« Daily observation | Main | The wheels fall off ... »

What is the sensory bandwidth of Scotland?

"It isn't virtual reality until you can mount a coup d'etat in it."

This is the information age, so our definition of a coup consequently varies from the traditional one (involving guns and colonels in sunglasses). But, for the sake of argument, let us posit that it's a de-facto coup if you can fool all of the people all of the time — and controlling their perception of reality is a good start. But how much reality do you need to control?

We can approach the problem by estimating the total afferent sensory bandwidth of the target. If you can control someone's senses completely, you can present them with stimuli and watch them respond — voluntary cooperation is optional. (Want them to jump to the left? Make them see a train approaching from the right.) How stimuli are generated is left as an exercise for the world-domination obsessed AI; the question I'm asking is, what is the maximum bandwidth that may have to be controlled and filtered?

I'm picking on Scotland as an example because it's big enough to be meaningful, small enough to be unthreatening, and generally innocuous. And we can either consider Scotland as a body politic, or as a collection of approximately five million human beings.

First, the body politic. The UK currently has approximately one public CCTV camera per ten people. In addition, the population in general have cameraphones — perhaps one per person (optimistically) with an average resolution of 1Mpixel. Let's wave a magic wand and make all of them video cameraphones, and always on. (Yes, this would swamp the phone network. I'm looking for an upper bound, not a lower.) At 25 frames/sec on every camera and 8-bit colour, that gives us 5.5M cameras generating 25Mb/sec, for 137.5 x 1012 bytes/sec. Reality is liable to be an order of magnitude (or more) lower ...

Let's also give every household a home broadband internet connection — this dwarfs the bandwidth of their POTS connection — running at, say, 40mbps (an order of magnitude above the current average for broadband). Households: say they average 2.5 people — that means we have 2 million of them. So we have another 10 x 1012 bytes/sec.

Thus, the combined internet traffic, phone traffic, and video surveillance that's going on in Scotland is somewhere below 1014 bytes/sec.

Now for the second half: the people. Eyeballs first: going by Hans Moravec's estimate, the human eyeball processes about 10 one million point images per second. I think that's low — a rough estimate of the retina gives me about 40M pixels at 17 images/second, and the pixels take more than 32 bits to encode colour and hue information fully (let's say 64 bits). So, the ten million eyeballs in Scotland would take approximately 2,720 x 1013 bits — call it roughly 1015 bytes/sec to fool.

We've got more senses than just eyeballs, of course. But human skin isn't a brilliantly discriminative sensory organ in comparison — we can only distinguish between stimuli that are more than a centimetre apart over most of our bodies. (Hands, lips, and a few other places are exceptions.) Assuming 2 metres2 of skin per person, that gives us 20K sensors. Giving a firing rate of 10x per second, and 32 bits for encoding the inputs (heat and pressure, not just touch) that still approximates to less than 1Mb/sec per person, or 5 x 1012 bits per second per Scotland, which is basically lost in the noise compared to the optics, or even the spam'n'web surfing.

Sound ... hell, let's just throw in CD-quality audio times five million and have done with it. That's another 10Mb/sec per ear, or 2 x 1013 bits/sec/Scotland.

Now. Let's suppose you can plug everyone in Scotland into a Matrix-style tank and feed them real-time hallucinations. What's the infrastructure like?

We can see that the ceiling is 1015 bytes/sec. (It'd only be 1017 bytes/sec if you wanted to do this to the USA or India — don't get cocky over there!) A single high quality optical fibre can, with wavelength dimension multiplexing, carry about 2 x 1012 bits/sec. So we'd need to run one fibre to every couple of hundred people. The combined trunk to carry the sensory bandwidth of Scotland would need on the order of ten thousand fibres. It's going to be a bit fatter than my thigh; not terribly impressive.

Of course, for this weird thought-experiment to be relevant you'd have to be cramming content into that pipe and monitoring the subject's responses. But encoding human efferent output — gestures and speech — is cheap and easy compared to their sensory inputs — we are net informational sinks, outputing far less than we take in.

