yet (still wading through all the other comments), but your post reminded me of a couple of ideas you might find interesting.

Stephen Salter (of duck fame) has a good essay about simplicity & technology called 'Simply Mistaken'. It can be found here:

At a very different level of technology, the modern trend of ever increasing tool use and population began about 45,000 years ago in western Eurasia, but 'modern' humans had been knocking around Africa for a long time before that, and appear to have tried and abandoned a tool-improving culture 45,000 years earlier. There is a theory that that early attempt was caused by a benign environment that supported a sufficiently dense population, and the advances were lost when changing climate reduced and dispersed that population. We then sat around getting nowhere for 45,000 years.

I think i heard this on the radio, but i've found a blog post poking holes in the idea:

Perhaps Ken Macleod's Romans on Mars will wake up to find they are being simulated by Africans on the way to Alpha Centuri?


If a society at the same tech level needs to be able to react to problems as well as us, then the 100 million number seems necessary, otherwise I just can't see why it's over a million, mostly in a city with few storms and coastal access, and a few colony towns at various precious resources. A lot of the factories would no longer be mass production, but more likely to create designs to order, but I don't think that there would be a major problem because of that.


One last thought. I suspect that somewhere in all these issues lies a problem of competency. How is it that some people gain expertise quickly, and become "old hands" at their new jobs in minimal time, while others, with equal intelligence and instruction, never become competent? Does it really take years to train someone to "expert" level, or is that an average that includes those who will never really become competent?

Regardless, I suspect that if you solve the problem of competence, your numbers for maintaining hi-tech civilization drop substantially.


You're looking at the mobile phone industry as a consumer. I'm afraid your model is incomplete.

First, you need to look at the infrastructure. Unlike the railways, we're still in the early days of building out the networks. They're upgrading from generation to generation, with 2-3 order of magnitude bandwidth improvements on each jump. The physics suggests there's some considerable way to go, but we need smaller cells to support higher bandwidth and more users.

Next, there's the back-haul. As folks in the developed world switch to smartphones, they want to download videos of cute kittens. This takes bandwidth by the double-handful, and the mobile phone networks weren't designed with that in mind. The best way to service them is to run high bandwidth cables out to the cell stations, but that in turn costs lots of money and involves ditch-digging.

Next, there's switching. See also "running an internet service provider 101".

Next, there's the developing world, where in many countries cellphone infrastructure is leapfrogging moribund POTS lines where the locals steal the copper wire if it's left unguarded, due to being dirt-poor.

This little lot amounts to about 90% of the cellphone industry. The plethora of handsets you pointed to is the chunk of the iceberg that sticks up above the ocean -- about 10% of it. And while you're right that most of the cellphones are just exercises in marketing-led wheel-spinning, we still don't have the ultimate handset, because we're still going through a period of maximum change.

I just got my iPhone 4 today. It's an upgrade from a 2 year old iPhone 3G, and believe me, the improvement in performance and utility is pretty damn clear. The iPhone 3G was in turn an upgrade from a detested stopgap, a Sony-Ericsson M600i, which in turn was an upgrade from a Palm Treo 650. I run my phones for about 2 years -- modulo hunting for a sweet spot between lineages, which led me down some dark detours: let's not talk about HTC and Windows Mobile -- and at each jump there's been a huge improvement in performance.

Once we stop seeing the performance jumps, then you can talk about downsizing the competitive R&D that's trying to find the next JesusPhone. And once everyone on the planet has got 1-20 gbps wireless broadband service to their smartphones, then we can think about cutting back down to maintenance levels for the industry ... around 10-20% of current investment, because electronics wears out. But until then? Nope.


Hm, "what advances would go furthest to reducing our modern societal footprint"? I like that question.

I think some sort of incredibly fast mechanism for training and retraining people would top my list. If I could develop expertise in a field and then record a "brain tape" that others could spool into their brains, that at *least* cuts down on training overhead. And if I can be an architect today, a bricklayer tomorrow, and a farmer the day after that...

I think next on my list would be some sort of incredibly efficient communication, something along the lines of telepathy. A smaller number of experts might be able to efficiently empower a larger number of non-experts via "telepathic consulting". But this has a huge risk if the communication method cannot be stored (like language can be stored in books), and if it *can* be stored that sort-of falls back to my pervious paragraph.

(Experts leveraging armies of "robot slaves" through telepresence just differs from that last one in degree, not fundamentally.)

I suppose safe, repeatable suspended animation might be another possibility... but that really changes the question into another question. With the right kind of suspended animation tech, you could basically have a huge "library" of people available to you, potentially reducing the number you need to have "hot" at any given time. Like, imagine if you could build NASA and then just freeze all the engineers, rather than keeping them employed constantly. Thaw out a subset on a rotating basis so they can keep some skills up-to-date, and then when a big project that needs 'em all turns up, thaw as many as you need. Similarly imagine if a drought reduced the amount of food you could output, but instead of having a starve-back and rioting and stuff, a large portion of the population just slept for 18 months and then thawed out.

I guess in a sense those are all just variations on a theme, that theme being: maybe you can do with a much smaller active population if you can sufficiently reduce the cost of "retooling" that population from one specialty to another on demand.


I think some sort of incredibly fast mechanism for training and retraining people would top my list. If I could develop expertise in a field and then record a "brain tape" that others could spool into their brains, that at *least* cuts down on training overhead. And if I can be an architect today, a bricklayer tomorrow, and a farmer the day after that...

You mean like Joe 90? :)

We'll probably have a robot society before that happens.


See also "Saturn's Children".

(It occurs to me that the minimum size of an interstellar colony for the robot folks in "Saturn's Children" is radically smaller than it would be for us unreconstructed Humanity 1.0 types ...)


or indeed like in the Matrix..
but no, it didnt work like that,bleh


"I guess in a sense those are all just variations on a theme, that theme being: maybe you can do with a much smaller active population if you can sufficiently reduce the cost of "retooling" that population from one specialty to another on demand."

