Every so often a news item grabs my eyeballs and reminds me that I'm supposed to be an amateur futurologist, because of course SF is all about predicting the future (just like astronomy is all about building really big telescopes, and computer science is all about building really fast computers, and, and [insert ironic metaphor here]).
Via MetaFilter, I stumble across the latest development in 3D printing (now that 3D printed handguns have gone mainstream). Mad props go to another printing startup, although that's not what they're marketing themselves as: Fabrican ...
Fabrican is a unlikely-sounding spin-off of the Department of Chemical Engineering, at Imperial College (which in case you're not familiar with it is one of the top engineering/science colleges in the UK; formerly part of the University of London)—at least, it's unlikely until you begin thinking in terms of emulsions, colloids, and the physical chemistry of nanoscale objects. It's basically fabric in a spray can. Tiny fibres suspended in liquid are ejected through a fine nozzle and, as the supernatant evaporates, they adhere to one another. If at this point you're thinking The Jetsons and spray-on clothing, have a cigar: you've fallen for the obvious marketing angle, because if you're trying to market a new product and raise brand awareness among the public, what works better than photographs of serious-faced scientists with paint guns spray-painting hot-looking models with skin-tight instant leotards? (Note: the technical term for this sort of marketing gambit is, or really ought to be, bukake couture.)
The real marketing value pitch is less ambitious, and buried further down the page. Fabrican currently amounts to spray-on felt; a loose mat of unwoven fibres that adhere to one another and naturally entangle. This is brilliant if you're an auto manufacturer, who wants to do away with the laborious hand-fitting of carpets in your cars (just have the paint shop spray the carpet on the floor panels), or a furniture manufacturer who wants to soften the image of those cheap plastic chairs you sell for lecture theatres or buses and commuter rail.
But the implications go much further, because this is just step one. What we're looking at is the first sign of the shift to 3D printing of clothing (and no, Victoria's Secret doesn't count, other than for novelty value, any more than the Honeywell 316/Nieman Marcus Kitchen Computer of 1969 was a sign of the personal computer revolution to come).
Here's the thing: we live in an age of plenty when it comes to clothing—but it relies on a dirty little secret. Clothing has gotten much, much cheaper over the past century; if you ignore the brand premium on Levi's jeans (which have risen in price in real terms, due to going from cheap workware for manual labourers to premium brand name fashion item), a pair of workman's trousers today cost less than a quarter of the equivalent price in 1900. But this fall in prices is local to us, in the developed world. Fabric is woven on mechanical looms, as it has been for a couple of centuries, and garments are still largely cut and entirely sewn by human hands—the greatest enabler of increased productivity was the sewing machine in the 1850s (and, later, the overlocker/serger and other specialised industrial sewing devices). Our cheap clothes are made in sweatshops by underpaid developing world workers, and as Bangladeshi wages rise, the factories migrate to cheaper nations.
A side-effect of separating garment manufacture from consumers (us) is that they don't fit well, either. There are legends of Chinese clothing factories whose first batch of sized-for-western-girth produce has to be rejected by the buyers because nobody on the shop floor believed that the people they were making clothes for could be so fat. Nor do we, in general, have our cheap clothes adjusted to fit. While it's worthwhile to have an expensive suit or formal gown tailored, who would bother fitting a $10 tee-shirt or a $20 pair of jeans? Yes, we have easy access to cheap clothes at prices that make them all but disposable. But we also have cheap clothes that don't fit particularly well and fall apart rapidly.
So, where does spray-on fabric come into this?
We are used to wearing clothes made out of woven (or knitted, or crocheted) fabric—lengths of spun yarn that are interlaced in two dimensions to form a flexible mesh. The individual fibres in cotton or wool or linen or silk may be quite short, but when spun they adhere to each other and this allows us to create thread or yarn many orders of magnitude longer than a fibre.
Right now Fabrican's spray-on felt relies on very short fibres in a liquid carrier that form a matted felt when the solvent dries. (I infer that the strands are probably quite weak, individually, requiring the matting to provide some additional tensile strength.) But I'd like you to imagine the same technology refined so that instead of coming out of a spray-can it comes out of an ink jet printer nozzle. And I'd like you to imagine the same print head also having a different "ink" to print with—a waxy masking substance that can dissolve in an oily dry cleaning fluid and be washed out of the finished garment. Print alternate layers of fabric and mask and the layers of fabric won't adhere to one another. Dry clean after printing and you have separate layers. Give it ink jet printer resolution and you should be able to "print" woven fabric, complete with the warp and weft in situ (separated by the mask layer). The rest of this picture is about ten billion dollars and ten years' worth of fine tuning, and then luxury fibres (synthetic spider silk, anyone?): but the basic premise is that we are between 5 and 20 years away from being able to 3D print woven fabric.
What are the implications?