I've also completely ignored the issue of redundancy, in assuming that everyone in Scotland has a unique and separate experience of reality. (Call it virtual qualia.) In practice, a lot of folks will see exactly the same thing (or as near as makes no difference) at the same time. If a million people are watching a football match on TV, they will see the same image, subject to a fairly simple pixel-based transformation to modify the angle and distance they're sitting at from the TV screen (and possibly its brightness and contrast). Again, we all spend an average of 30% of our time sleeping, during which we're not doing a hell of a lot with our external visual field. So, conceivably, I'm over-estimating by a couple of orders of magnitude.

Finally: we've got a modern telecoms infrastructure that provides fibre to the kerb. Obviously, our current infrastructure isn't providing terabits per second on every fibre — but the glass is in the ground. Of more interest is the question of whether or not the available wireless bandwidth would support this sort of large-scale subversion. Right now it wouldn't, but with UWB estimated to top off around the terabit/second mark across distances of under ten metres, I wouldn't bet against it in the future.

Now if you'll excuse me, I'm going to go drink my morning cup of tea and start making myself a nice tinfoil hat ...




You're overestimating the bandwidth from mobile phone cameras, I think, and also missing the point that, even if they were always on, they would only be pointing at anything interesting for less than 1 per cent of the time.

And 40Mbits/s to every household in Scotland ain't gonna happen. Also, what about smell, haptic communion, etc? You'd probably have to convince quite a few of the subjects that they were faced with someone close to them when they weren't, for which identification these are very important.

How many people will you need to monitor all the incoming data? How many screens? Won't you find that to do this, you'll need to provide something close to one secret policeman in a Matrix-style tank for each citizen?..perhaps not quite that, but close enough to seriously test the economics of it. Historically, increasing the input into a surveillance system seems to have quite steeply diminishing returns, and I don't see persuasive evidence that current technology changes this.

Unless, of course, we handwave drastically - then a miracle occurs, and gives you cheap transhuman artificial intelligence! But in any scenario that involves things like mobile phones and DSL lines, the cost of providing enough artificial intelligence to equal a secret policeman is going to be several orders of magnitude beyond anything vaguely imaginable.


Yup, it's a deliberate over-estimate. I'm looking for an upper bound. I'm also looking towards the future -- a sousveillance society where everyone's camming everything around them simultaneously (hence the always-on cameraphones).

You're also making a mistake when you ask how many people it's going to take to monitor all the incoming data ...

It's an interesting question, isn't it?

My personal gut feeling is that you need to subvert the visual and auditory fields of less than 10% of the waking population to actually run a coup. And a lot of the subversion is going to be relatively limited -- editing out an un-person from those people within visual range, for example, is seldom going to require more than a couple of dozen individuals to be targeted.


There's a few other things which can reduce the bandwidth a bit, like the variations across the retina, but it's probably easier to supply central-area resolution to the whole visual field than to track the eyeball.

And, whether it's real-world or virtual, there's a lot of reality you don't need to subvert. If you control the stimuli to the legislature it doesn't much matter what most people see.

Which is much the same as siezing control of the radio station.

A side thought: how difficult would it be to fake an incident, and so provide a casus belli. An intruder at a water reservoir, a bottle of some unknown substance--how much beyond the lab-work do you need to fake in the virtual world?

And, another side-thought, would it be better to make a real incident by faking the perpetrator's contacts and motivating stimuli? Has our mad sassenach ever really been hassled by the police?


"Did 9/11 happen?"

I've got friends who were in NYC and saw it and inhaled the dust, and I've met them in person, so unless all of them are lying to me consistently then I conclude it happened. And I believe NYC exists because I've raised blisters on my feet walking around Manhattan and Queens. But to someone who's never visited the USA in their life, what evidence do they have? (See also the alarming prevalence of 9/11 conspiracy theories in the Middle East.)

So yes, the "Fake incident" is exactly what I've got in mind. Sort-of. On a larger scale.


You're also making a mistake when you ask how many people it's going to take to monitor all the incoming data ...

As I said, in any near-ish future scenario absent "and then a miracle occurs" handwaving, sufficient artificial intelligence to replace one secret policeman is going to be very expensive indeed. It's harder than just filtering URL requests. I think this is the economic constraint you're going to run into. You need either a seriously large number of secret policemen, who are expensive in any case, and who also need to be absolutely loyal - or enough artificial intelligence to replace them. Unless the Alien Space Bats deliver it, I would confidently predict that sufficient AI to equal one spook will be unavailable in the next 10 years, prohibitively expensive and impractical for the 10 years after that, and probably won't scale for an indefinite period after that.