But isn't this slightly dodging the question? You still need to establish the minimum number of specialisms and the number of specialists in each. Say there are a minimum of N specialisms and for simplicity we only need one practitioner of each (and ignoring succession - just looking at the instantaneous system), you could meet that need by N individuals. Or, yes, you could meet it by fewer than N individuals who can either learn new skills quickly or who have several to begin with. But you still need to estimate N to begin with.

(Now where's that spherical cow gone?)


I think, if you're talking about "self-sufficient" colonies, here's the best comment I've ever run into on the problem:

Our first challenge is to create an entire economic infrastructure, from top to bottom, out of whole cloth. No gradual evolution from previous economic systems is possible, because there is no previous economic system. Each interdependent piece must be materialized simultaneously and in perfect working order; otherwise the system will crash out before it ever gets off the ground.

From Sid Meier's game "Alpha Centauri"


The coordination is important. Even if we started with a huge library of knowledge and had a goal of getting technology from low level X to high level Y over Z generations, it is not going to work on some schedule of training and infrastructure development over generations without an intolerable level of control. Unpacking a "shake & bake" technological civilization is much harder still.

More than likely it will look like biological evolution with civilization taking a new path, possibly not even technological.

And yet, I still think it is possible. If we posited the US, Japan or Europe as being moved instantaneously to some other place ("Missile Gap"?), even with some disruption in resources it seems not unreasonable to believe that society would go on much as before, given energy wasn't limiting. The coordination function in this case is unneeded.

So could we posit some way to do this, recreate the old world in the new, either by transport or de novo, and then populate by reawakening our human travelers and having them carry on their lives?

Of the three entities, Japan is ~130M. Could S. Korea work, pop. ~ 49M? What about taking 1/5th of S. Korea only, ~10M, might that be enough?


There's another variable to consider: institutional momentum. As this blog reminded me today (, assuming that an institution == an instrument to accomplish a goal is simplistic. These things keep going on their own, and in the absence of active pruning, they grow; usually in ways that evince diminishing returns (bureaucracy at large) or even negative returns (hi, financial sector).

You aren't going to cap this discussion off with the Tainter solution to Fermi, are you?


The quote doesn't seem to match the comment you're replying to, Alexander.

Though I did make the point about AI people when I thought we were discussing colonies.


Exactly, the world of the mobile phone is going through a period of change, and we are spending a lot of engineer-hours on making that change. A few of those have been mine, so I'm not exactly seeing it as a consumer.

I'm not trying to say that that change is pointless, but that the rate of change is (depending on exactly which question you want to answer) optional. Are we looking for the number of people who can maintain the current level of sophistication, or are we looking for the larger number who are required to also maintain the current rate of change ?

After all, the current rate of change is not a given. We could get by with less, and we could use more. The same is probably true of all of the technical industries.


"Our first challenge is to create an entire economic infrastructure, from top to bottom, out of whole cloth. No gradual evolution from previous economic systems is possible, because there is no previous economic system. Each interdependent piece must be materialized simultaneously and in perfect working order; otherwise the system will crash out before it ever gets off the ground."

I've gotta disagree. "Farmer sells food to blacksmith, blacksmith sells repair services to farmer" is, in fact, an economy, albeit tiny and artificial. This assumes that we have planted farmers and a "blacksmith" somewhere and given them ploughs and a machine shop, and adequate supplies. The next step, adding (for example) miners - improves the economy - but it doesn't require insane amounts of additional complexity.

Replying to the next poster, this does not require tons of overwhelming social control, assuming that our person who "definitely will fulfill the social role of blacksmith" is in fact a blacksmith who has volunteered to go off-planet. Children of the first generation might have some angst, but I can think of worse places to grow up than a successful and growing colony...


The DMCA has been changed to allow "jailbreaking" for the iPhones and other smart phones.


The DMCA wasn't changed; there was, instead, an exemption granted for that. It will last three years, at which point someone will have to present it again. (Unlocking was granted an exemption three years ago, and was presented again and again exempted.)


The world is already an Idiocracy.

The vast majority of resources go to sustaining and controlling such people.

The vast majority of people across all races and nations contribute nothing of lasting value their entire life.

When I say nothing, I'm not judging one invention, book or song better or worse - I mean most people produce nothing but pollution and kids. A progress would happen faster without them.


People went extinct on Kangaroo Island too, which is a decent sized chunk of land (or is that the one you're talking about?). Significantly, they appear to have lost boat-building skills pretty early on, because although Kangaroo Island is too far from the mainland to swim, it's not THAT far.

Actually I'm surprised no-one has mentioned Easter Island yet.

My understanding is that the Tasmanian population went through a really severe population bottleneck, perhaps as bad as a handful of children and no living adults. That seems to be the best explanation for them losing the knowledge of how to make fire, amongst other important techniques that were nearly universal in stone-age cultures.


Charlie- here's the analogous question that helps to understand the issue: Would you support the complete political and economic union of the UK and US in the name of increased efficiency and economic prosperity?

I'm guessing no-- no matter how compelling the economic argument, the political skew to the right would be sufficient that you'd trade some prosperity to avoid it.

The question might be what level of wealth you're willing to accept to get the desired political state you prefer.


...does not require tons of overwhelming social control...

Depends on the desired goal. The level of technology an social structure being desired as the target, e.g. C23rd earth may not be the goal of our test civilization. They may wish to take a different path, perhaps even give up most of the technology. This may conflict with the plan to restore C23rd technological civilization. If we re-ran history from 1950, I would bet that many people would desire a different outcome than the one we are in now, and with knowledge of how technology did unfold, could change that development.

I'm not even going to argue that we don't have a lot of social control even now. :)


Why is the minimum size for robots less?

And what are the constraints? I'd have thought that the number of people you could plausibly send in upload (or true "AI" would be significantly more.


But do you need neurosurgeons at all?