If you don't think printing woven fabric is a big deal, DARPA beg to differ; DARPA is pumping serious money into robot sewing machines. But automating garment assembly from traditional fabric components turns out to be a really hard problem (as this possibly-paywalled New Scientist article on a €23M project to build a sewbot explains). Cloth is slippery, changes shape if you drop it, wrinkles, and has to be stretched and twisted and folded as it is sewn. Note that final word: sewn. If you can print fabric in situ out of fibres in a liquid form, you don't need to sew components to shape—especially if you can print more than one type and colour of fibre at a time: you can fabricate your "stitches" (inter-layer connections) as part of the process, with minimal hand-finishing to possibly add fasteners (zips or buttons).
Add in a left-field extra: the rapid spread of millimeter wave scanners for airport security. These devices caused a bit of a to-do, earning them the nick-name "perv scanner" in some circles, because of their ability to see through clothing to the skin beneath, in order to check passengers for hidden contraband. But if you put the same machine in a clothes shop, it allows the establishment to obtain extremely accurate measurements of its customers without requiring a strip-tease and manual measurement of all the relevant saggy, lumpy bits and pieces. By use of surface-penetrating wavelengths (possibly high-intensity laser light, or infrared) it may also be possible to automatically distinguish between fatty tissue, musculature, and underlying bone structure. All of which are relevant to the construction of clothing.
So here's my picture of the chain store of the future. You go in, go to the scanning booth, and do the airport-equivalent thing in a variety of positions—stretch and bend as well as hands-up. You then look at the styles on display on the shop floor, pick out what you like, and see it as it will appear on your own body on an avatar on a computer screen. You buy it, and a machine in the back of the store (or an out-of-town lights out 24x7 robotic garment factory) begins to print it. Some time later—maybe minutes, maybe hours or a day or two—the outfit you ordered comes to you. And it fits perfectly, every time. Some items are probably still off-the-shelf (socks, hosiery, maybe even those cheap tee shirts), but anything major is printed, unless you can afford to go to the really high end and pay a human being to make it for you out of natural fibres. Oh, and the printed stuff doesn't have seams in places that chafe or bind.
Now, here's the down-side.
The fabrics on offer to start with will be fugly. Maybe not as bad as the bri-nylon shirts and terylene and other crappy synthetics of yesteryear, but it's still going to be fairly obvious (at first) what you're wearing. Figuring out how to make a sprayable matrix that uses cotton or silk or wool fibres has a multi-billion dollar pay-off at the end, so I expect it to happen eventually, but at first the stuff is going to look and feel like felted nylon. The styles on offer at first will also be fugly. I've spent a few years watching my spouse make her own clothing, and it's worth noting that dress patterns are complex and don't scale linearly: going from a size 12 to a size 18 isn't just a matter of blowing every dimension up by 50%. Clothes that are some variation of a simple tube or tubes will be easier than, say, a pair of jeans (with pockets and decorative seams) let alone an underwired bra or a sports jacket. Nor are there going to be many chain stores left to buy this stuff from. The job of a high street store in this scenario is to take measurements with a scanner and handle order fulfilment. Maybe also to act as a showroom. Today they have changing rooms and act as edge-of-network distribution centres. Tomorrow? Expect tumbleweed where the likes of Macy's or Primark have their bigger stores. Let alone T[J|K] Maxx—that business model is on the way out.
But back to the product itself. The first printed garmets aren't going to eat into the high end fashion market. Rather, they're going to displace sweatshop low-end produce. No, scratch that: initially this stuff is going to be something you spray on conference seats and car body panels (and maybe horrible 70s style flock wallpaper). But sooner or later it'll get good enough for really cheap, semi-disposable clothing. And then the pressure to improve the processes and recapture some of that $100Bn imported-from-China garment market will be irresistible.
So I expect 3D printed clothing will take time to catch on. But as it catches on, a lot of developing world factory workers are going to find their jobs are as obsolete as the half million men who used to work down British coal mines, or the million who worked in iron and steel foundries, or the other countless millions who used to pick crops and plough fields by hand and by horse. People who use sewing machines for a living will find their jobs have gone the same place as people who used to work in office typing pools with carbon paper and manual typewriters. Low status jobs, mostly women, with negligible social safety nets to catch them when they fall. On the other hand, this will hopefully be as much a thing of the past as this.
When garment manufacturing returns to the countries where they're consumed, the pace at which fashion trends turn over may actually accelerate: currently, there's a limit on how fast high street fashion can change imposed by the time it takes to send pattern blocks to a factory overseas, verify that the product is of satisfactory quality, then ship the loaded TEUs to market. It'll be like going from batch processing of punched cards on a mainframe in a computer bureau to using a time-sharing terminal: expect flash fashion trends to take off like a rocket once the tech gets cheap enough and good enough to fit the budget and taste of the vital high street 14-24 year old female demographic (and once the design software gets accessible enough).
It'll take a while longer (if ever—there are strength/durability/flexibility issues here) for 3D printing to revolutionize footwear (but, oh my aching feet, I can't wait).
The hand-sewn couture market (which still exists) will be joined by the not-as-high-end machine-sewn-by-real-people somewhat-more-durable market in the middle end. But it won't be a mass employer.
Now. What am I missing?