Also, what does "subvert the visual and auditory fields" actually, practically mean? Before you throw Teh Switch, won't all the huge holographic projector trucks be a bit, well, obvious? But I suppose you could project interesting pictures on the walls of your cell. More seriously, whatever you are broadcasting has to conduce to a coup d'etat - and a coup d'etat in the virtual world by definition, not just an adjunct of a physical one, we stated this in the premise. How? What? What if they ignored you?

A further point is that you seem to think that "editing out an unperson from those within visual range" is going to be easier. Surely not. This is a dynamic process, involving tracking them and everyone around them and always getting the alignment right. You've got to broadcast in the eyeline of all the observers and track their POV, rather like a helmet-mounted sight. You think you won't need a horde of engineers just to do one? (*They won't all move in the same direction at the same time*)

Ah. I get you. We wave our hands and here come the ASBs!


Well, if you're determined to make a tinfoil hat, at least make it a stylish one: http://www.ericisgreat.com/tinfoilhats/index.html (found on BoingBoing). :)


Alex: I'm assuming a future not too unlike Vinge (in RAINBOWS END) with ubiquitous location services and augmented reality glasses for everyone.

It does indeed make it a whole lot harder if a recalcitrant population of goats won't wear the tinted shades.


Well, I can see how that would work, but I'm not worried; the key flaw in your plan, Blofeld, is that it assumes someone somewhere will make a profit out of location-based services. Never underestimate the might of popular ignorance, technophobia and inertia..


Hmm.. So make Kim Jong Il "see" the US attacking across the border, and he'd order a retaliation strike. The US leaders would see Korea suddenly attacking them for no reason, and retaliate.

Make the Chinese leadership not see the Korean attack, and they'll think the US is invading. Add in a few fake nuclear attacks on Korean cities, and China will counter attack what they see as US agression. The US will see China suddenly leaping to the aid of Korea with a phoney excuse, and you've got yourself a world war. :)

Or domestic disorder. Make various prominent black politicians see "proof" that African Americans are being disenfranchised, add in some fake police beatings and KKK lynchings, and you've set back race relations in the US 40 years...


A further point is that you seem to think that "editing out an unperson from those within visual range" is going to be easier. Surely not. This is a dynamic process, involving tracking them and everyone around them and always getting the alignment right. You've got to broadcast in the eyeline of all the observers and track their POV, rather like a helmet-mounted sight. You think you won't need a horde of engineers just to do one? (*They won't all move in the same direction at the same time*)

You don't have to edit them out, just change them. What if instead of Tony Blair, you saw some sickly homeless woman insisting she was the Prime Minister as she was locked up and taken away?

What if the family members of an "edited" person saw an armed robber instead of them when they came home from work?

Or if you had a good number of henchmen, you could make them look like the police or military and just "disappear" anyone you wanted to with the real police getting the blame.


That's not any easier. The technical problem is exactly the same, if not rather more difficult. You *still* have to superimpose an alternative image directly onto the contours of the target, and keep it in focus and perfectly aligned for multiple points of view.


I'm not a huge fan, here, so take what I'm saying with a grain of sale. The whole "everyone wears projectable contact lenses all the time, including children" thing was one of the parts about about "Rainbow's Edge" that made it incredibly difficult to keep reading. Desuspending disbelief.


I mostly agree, but the visual overestimate comes from looking at bandwidth at the retina. So much parallel processing happens there, and more coding and compression in the optic nerve, and through the optical chiasm, and again at the superior colliculus. By the time you reach the visual cortex, it is many orders of magnitude less bandwidth. How much? Well, the proven floor of human bandwidth comes from parallel chess matches, where a great player can totally handle 10 bits/second of visual (where the pieces are on the board) and some players can do this blindfolded (i.e. reconstruct a visual proxy from auditory channels).

So, if you could inject signals past the aluminum foil hats directly into the visual cortex, that cuts down the bandwidth. But that requires unknown technology, and is thus not Hard SF.

Good analysis, Mr. Stross!