Well.,. if you want to have a society where people do things like ride bikes, ski, play rugby or drive over 30 km/h the answer is yes, within 2 -- 4 hours of the accident site, with intensive care backup. Sore point where I live because the Ministry of Health is trying to shut down the service for 300 000 people, and I happen to be one of them.

But... onto the big issue. I'm a psychiatrist. Many people (eg. psychologists and nurse practitioners) reckon they can prescribe. Yes, I can do that -- and the only people who have as much knowledge around psychopharmacology tend to be pharmacists, BUT I can run a team, treat with Psychodynamic psychotherapy, cognitive therapy, interpersonal therapy competently, run a research project, critique a paper, write a meta analysis, set up a new service, audit same...

And I can (and do) train others to do my job.

And take the responsibility for the practice of everyone else in the treating team. And that responsibility is what you are paying me for. Or the neurosurgeon.

No doctor can work alone. We need teams. And (coming back to where this started) these teams are complex and have to be perpetually renewed. In the commonwealth, this is done as a series of apprenticeships (the medical model is the most elaborate, but nurses are mentored over the 4 -- 5 years it takes you after qualification to learn your job).

So... I agree with Charlie that you need some redundancy. I don't think you need 100 million to do this: I still think "Australasia alone" would (barely work).

And it is easier to write a scenario that gets to 20 mill over 40 years than 100 mill.



It depends on what you mean by "our current level of technological civilisation". A culture that accepted a Victorian level of sporting injury might be able to manage without neurosurgeons, at least for the scenario you describe, even if it had the smartest of smartphones.

If the level of preventable injury is one of your measures of civilisation, perhaps the automotive sector could stand some radical pruning/ elimination? ("But how do we get the uranium to market by bicycle?" whine whine)


If you accept a Victorian level of injury and death, you are no longer looking at a modern society. You're looking at a society where most people will have 5-10 children and expect half of them not to make it adulthood, where the life expectancy of the working classes (remember them?) is less than 50. The social and demographic effects of this spill over into almost everything: woman (and the working poor) don't get educated, etc, etc, etc. No thanks...


It depends on if your society accepts that people will take risks and has some methods of dealing with the consequences. Which is what happens in Modern Societies -- we let people take real, life threatening risks and then patch them up when they break.

But the big problem in world building is that acute health care requires a team of people and each person in that team needs to be training up both an apprentice and a journeyman. I suggest you look at the royal college of (insert specialty) medicine. There is a reason why the training is so long -- you need to do enough so that when things go wrong you have been there beforehand.

Things have to be complex for this kind of infrastructure. You can simplify -- run a Holden (Vauxhall senator) system rather than a BMW one.

So..(what I called the ANZAC model) 20 million, about current level of tech, some development, as bottom end 50 million gives you redundancy - which is rhe Aussies, Kiwis and Canucks.


But regardless of the No of people we seed the colony with we can provide certain facilities that are, to all intents and purposes, limitless from day one.
For example, communication: We can put a bunch of satellites in orbit to provide essentially free coms, plus whatever else is available (GPS, etc).
For example, power: By the time we're shipping the convicts to Mars we ought to have a reasonable grasp of some essentially free power infrastructure, be it wind/solar/nuclear, that can be shipped out and setup first providing essentially free power to the colonists.
For example, knowledge: Others have touched on this already, but the whole range of human knowledge would just be a phone call away.

Given the 'essentially free' things we can provide to the start-up, we can remove the need for many of the local skills, and similarly provide a strong foundation for the new empire ...


1. You can change your technology to substitute energy for complexity of mining technology. The real cost of metals would double, though.

2. Remember, the same technology that keeps people alive and nonproductive a few extra years is the exact same technology that keeps people alive and productive for a few extra years. You don't know which is going to be the case until after the operation.


Will: it's a work of fiction. This debate just reminded me that a side-effect of the setup in $WORK_OF_FICTION would make establishing an interstellar (read: no resupply from home in less than centuries) colony much easier. Remember: it's a work of fiction, i.e. amusing lies written for money. Don't take it too seriously (and pay no attention to the SF author scribbling notes for a sequel; we and ours have no place in that fictional universe).


I am not convinced that information is substitutable for labour in all (or even most) cases.


It is in all those cases where the information you're referring to is a matter of prognosis of the supply of labour from elsewhere.

If you don't know how many workers you'll be able to hire to do $important_stuff, you substitute this lack of information by simply doing the work yourself.

The result is that a whole lot more work is being done than strictly necessary. The difference is the value of information.


The large death-toll of children in Victorian industrial society was mainly due to "simple" diseases such as diptheria, scarlet fever etc. and occasional epidemics of cholera, typhoid and the like with the death toll enhanced by the debilitating effects of malnutrition and pollution. Industrial accidents and violence were a minor cause of childhood death especially in the 0-5 year old clade. I would presume that any colony would take a vaccine factory (or three) along with them or at least make strenuous efforts to select their initial population of colonists to eliminate carriers of these diseases. At the other end of the journey preventing malnutrition of children is going to be an on-going process as is limiting the pollution of air and water supplies.

The average lifespan of adults hasn't really improved that much over the past few hundred years; the number of people who died in infancy in previous generations dragged the average down but if someone survived into adulthood then they had a fair chance of making it to 60. They may have been worn out by unrelenting toil or disabled by accident or disease but they would still be alive.


As others have already pointed to, I think that a really important parameter to your question is the timeframe of the passage to a smaller population.

If you can plan for it well in advance, using all the resources of the current civilization to identify a core subset of our civilization, you can go down a lot, cutting out the wastes, the professions kept only because right now at some levels labor is cheaper than automation, compacting living spaces, and taking into consideration the natural need for less infrastructure to keep care of a reduced population.