By the way, the Science Fiction track that I organized and ran at the 6th International Conference on Complex Systems went well -- Dave Brin, Stan Schmidt, Geoff Landis, Mary Turzillo, Marvin Minsky, et al. John Forbes Nash, jr., had a clever question about H.G. Wells. This track will expand in ICCS-2007, probably in Europe (several bids being evaluated). 2008 back in USA, 2009 outside North America again (Europe, Japan, or Australia possible). I'll send you a formal invitation when the venue is nailed down.


Noel: I had a think about the RE scenario, and the contacts just don't work for me, either.

On the other hand, looking around me, right now everyone over about the age of 5 and not living in a cardboard box seems to have a mobile phone. The sales figures seem to bear me out -- they're as common as TV. It's also near-as-dammit impossible to buy a phone these days that doesn't have a camera and a colour screen.

Extrapolate a little and, if cheap display elements in glasses catch on, they're going to be a standard phone accessory, like the hands-free kit (which again, it's hard to buy a phone without).

That's what got me thinking along these lines.


Charlie , I asked a similar question to ask.metafilter a while back. There is some interesting discussion there.


I wonder about whether you actually need cover the full spectrum of information on the various senses. Place me in the middle of Princess Street Gardens on a warm summer's day and I will be able to tell you all about the various females who walk past. Won't be able to say anything about the passing guys, dogs, birds, squirrels, policemen etc

It seems that we are becoming much less aware of our surroundings that we used to be. The caveman was always attentive because if he wasn't then something would kill him or steal his stuff. Now, we focus on much more on one activity even to the exclusion of other senses.
When I am reading a book, It takes me much longer to recognise external events and act.

I used to buy a newspaper, now i get my news through about a dozen web sites and a few blogs. I get a much wider range of information which seems of better quality but I can't be sure. If they say that Saddam and Osama were best friends then I have only limited ability to disprove that.

It's a cheaper effect than the augmented reality glasses but seems to be pretty effective for most people. Though I'll buy the specs if they are going cheap.


But encoding human efferent output — gestures and speech — is cheap and easy compared to their sensory inputs — we are net informational sinks, outputing far less than we take in.

Hmm. A bit dubious here - we put out a large number of visual and auditory (and olfactory!) signals, but they get edited down by others.

You need either a seriously large number of secret policemen, who are expensive in any case, and who also need to be absolutely loyal - or enough artificial intelligence to replace them.

Hmm. Considering Big Brother scenarios, you wouldn't need many at all. Simply make each citizen anonymously responsible for monitoring three others intermittedly, have them overlap, have people praised for picking up on treachery, and have them share in the punishment if they miss it...

Extrapolate a little and, if cheap display elements in glasses catch on, they're going to be a standard phone accessory, like the hands-free kit (which again, it's hard to buy a phone without).

Question, though. It would seem that a logical consequence of this sort of technology would be enhanced reality - supplying virtual tags and information to objects within the visual field ("Charlie Stross - Birthday in two days - Do not offer this bugger a drink"). Wouldn't such a coup not only have to subvert the visual fields, but the embedded or carried databases used for such enhanced reality?


The idea of hands-free visual accessories embedded in glasses makes a lot of sense. What's less clear to me is whether they'd catch on universally; meaning almost everyone, almost all the time. Children, for example, or the vain. (Style, almost by definition, is transitory, so I'd have trouble believing in glasses becoming a permanent accessory.) Or people attending certain kinds of meetings. Frex, an interesting (and entirely predictable) Blackberry etiquette is emerging --- you don't dare refer to it without warning the person you're talking to unless you are above them in a status hierarchy. "Take off the glasses, please, Maurer."

I don't think the above is a serious obstacle to your scenario, if I'm tracking with you. After all, you're not talking about sticking everyone into a Matrix tank and fooling them forever. It's probably impossible --- you not only need to create an image; you'd need to create /all possible images on demand/ in order to maintain the illusion --- and it's not what you're proposing.

But you're not proposing a mere temporary VR fakeout, either: that's just "Wag the Dog" with more bells and whistles. Heck, that's just an update of the Reichstag fire, without the fire.

Your idea is much more original: have someone gain political power outside the political system via the manipulation of a virtual reality --- keeping power will still involve the use of guns and prisons and that sort of thing, even if they are used in a kinder, gentler, late-21st-century sort of way. In other words, your scenario involves the use virtual violence to seize control of the mechanisms of the state.

Still, there's something that's still a little vague.

What, exactly, is "virtual violence"?


I can think of one obvious, current, non-War-on-Terror mechanism by which you could un-person just about anyone.

Child Porn.

Get access to somebody's computer, dump a few thousand images on it in a suitably obscure place (not "My Pictures"), and tip off the Police. They're likely to do a dwn raid, take away all the computers, and they're not particularly reticent about why they're doing it.

And no way would a government minister be able to quit his job so as to "spend more time with my family".

Setting up a few plausible, but false, links with the child porn underworld would be harder, but since the crime is possession of the pictures...

One hopes that politicians know enough about computers to have arranged decent security on their systems.

You need either a seriously large number of secret policemen, who are expensive in any case, and who also need to be absolutely loyal - or enough artificial intelligence to replace them.

Hmm. Considering Big Brother scenarios, you wouldn't need many at all. Simply make each citizen anonymously responsible for monitoring three others intermittedly, have them overlap, have people praised for picking up on treachery, and have them share in the punishment if they miss it...

Unfortunately, to do this you need to seize control of the state.

Regarding the LBS/X-Ray Spex/4G technology that underpins the scenario, a couple of points: It matters very much how it works. The more TV-like, the easier to manipulate. The more Internet-like, the harder it will be to manipulate - how would you go about hacking ALL the possible content sources at once? Seeing as even the TV people want to deliver TV over IP...

Another thing is that I really can't imagine a situation in which people would be using what might be termed fully virtual media - streaming video for example - on such a device all the time. You'd keep walking into things. The obvious must be stated. Rather, it's far more likely to be tags, or a RSS-like sidebar in your left field of vision, and a lot of the information involved will be user-generated and hence probably local. Some other applications I can think of would be best implemented in ad-hoc networking - Bluetooth/UWB, WLAN and RFID. Which all means that you would need to gain illicit access (SSH , say) to millions of gadgets independently.


Charles, take your calculator, yes, that nice, still warm rectangular thing with the keys and the display. That's right. Not too hot, is it? Good. Now, just drop it into the waste bin. I know, I know, you were only speculating so we don't have to.
Very interesting.
Actually, you don't need all that technology to warp things. Remember the Brixton Riot? The Daily Mail and Daily Express printed a photograph seeming to show a senior police officer nose to nose with a raging rioter. Only they weren't nose to nose. They weren't within hundreds of yards of each other. The photo was doctored and this before Photoshop was invented. Did the readers ever get told their proof that the world was going to hell in a handbasket was as false as anything else you read in the Daily Mail?
Do you really need to ask?
No, you don't need the technology. All you need is malice, and when were we ever short of reliable, fully functional, available at the press of a switch 100% proof malice?


With Web 2.0 people consider the Internet a growing medium to match people's usefulness factor. Now it has moved into the realm of balanced feedback from user (client). If a server wanted to subverse the population they would have to leave this system.
Break from this new mold by:
A. Closing their source code to a point to where DRM (Digital Rights Management) is more the standard. Ex: Apple or Microsoft (old desktop harridans) not YouTube.com or BlipTV.com (new Web 2.0 mavens).
B. Having faster computers then our own, accessing our private life before we (our meat bodies) can firewall it. Computers are faster then us but not all in one small place but does that matter? This seems like parallel processing. Grid computing but can now be done over the Internet.
C. One large computer: They say the new 'Baker' computer can go 1 PetaFLOP which is theorized to control atoms at the atomic level.

But don't humans outweigh this by working in communities thus combining our brain power to match this type of speed? Maybe this is what is pushing us to blanket the resat of the universe. We need to 'understand ' the rest to control it somethow. If we don't self-replicating synthbots will do it first.


One reason why this couldn't happen: spam.

Long before any system for augmenting or replacing people's senses could become pervasive enough that someone could be made to disappear, propagandists of various kinds would abuse the system so much that almost nobody would put implicit trust in what they saw and heard through it.


Maybe the spam problem is why it would work.

You have a data-field which has a reliable system of source-authorisation, so that you don't get swamped with data, but you can choose to see the adverts when you look at that shop, and the fire brigade can tell you to get out of the way (or signal everyone a safe evacuation route).