Depending on the level of cuts you're willing to consider acceptable, and given tens of years to adapt for it, I think we could maintain tech equivalent to today but products much more expensive and much less options even with a population of 400-500.000. I could be off of an order of magnitude obviously, but I'm thinking about a medium city like the one I live near: net positive food production, university with all major departments and some small research centers, a tissue of hundreds of small industrial companies specialized on on-demand special equipments, an hospital that's able to take care of all the resident population for all but the most exotic contingencies, a small automotive industry concentrated on agricultural equipments, some small pharmaceutical and chemical, some electronics (as there's a material science department at university, I know they've also a small tirature chip manufacture)... obviously right now it's *not* independent, and for sure it could not be competitive with a full scale civilization, but given time and effort, I think it could be made to work.

Instead, if the timeframe for adaptation is sufficiently short, for example like in a disaster, epidemic, or time-relocation of a random selected group, I think you're even too optimistic: even losing in a strike 10-20% of the current total population I guess should be enough to drive economy in the ground and risk triggering a cascade of collapses... and I would not so easily exclude from my calculations third world countries, as right now our civilization depends a lot on them as a source of cheap labour, and cheap natural resources.

All this is obviously strictly imho, I doubt this question could be really settled without trying it out for real.


With respect to health-care, the 20/80 rule absolutely applies.

Get basic sanitation, the usual childhood vaccination-schedules and some broad-specter antibiotics in place, and you have a pretty healthy population, if you can keep them from the fries.

Most of the stuff we dump fortunes into these days, are attempts to cure ailments of old people which could have been prevented at practically no or even negative cost 40-50 years ago.

The final bit is to realize that humans are a renewable resource, and that it is better to replace them than repair them beyond some certain point.



@Robert Sneddon

You wrote "The average lifespan of adults hasn't really improved that much over the past few hundred years; the number of people who died in infancy in previous generations dragged the average down" and I have to disagree.

Regarding men, "they had a fair chance of making it to 60" if they managed to avoid infectious illnesses and famines, but that was a big if. Women added to those risks giving birth, which was dangerous indeed. Actually I read somewhere that statistics show the average medieval/renaissance marriage did last scarcely more than ten years, and relatively few lasted more than fifteen: by then at least one member of the couple would have died, her quite more often than him.


Figures from the US state that a white male newborn in 1900 had an average life expectancy of about 50-odd but a ten-year-old might expect to live to 60 plus. By comparison a newborn in 2004 has a life expectancy of 75 years and a ten-year-old 76 years.

By the 1900s pregnancy and birth wasn't as risky as it used to be, in part because of sanitation improvements, in part because of diet. Even poor women could have seven or eight children during their fertile period but the conditions prevailing then would mean they might expect to lose half of them in their first few years of life. The children of richer folks had a greater expectation of surviving to adulthood due to better food, medical care, sanitation and a healthier environment generally but they still suffered from levels of infant mortality that we would find astonishing today.


I agree with Isidro; the increase in life-expectancy for a person in early middle age (30-40) is around fifteen years, for much of which a fair proportion of the population remains productive. (Especially in clerical tasks; ability to perform heavy or highly dextrous labour tends to decline rather earlier.)

If you're 40, the change in life expectancy from 20 years (death at 60) to 35 years (death at 75, although these days it's nearer 80 in most high-tech nations) is a pretty significant one.

(And the health of a 40-year-old non-aristocrat (never mind a 70-year-old non-aristocrat) is likely to be much better than it would have been a hundred years ago. I agree that this is a separate issue, but in terms of making a self-sufficient society, it's definitely not a non-issue.)


I don't think this is the right question. Instead of asking, "How many people does it take to generate our current industrial base?" you should be asking, "What's the minimum set of technology that's needed to allow the current industrial base to be regenerated?" A related question might be, "What's the minimum set of technology that's needed to ensure a self-sustaining civilization?" but I'll assume that it's a subset of that necessary for full regeneration.

Once you know the answer to these questions, then you can work on figuring out how many people you need to accomplish it.

I too was a bit puzzled about your implication that complex civilization implying that hatred of big government was delusional. Big government is notoriously awful at responding to the needs of a complex system, while self-organizing solutions work great. This isn't to say that you don't have lots of infrastructure and regulatory things that you do need from government, but a well-understood--and bounded--set of these services seems far preferable to an unbounded response.


Read the last few comments and noted some good thinking taking place. But what is the ultimate "best practice?"

It seems to me that if you're planting a colony, (or planning for a controlled collapse,) the first thing you're going to do is create the best simulation money can buy. This is cheaper by huge orders of magnitude than even the simplest colonizing scenario. If possible, you release your simulation as open source software and encourage people to download it, perhaps even offering rewards for solving certain critical problems. One such reward might be a ticket to the colony/human preserve. If you build your software correctly, you can drop in the database for a new colony world and not make too many other changes to the software. You can run multiple scenarios with multiple different participants; i.e. people living in Maine have a scenario where there are multiple heavy winters in a row, people in Scottland get to figure out what to do when the local equivalent of Godzilla drops by.

As you actually colonize the world/human preserve, administrators have a whole range of possible scenarios with pre-existing simulations that can be used to predict outcomes. Not perfect by any means, but still very useful.

I'm not going to reread the whole thread, but acting from memory it seems we've identified an enormous number of "best practices," including the exclusion of sick people, good public health practices, (no tobacco or ice cream for the colonists...) leaving behind bad infrastructure decisions made in the past, leveraging computers and communications, setting up "fallback" points our society can drop to, a good library/university system, building stuff to high survival specs, eliminating jobs that could be automated or are currently reserved for numbskulls, (making sure not to eliminate the telephone sanitizers) eliminating redundant products, etc.,

All that being said, I don't think we've cut more than 50% off the necessary numbers to maintain a high-tech society, and that's mainly accomplished by eliminating redundant products. (We don't need both the Android and the Iphone in the first century or so of our colony's existence.) In the very best scenario we still have to transport a bare minimum of 2-3 million people to wherever we want them and pay to set them up.