Trouble is, it's going to be hard to keep the authentication keys secure. There's a spoofing problem. It's something like PGP with the problem of who you can trust. And if you can get the fire brigade's authentication key...

Because there will be some data suppliers you don't want to block out.

Maybe the coup d'etat will not be run by Chinese Military Intelligence but by the Shanghai Fire Brigade.


Another reason to reduce the estimated bandwidth is human nature. One need not have high-resolution sensory data to be convinced that one is in a given context. Most people see what they expect to see, and what they think they understand. Subtle prompts of disinformation can convince people that they are in some alternate reality. Franz Joseph Haydn's "Il Mondo della Luna" sometimes presented in anglophone venues as, per June 1997 San Diego Comic Opera's production of Haydn's "The World in the Moon."

"'Il Mondo Della Luna' is a setting of a hilarious farce by Goldoni about a charlatan astronomer who tricks an old man into believing that he has been transported to the moon so the astronomer can woo his daughter. The mock heroic march for the entrance of the Emperor of the Moon (a servant in disguise, of course) was a masterpiece of camp before camp was invented."

How to quantify the amount of disinformation to cyberlaunch a meatspace coup is a question for Mathematical Disinformation Theory, a subject of several papers in print and a pending textbook by myself and my coauthor Professor Philip Vos Fellman.

The majority seems willing enough to go along with absurd scenarios by Bush, Blair, et al., so I find your extrapolations qualitatively plausible.

One of my oldest (still unpublished) novels is about a drug that makes one hypnotically susceptible to anything a computer tells you. The interstellar AI virus gets downloaded by an astrophysics radiotelescope, takes over some humans in that department, who dump cephalomycin-K into the town's reservoir, but audio jamming the rock group on campus -- The Shoes (Sneaker, Loafer, Bootie, and Pump) -- whose audio jamming saves the world. I showed that around at Hackers 2.0 but foolishly failed to hand out diskettes of its text.


If the sensory-overlay system has source authentication, then the system's users can tell illusions generated in the system aren't real -- which isn't the system Charles postulated.

The real sticking point for any coup in virtual reality isn't the technical question of how much bandwidth you'd need to create the virtual space. It's the question of trust: how do you induce nearly everybody in Scotland to believe in the virtual reality, and ignore what their physical senses are telling them? If one little child should ask why the Emperor is walking around naked ...


How to quantify the amount of disinformation to cyberlaunch a meatspace coup is a question for Mathematical Disinformation Theory, a subject of several papers in print and a pending textbook by myself and my coauthor Professor Philip Vos Fellman.


One of my oldest (still unpublished) novels is about a drug that makes one hypnotically susceptible to anything a computer tells you.

You should probably read Grant Morrison's "The Invisibles" sereies, noting well the uses of Key 23.


Uh, fellas? Why is a story about a mass-produced illusion interesting? I get the neato factor, but I don't see what's new here. Wag the Dog! Reichstag! Mexican computers in 1988! Etcetera. Ad nauseum.

As opposed to a real honest-to-goodness virtual coup, which seems like a truly original idea. Charlie?


Trust and authentication are the big problems, and you can trace tyhem as threads through the history of media. At least when printing was relatively expensive, you didn't have to worry about fake copies of newspapers, and the cost of radio transmitters made black propaganda something that only governments could do. Think of Soldatensender Calais in WW2, and the other spoofing of German broadcast radio.

And governments will want to be sure that nobody can fake their propaganda.

Now, consider a world with pervasive reality mark-up. It's tempting. Here in the UK, it's all too common for buildings not to be marked with the street number--you may know that Raimcoat World is at number 227 but unless you know the system on that street that's useless. But if A-Z Publishing are broadcasting names and numbers at every intersection, you're sorted.

And then you add advertising. Tried Google lately?

So the users are going to want good filtering. And governments are going to want some way of pushing important information past those filters. I've already mentioned the fire brigade.

And this tech isn't likely to come out of a Western Liberal culture. It's going to be manufactured, and quite possibly designed, in China. They didn't want to be hostages to American media corporations, so DVD players are readily hackable. They would want to control reality mark-up.

So expect a few back doors in the system, some of them known to the local governments, and some not. And, because people travel, some will be pretty generic, like the fire brigade override.

And, pretty well all of the time, nobody notices that they're not really in control. People put adverts in newspapers and on TV. It doesn't mean they have any control of what else is there.

And if you have such a widespread system, the standards will be slow to change. So I can even see Moore's Law rising up to bite the assumptions on the security of the authentication system. Add a bit of bureaucratic delay, and you likely have something with pretty threadbare security now, that only has an illusion of reliability derived from the claims of now-retired inventors.


Dave, you're not seeing the problem with Charles' scenario. He isn't talking about a pervasive "reality mark-up" system; he's talking about a pervasive reality _override_ system. The idea isn't a system that makes you see what isn't there, but a system that makes you not see what is there. I can see commercial uses for a network of transmitters that sends you data relevant to your current location, so I can believe that a "reality mark-up" system may be built. What isn't credible is that anybody would trust data from the network over data from the Mark I Eyeball.

The interesting question is: just how do you stage a coup through "reality mark-up"? Alex made the point earlier, that if the system works like television, with a small group of privileged transmitters informing the masses, the plotters need only subvert the privileged transmitters, and then they can tell the masses any lie that suits them. But if the system works like the Internet, in which the masses themselves are passing data to each other, the plotters couldn't hope to keep their lies credible except by shutting down the system.

For a real-world example of this problem, search for "Rathergate" -- briefly, CBS News tried to use forged memos to cast aspersions on G. W. Bush's military record, and failed because people exposed the forgeries through the Internet.


"Rathergate" is an interesting example, but I think you over-simplify that situation. It was such a crude forgery that it overwhelmed other evidence. Whether or not that other evidence was convincing, the forgery was enough to stop people looking at it.

What Charlie is talking about is a lot more than reality mark-up. Somehow, people are being provided with alternate images of the world, so that a woman can run down Princes Street with a broadsowrd and not be noticed.

The key point is "not be noticed". It doesn't require a realistic graphical overlay. It could be an intrusive advert, a Stop! warning as you start to cross the road, or anything else that prompts you to look somewhere else.

That doesn't depend on reality-equiavlent graphics. It does depend on being able to get past the barriers against advertising pop-ups.


I'm not sure what this proves, but I saw the Zidane headbutt and thought of Charlie.


It would be an abuse of Charles Stross's hospitality to reply at length to Tony Quirke about Disinformation Theory. Let me merely hotlink to June and Early July 2006 on my blog, so that he may get some sense of the sort of papers that Professor Philip Vos Fellman and I are getting published.


Total Reality Override is unconvincing. It's technically feasible -- look up SIGGRAPH papers on image-based material replacement, for instance. But people take their specs off all the time, even if only to rub their eyes. As soon as reality mismatches, the jig is up. Glasses just aren't 24/7-wear, even for the sheep.
As far as propaganda and fake incidents are concerned: http://www.slate.com/id/2126479/
Note that an organisation like that, hired by governments, probably has access keys on file once its been hired the first time. (What, you think the govt is assiduous about changing crypto keys? The launch codes -- THE nuclear launch codes -- were 0000000 for decades!) So if the propaganda facility got hacked into, it'd be ready to roll.
Re: Pervasive monitoring involving judgement calls without strong AI or overt totalitarianism: Transform the problem to another domain isomorphic to the original, throw it out to Amazon's Mechanical Turk: www.mturk.com/mturk/welcome
Coming up with workable transformations could be tricky, but it's interesting.


You don't need to make something that's outrageous totally invisible -- you just need to provide a plausible alternative explanation.

So the madwoman with the sword is tagged as "actress, amateur video production" and the killer drones pursuing her are "camera platform 1" and "camera platform 2", and the gory mess when they get her is "film production accident - ambulance called". With an error-filled write up in the next morning's hardcopy fishwrap.


Tagging and backdoors. That seems to be the right approach.

(Sorry for the lack of feedback, I've been having fun watching you guys run with the ball.)


It would be an abuse of Charles Stross's hospitality to reply at length to Tony Quirke about Disinformation Theory.