That being said, remember that past colonies were not an attempt to set up high-tech societies (whatever that meant at the time) but instead were profit-making ventures to cheaply extract resources from other lands and take them back to the mother country. How that might work is entirely a different kettle of fish. To begin with, most big corporations couldn't care less if 70% of your children die or you're dressed in rags or you don't have a library available. If you have no way home and the world is otherwise awful, you're totally at the mercy of The Company, which leads to a very different scenario indeed...


Au contraire, corrupt government is really crap at responding to crises. I suspect this is why so many Americans seem to have a down on government; they've never seen a really good one in action.


Radicalmoderate - how many self organizing solutions automatically dissassemble when their purpose is done?


Au contraire, corrupt government is really crap at responding to crises. I suspect this is why so many Americans seem to have a down on government; they've never seen a really good one in action.

Speaking as a concerned American, you are so right - as far as it goes - and so wrong.

In the US, "we want a smaller government" is essentially code for "we gotta get more White Power so we can have Jim Crow again, and not have to worry about those scary brown people who steal our jobs and wimmen."

Or it's code for, "We don't want to follow health and safety regulations, because keeping rat entrails out of the hamburgers is expensive and our obscene profits aren't obscene enough."

All that said, you're 100 percent right, particularly about the dynamic involved. I've read some commentators who believe that Conservatives deliberately make government worse so people will hate it. (You may pound your head on the desk now. In fact, we can beat ourselves bloody together, just to make the pain of this idea go away...)

Corruption is much more a Liberal concern than a Conservative concern in the US, but neither side is willing to do hard work of rooting it out. Liberals tend to want more government, and conservatives want less government, but nobody wants to thoroughly purge allthe governmental and business elites who are on the take regardless of party affiliation.

American conservatives, particularly those who make less than $150,000 a year, are the most lied to and credulous people in the world. They've got their mental headsets plugged into Faux News and Rush Limbaugh 24/7 and they're as stupid, shortsighted, racist, and religiously delusional as you can possibly imagine. In any discussion of substance they can safely be ignored. Even if their heart is in the right place (and it seems so for the poster above,) they'll never do the work needed to get outside their own carefully constructed reservoir of false data.


American conservatives, particularly those who make less than $150,000 a year, are the most lied to and credulous people in the world.

It amuses me no end that on the odd times I find myself behind a vehicle with Conservative stickers on it (I live in a major city in the Pacific Northwest), it is usually, as it was today, a 15+ year old Buick, and not a nice shiny new car.

If the market works, it certainly seems to be mysterious in it's ways ;)


Ahem: this isn't a thread for discussing US domestic politics/ideology. Please bite your lips, or if you can't hold it in, go somewhere like Little Green Footballs or FireDogLake instead.

I'm on the road this weekend with an ipad, so I don't have time to ride herd and issue polite warnings: further political derailing will result in censorship (which is easier - fewer keystrokes).


That's why he said 'sporting injury', and later preventable injury. There's no need to accept victorian rates on everything. Just on dangerous (and rare) activities. I don't think it'd fundamentally change society if people having accidents whilst skiing will most likely die. It'd just discourage people from skiing, keeping net accident rates about the same.


Sorry, Charlie. Your initial comment re: corruption was completely brilliant, and I guess I got overexcited.


We should be able to gain some perspective on this problem by
looking at how much various actual industrialized nations
import. The more they import right now, the more their standard
of living would fall if they had to go it alone, since they
would have to find substitutes for what they used to trade for.

Starting at the top end, the US has 310 million people, a
GDP of $14.26 trillion, and imports of $1.563 trillion.
Imports are 11 percent of GDP. Just eye-balling it, closing an
11 percent import gap doesn't seem particularly daunting;
an isolated US could probably roughly maintain its living standard.

Stepping down a factor of three, to around 100 million people,
we have Japan (127 million people, GDP $4.137 trillion, imports
$499.7 billion (12% of GDP)) and Germany (82.3 million people,
GDP $2.811 trillion, imports $966.9 billion (34% of GDP)).
Japan is in about the same boat as the US, but Germany's situation
is rather different. If it were isolated, it would have to find
substitutes for a third of its economy, which seems daunting,
particularly since advanced economies tend to have between a
quarter and a third of their workforce in industry. An isolated
Germany would have to roughly double its industrial sector,
at the expense of services.

Stepping down another factor of three to around 30 million people,
we have Canada (33.7 million, GDP $1.285 trillion, imports
$327.2 billion (25% of GDP)) and Taiwan (23 million people,
GDP $717.7 billion, imports $172.7 billion (24% of GDP)).

Another factor of three gets us around 10 million people, with
Sweden (9 million people, GDP $333.5 billion, imports $120.5 billion
(36% of GDP)) and Belgium (10.4 million people, GDP $381 billion,
$253.1 billion (66%)).


This is not a lot of data points, but there seems to be some
sort of transition around the 100-million mark, where the
portion of imports shoots up from the tweens into the twenties
or thirties. If this corresponds to the boundary where trade
goes from an efficiency to a necessity, then it seems roughly
100 million people are needed to run a modern economy.


I would like to reply to a few comments:
1) The number of people initially needed for "The Colony" to keep up our level of civilization is dependent on the mean IQ of the population and it's distribution. With a mean of 100 you probably need your lower estimate of 100 millions. With a mean IQ of 120 you can lower it to 30 millions and with a mean of 140 you can lower it to 10 millions. An mean IQ of 160 would lower the limit to 1 million.
Other traits, like neuroticism, trust and so will be useful.

2) Depending on the size of the initial population you need to factor in the capital goods. Less people, more capital goods. The know-how, library, video, voice and text archives are part of this. But less people you will sens, more per capita goods will be need. Foodstuff for sure, but also tools, boxed factories, etc.
It is entirely different to drop a billion of dull people on a virgin world naked or drop one million of bright people fully equipped to last a decade or a century before the starting capital run out.