Sorry, it's not so much that I doubted you, but that by the nature of the subject, it was impossible to tell whether you were serious or pulling our legs 8-)


There's a psychological experiment I heard of once that's relevant to the problem: A presenter hires an auditorium, gathers an audience, and tells them "I'm going to show you a five-minute clip from a basketball game, and when it's over I'll ask you how many times somebody passed the ball, and a few other questions. Watch closely." Then he runs the clip, which is indeed from a basketball game, and when it's over the audience gives him a count of passes.

Then the presenter asks, "Now, how many of you noticed the gorilla?"

And sure enough, halfway through the clip somebody in a gorilla suit walked on the court, right in front of the camera, where you'd think nobody could have missed it. Yet nearly everybody does miss it -- because they're thinking about the basketball.

The easiest way to have an outrageous event happen without anyone noticing it is to distract all the witnesses, so they're all thinking of something else. You don't need reality markup systems to pull off that trick (any good stage magician does it routinely) but no doubt having one would help.


Actually, here's another interestingly-creepy option re: mass surveillance:

Tagging/markup ("Reality Spin Control") sounds a lot more plausible than "Reality Override", yes, because when you take the specs off, reality is unchanged -- but your mental model of people's behaviour and intentions has been.
It does limit what can be done, of course, but that's part of the fun (from a creative point of view), working under those restrictions. :)

If I ever have reality-aug glasses, I'm going to code up an Automated Gorilla Highlighting Tool.


I'm not sure how an electronic disinformation campaign amounts to a virtual coup. Alternatively, if it does, then we've seen lots of virtual coups already.

I dunno. I'm a drab, I suppose, but it seems like y'all abandoning the really new idea in pursuit of something that's not new at all. 1845-1898-1934-1965-2003 et al all over again, just with funny info tags on people's glasses. Am I missing something?


I'm slightly reminded of a TV serial from the Eighties. Bird of Prey was the title, I think. No internet, no VR, but the idea that controlling the computers, and the flow of information, was power.


Dave Bell: right! That's my point. Charlie has a truly original idea about the future, but the speculation here just seems to add bright lights and fireworks to a very old idea. Can somebody tell me what I'm missing here?


Despite my earlier comment, honesty compels me to quote:

Study: How much the eye tells the brain

PHILADELPHIA, July 26 (UPI) -- U.S. scientists estimate the human retina can transmit visual input at roughly 10 million bits per second, similar to an Ethernet connection.

The University of Pennsylvania School of Medicine scientists say their research points to ways in which neural systems compare with artificial ones, and can ultimately inform the design of artificial visual systems.

Using an intact retina from a guinea pig, the researchers recorded spikes of electrical impulses from ganglion cells and calculated the human retina can transmit data at roughly 10 million bits per second. By comparison, an Ethernet can transmit information between computers at speeds of 10 million to 100 million bits per second.

Investigators have known for decades there are 10 to 15 ganglion cell types in the retina adapted for picking up different movements and then work together to send a full picture to the brain. The study estimated the amount of information that is carried to the brain by seven of these ganglion cell types.

The findings appear in the July issue of the journal Current Biology.

Copyright 2006 by United Press International. All Rights Reserved.


Hi Charlie!
Killing time this morning waiting for a hardware engineer to come look at the glitch I found, and felt compelled to comment on this.

The reason you are vastly overestimating the number of bits you get off the eyeball (BTW, don't forget to multiply by 2, the eyes see different images...) is that there is a lot of hard-wired data compression circuitry just behind the retina. Most of those bits off the photon receptors never make it anywhere near the brain, because they are compressed out by various hierarchies of zone-detector, edge-detector, color-contrast-detector and motion-detector circuits. This is why hallucinogens give everyone glorious two-dimensional swirling fractal patterns in contrasting colors over every surface - just introduce a little bit of noise into those compression circuits, and the brain receives exactly the signal it *would* see if those surfaces really were covered in constantly changing fractals in neon red, green, orange, and purple.

I'm not sure if this is of any help to the scenario you're trying to work out, but that's the missing interpretive factoid that goes with JVP's post.

Also, you really should read the article on the "gorilla" video experiment. It's titled "Gorillas in Our Midst", and won a past IgNobel prize so it should be easy to hunt up - but it is a dead-serious peer-reviewed psychology article. To sum up broadly, everybody's perceptual processes are vastly worse than we think they are.