3) Just after securing the needed food sources for the initial population, and the basic resources, what the colony need is to increase the population as fast as possible. This is need to allow the economy to grow and allow to more specialization.
Foragers, hunter/gathers, were able to raise food enough for 7 other people (they were unable to store it and bring it along with them when forced to move). Farmers were able, in antiquity, to raise food for 30-300 people (preserving it was a different matter).
So, if the technology used to begin and the conservation of the food is not from the Roman Age but a bit later, the production is enough to allow for a large increase of the population in relatively few decades. Genetic testing for avoid low IQ and other bad traits in new-born (before embryo implant) would be useful to keep the population fit and fitter and avoid surely unproductive individuals to be born.
A population selected to be quiverfull like the Duggard would be useful for the first generations. Sorry they are Republicans, Evangelic and White.

4) There is no need to create money at the start of the colony. As the size of the economy grow, something like gold, salt, fur or butter will be used as money. It will happen without the need of some government busybody to decide what is "money". And please, if someone come up with some piece of paper backed by nothing apart the violence of the government, kill him with extreme prejudice (but without unneeded pain), this would save the new colony from many financial sector problems like the current one. Had the Various Lehman, Goldman and the Fanny Mae be force to only lend gold the had in their reserve, the boom and boost would not happen. Gold backed and full reserve bank system. And private money.

5) A free market would be a must (like in the day of the early colonist of America), as it force people to adapt to changes and reallocate the resources to the best use possible as determined by the buyers. The Pilgrims of the Mayflowers were collectivists, but it don't worked well and famine was the result. But when they switched to private farming the people with bad health started to work the farms for themselves and abundance of food followed and the colony flourished.

6) As a nurse, I can say that half of the cost of health care is spent during the last year of life. And, of the rest, 50% is due to 1-5% of the population.

7)Home costs are a huge part of the costs of living in developed society (25% of the life earning of common people). If the price of a home is the cost of building it, we can reduce the price paid to 5% of the life earnings at max (given the current lifespan and working years). Modern technologies could reduce the cost of building homes in this way. In "The Colony" land would be basically gratis (free).

8) The Arthur B. Robinson's Family have the children schooled at home at a cost of 300$/year/head and 15 minutes at day of help. All of them successfully completed their curriculum, went to hard science universities and earned their Ph.D.
Surely the genetics there helped a lot, but the method was successfully replicated by others less gifted. Given the cost of public schooling of a child (with barely adequate outcomes) is around 10.000$ (US) or 7500€ (Italy) at year, this allow massive saving. At 7-8 years old you can have barely educated foragers or Mozart; it depend on the quality of the people you start with and how well you nurture them.

9) The Survivors (Godwin novel) is a good story. They manage to survive in a very hostile habitat (with heavy losses initially) and keep enough know-how and informations to be able to free themselves from their prison. The population was forced to evolve at a fast speed and in a few generations developed near superhuman abilities.


# @ 360

Right and very wrong ....
1 Possibly - requires very careful handling
2 Agreed and obvious
3 end - no, quite unnecessary.
4 Codswallop. Gold has no intrinsic value, either. ALL money is fiat. Try reading Mr Pratchett.
5 Pass
6 ONLY in the USA. Does not apply outside USA. Here (UK) we throw money at children's care, ditto most of Europe, and we're healthier, overall. US-parochialism strikes again.
8 ditto. The US non-education system is so inefficient, because the rethuglican christians are deliberately sabotaging it, and simultaneously pointing out how bad it is, whilst trying to insist that their religious lies are taught in the same schools.
Grrrr .....
9 Pass....


I have recently come across some people making the claim (I don't know yet whether or not to give it any credibility) that the power elite are on a mission to eliminate a large part of the worlds population to make it easier for them to control the rest and have a bigger slice of the pie as well. For the reasons given in your article I find this hard to believe. If they "culled" too many people they would have to give up much of what they are used to. And the numbers they would need to leave intact woulld be much higher than yours also since they would not be abled to be as selective. All in all, I find such a conspiracy theory to be solidly in the realm of the tin foil hat crowd. In the real world it would not be able to work out very well for the ones controlling it.


Greg: I just banned painlord2k from this blog. (Reason: other postings of theirs -- moderated -- with racist content.) No need to feed the troll with your attention.


You need to consider exports as well as imports, because your existing data is not adequate to support your conclusion. Notably, you assume a dividing line whereby the US (a net importer[0] of goods) is able to be self-sufficient but Germany (a net exporter[1] of goods) is not. Why do you assume that Germany would be incapable of turning existing export industries round to provide the goods that are currently imported?

[0] in 2009, US imports exceeded US exports by ~500 billion USD
[1]in 2009, German exports exceeded German imports by ~200 billion USD.

(Of course, in a real-world context, the types of imported and exported goods/activity are significant too: US imports are mostly fuel and manufactured goods, while US exports are a mixture of food, manufactured goods and financial services. Very hard to maintain your technology level without external contact when you can't actually power everything without fuel imports! But this isn't strictly relevant to the question Charlie asked.)


Charlie @ 363

The boss has just read "painlord2k" and my response ....

She has grokked that the writer is not a English "native speaker", and is probably from E. Asia, and most likely Chinese in origin ....
Interesting, if only in passing?


>>> This is not a lot of data
>>> points, but there seems to be
>>> some sort of transition around
>>> the 100-million mark, where the
>>> portion of imports shoots up
>>> from the tweens into the
>>> twenties or thirties.

That's just an artifact of the particular data points you've looked at, namely countries with tight economic integration to a much larger economy (EU or US).

If you look at more isolated economies such as Britain (63M pop, 17% of GDP as imports) or Australia (22M pop, 16% of GDP as imports), you see that relatively low levels of imports are possible with populations far lower than 100M. Due to their relative physical isolation, I would argue that Australia is a much better data point than EU-embedded Germany or NAFTA-embedded Canada.

Even New Zealand and Norway (~4M pop) have imports of only ~25% of their GDP, suggesting that larger countries have higher levels due to convenience rather than necessity.


The type of constraint that you're discussing is somewhat beside the point to the type of autonomy we normal consider when thinking of space colonies. Normally we're talking about autonomy in the form of physical resources when we talk about colonization. Specifically, we talk about these things because physical resources are very expensive to ship around the solar system. But information resources, consulting, engineering, are different entirely. We wouldn't need many people on-site on a mars colony to have an entire planet of engineering expertise available to those colonists. Bandwidth, at least within the solar system, is a very manageable problem. All we need is big dishes on each end; just scale up the DSN, maybe put it in orbit. We could move things along even further with advanced robotics and teleoperation.

It's true, this isn't complete autonomy, but it is certainly enough autonomy to bootstrap a mars colony up to self-sufficiency at a reasonable cost.


A more fruitful way to think about the problem, and more importantly about the political implications that flow from it is to think about how thin your supply lines could be while maintaining modern civilization, and what the costs associated with making those supply lines thinner would be.

(The main problem with space exploration is not autonomy, but finding someone for whom there is an economic benefit to living there rather than Antarctica or the Sahara or a boat on the Pacific at a tiny fraction of the cost and difficulty.)

For example, most of the beer consumed in all of New Zealand is produced in one factor that employs twenty people per shift. Complete beer autonomy for New Zealand wouldn't increase the price of beer much, but would greatly reduce the available choices of high end beers, or require Kiwis to develop lots of micro-breweries of the kind common in Denver and Germany.

You can make those choices. Germany is committed to beer autonomy at the local level. Japan is committed to making milk locally. It costs more, but in a prosperous society, the greater quality may be worth the cost in something that isn't a big portion of the GDP anyway.

If you want to stop importing oil and you don't have a lot of it to start with, you have to make bigger adjustments. You need to create an electric or biofuel vehicle industry, an infrastructure to support the vehicles, public transportation infrastructure, and urban planning choices that encourage population density. For London or New York City or San Francisco, this may not be all that hard. For LA or Nebraska, this is a real problem.

There are ways to compensate. Car imports to New Zealand and Cuba are scarce, so the mechanics there are very adept at maintaining late model cars (road salt free roads that prevent rusting also helps).

If you are willing to fly anyone who needs a high end medical specialist to another country (as, e.g., Saudi Arabia does), you don't need to have so many local medical specialists. The line if flexible. Perhaps you decide that you need a home grown, in house cardiologist, but not an in house expert on pediatric endocrinology. Three contiguous U.S. states (Montana, Wyoming and Idaho) lack a single pediatric cardiologist and rely on Salt Lake City and Denver in the rare cases where one is needed.

Politically, it is less worrisome to rely on someone else for something that has only one foreign vendor or only a small number of foreign vendors, particularly if it is vital to your society. Your society may be able to endure a temporary shortage of gem quality diamonds, but not a temporary shortage of oil, so oil self-sufficiency may be a greater worry than diamond self-sufficiency. But, self-sufficiency in salt is hardly a worry, because even though it is necessary, many vendors can supply it.

Like anything there are marginal cost issues. Most communities can probably be much more self-sufficient at a pretty modest cost of living increase or cost in choice or in quality. Saudi Arabia wouldn't have to import nearly as much foreign labor as it does if its people were willing to do low caste but low skill work for themselves; but it has no choice but to import highly skilled labor that it can't produce locally. It could reduce its demand for foreign labor by 60% with only modest adjustments, but by 90% only with great hardship.

If the U.K. wants to stop having its own right hand drive automobile industry, it needs to rework its entire traffic infrastructure. But, it could probably end all imports of sweaters relatively painlessly, although sweaters would be more expensive as a result and there would be fewer choices.

Even a slight difference in openness at the margins makes the difference between a Cuba or Yugoslavia on one hand, and a North Korea or Soviet era Albania on the other.


"Standardize, e.g., software and machine tools to eliminate extra redundancy. For a concrete example, I'm sure a colony doesn't need 45 different Ethernet chip designs.

How about a related question? Using early 21st century technology and hindsight to design and prepare, what is the smallest self-sufficient colony capable of maintaining, not extending, early 19th century technological civilization?"

The hindsight point is a good one, although the more common practical impact is "leapfrog technological development."

Much of the world will never have an extensive land line telephone system because they started with cell phones. What country being developed today would put the money into developing a cable TV network when satellite dishes could be used instead? New houses aren't built with coal chutes and milk delivery doors. A society with reliable e-mail doesn't need the same kind of postal system.

Technology is path dependent. But for the Hindenberg disaster rural communities in places now served by expensive road and rail connections might have been built in roadless areas served by airships. But for Three Mile Island, Hawaii might have a nuclear power plant instead of an oil fueled system. If Apple had been more relaxed in licensing its operating system, Bill Gates would probably be just another entrepreneur who failed.

Much of the American West would have had much larger farms away from major rivers, and far fewer of them, and might even still be an open range, if a few more members of Congress on the relevant committees had understood at a gut level the water scarcity issues associated with farming there when they wrote the Homestead Acts. Those same acts are critical to understanding why American farmers live in isolated homesteads and have an individualist view of the world, while European farmers live in villages that were once home to feudal lords who share cropped out the land that the lord owned to them until there was land reform.

You also see examples of this in comparative law. No newly independent country that didn't have centuries of experience with English law adopted the common law legal tradition. Most of the world has copied wholesale the civil codes of France, Germany or Spain as they existed at the time of adoption, specifically because the instruction manual is smaller, and each of their civil codes were in turn based on the moral equivalent of bar exam prep course textbooks from Rome, for the same reason, to the detriment of their own indigenous legal systems (the fancy word for doing that is "reception" of the Roman legal tradition).

Simplification would be an issue too. If one opts for electric cars in a society (a good idea if your world lacks fossil fuels), you probably don't develop any kind of internal combustion engine industry at all, and you probably invest more in electric intercity rail. Clearing payments through postal accounts as most of the civil law world does, is a lot less involved than the system based on personal checks used in the United States.

Your second question is really the wellspring of Steampunk. The Zeitgeist cares about 19th century technology developed with 21st century insight, because the 20th century technologies based on oil seem doomed in the long run. Put another way, your question is, how will we maintain our society without oil and natural gas reserves.


@ 368
Minor (?) political point.
You said: "Even a slight difference in openness at the margins makes the difference between a Cuba or Yugoslavia on one hand, and a North Korea or Soviet era Albania on the other."

Erm, wrong order ....
Yugoslavia, Cuba, Albania N. Korea, in increasing order of control-and-oppression.
Tito was more concerned in keeping the lid on intra-ethnic tensions, and even after his death, it only came apart because Milosovic was determined to do so, for personal gain.
Cuba is still an oppressive dictatorship, in spite of the "left's" love-affair with it, as part of a gut unthinking anti-Americanism, Albania was quite mad, but "only" internally so - as soon as the Hoxja (sp?) went, it started to unravel ...
Then there is N. Korea - really scary. They are actually believeing (at least some of) their own propaganda, at the same time as starving their population.
I don't know which is scarier, really - an internal implosion, or a "limited" nuclear excange in that area.

I wonder if the John Brunner story "Who Steals my purse" could be used as a model to fix that problem?


That's one reason I was impressed with Stephen Baxter's Titan. The best way to colonize is to take the long view: send DNA.

I once suggested a SF story backdrop with the assumption that the colonists would suffer from technological regression, and the idea was to figure out social structures that would be appropriate. My particular suggestion was perhaps a bit unusual, but the idea of designing for regression is an interesting one.

As far as minimum numbers, I suggest 500 million. At about that world population in the 1500s things started to "take off" in terms of technology.

One should also consider a maximum population to keep the rate of technological progress at a level that society can properly adapt to it. I believe we are above such a population today, and the likely result will be Atwood's Oryx and Crake scenario.


For my day job I buy metals. I often get to visit factories and see the production facilities.I am quite often struck by the lack of bodies on the shop floor compared to what they would have been historially.

A good example is a Aluminium plate manufacturer in Birmingham, the plant was set up in the 1940's and employed over 2000 staff. Now with modern automation they employ a tenth of that number.

Taking this as an example, isnt the number of people required reducing, not increasing ?


Phil @ 372: I think the total number of individuals is decreasing, but the number of specialisms increasing. So you need fewer people to do most given things (particularly that involve physically producing something) but you have a long-tail of different specialisms that have all contributed something to the process.

The surprisingly few bodies thing holds good for major construction sites as well, incidentally.

Surprisingly few people on the ground digging holes and banging things with hammers if you wander around the building of a major structure. Very little activity, very carefully focussed.

However tucked away round the back somewhere will be a massive site office doing detailed design and project managing the build in intricate detail. You get really high productivity through good planning and very careful application of your resources.

Think back to a medieval cathedral construction site: the ratio of workers to managers/designers/planners would be totally different.


The UK doesn't have it's own right hand drive car industry, quite a lot of the world also drives on the left.

UK, Ireland, Australia, Japan, a lot of Africa etc...

It's actually quite easy to do at a factory level.


Damn this article. Now I cannot stop thinking about it. Many of the comments here make excellent points, but I think Charlie is still mostly correct. One thing that might change the nature of this debate is self sufficient technology. We are a long, long way from there, but what about self reliant technology? Technology that can not only maintain/repair itself, but also improve on itself. Maybe that is the next major evolutionary state, the emergence of self sustaining technology.


Well, I'm pretty far down the comment list by now, so commenting for my own sake ;-)


I agree that you need some large population in order to continue to expand technologically. However, I suspect that the level needed to *maintain* a technology is dependent on the relative intelligence of human substitutes, robots, or whatever. Look at a modern factory, for example. Far fewer people to run a robotized factory than it took to create it.

An alternative question for today's society might be thus: what is the *maximum* number of people a society can tolerate who cannot make to leap to the engineer/designer class, and for whom society has replaced by other means? In other words, is there a maximum number of people that can be tolerated by our current level of technology?



Why would a successful space colony - say, one on Mars - need to make significant contributions to our civilization? Why would it even need to be an essential part of the (trans-?)global economy?

It seems like we can make a similar argument against a remote oil rig in the North Atlantic. Woudln't we need a hundred million people to reproduce the basics of our civilization on that oil rig? Of course, we don't want to do such a thing. We want to colonize for the sakes of exploration and expanding human frontiers. And maybe for raw materials, if we can get them and ship them back home.

And what about the ultra-, ultra-, ultra-rich of the future looking for the most exotic exurb to build their homes? Or religious groups who want to start over on another planet and claim their new homeland? These people seem to be completely irrelevant to global capital flows, international trade, parts sourcing, the technoogical and social complexity behind building the craft that gets them to their new planet.


The *State of Israel* might have been set up by a few thousand people, at first (i.e., decades beforehand). There was already a population living there, with which to trade (and fight). And a whole Western civilization bringing goods and services into the ports (which had been made centuries ago).


If I consider Germanys imports and exports in details, it seems 80 million people might be enough. Germany is tightly integrated into a supernational economy, but the only good with a real import imbalance are raw materials (coal, petroleum, gas, ores), agricultural products and clothing.

On the other hand Germany exports a lot more machines and cars than it does import. With all labour and capital that is now devoted to producing good for export rerouted and an entire planet at their disposal (so no need to import raw materials and agricultural products), I suspect that 80 million people could keep our technology running.

The standard of living and the variety would be lower, but the "technology level" could probably be the same.


About this Entry

This page contains a single entry by Charlie Stross published on July 23, 2010 4:37 PM.

Five reasons to envy the French was the previous entry in this blog.

Mediocrity is the next entry in this blog.

Find recent content on the main index or look in the archives to find all content.

Search this blog