The article comes from a study by the International Labor Organization.[1]<p>All this seems to come from one startup, Softwear Automation, which has just some lab demos and can't put a whole garment together.[2]<p>First of all, the textile industry has been mechanized for a long time. It was the first industry to mechanize, over 200 years ago. Everything from picking to weaving was mechanized in the developed world by 1940 or so. Cutting and sewing are now semi-automatic, and slowly going automatic.<p>If you've never seen a modern loom, here's one in action at a trade show.[3] Here's a working textile factory.[4] The looms pretty much run themselves. Note how few people are in the textile factory. Even third world countries have this technology. Low labor costs cannot beat a 1000RPM Toyota jet loom.<p>Sewing has advanced to the semi-automated stage. There are lots of machines where a human lines up the fabric and the machine does the rest.[5] The main remaining problem is handling and placing fabric automatically. The T-shirt and jeans manufacturers still haven't solved that problem, though they're getting close.<p>Production of simple shapes such as towels and sheets is almost totally automated.
Here's a factory in Bangladesh. Massive machines, few people.[6]<p>Summary: most of textile automation has already happened. The last step, sewing of complex garments, is the only part that hasn't been automated yet. Expect T-shirt and jeans production to go full auto in the next decade.<p>[1] <a href="http://www.ilo.org/public/english/dialogue/actemp/whatwedo/aseanpubs/report2016_r1_techn.htm" rel="nofollow">http://www.ilo.org/public/english/dialogue/actemp/whatwedo/a...</a>
[2] <a href="https://www.youtube.com/watch?v=BA96-WX-oXc#t=163.352979" rel="nofollow">https://www.youtube.com/watch?v=BA96-WX-oXc#t=163.352979</a>
[3] <a href="https://www.youtube.com/watch?v=Vl2rmup2dVY" rel="nofollow">https://www.youtube.com/watch?v=Vl2rmup2dVY</a>
[4] <a href="https://www.youtube.com/watch?v=SCcZW91Ub38" rel="nofollow">https://www.youtube.com/watch?v=SCcZW91Ub38</a>
[5] <a href="https://www.youtube.com/watch?v=aJw6oMwuJ94" rel="nofollow">https://www.youtube.com/watch?v=aJw6oMwuJ94</a>
[6] <a href="https://www.youtube.com/watch?v=RXiEU1Dhc9M" rel="nofollow">https://www.youtube.com/watch?v=RXiEU1Dhc9M</a>
"Who will benefit from automation?"<p>Easy answer the companies profits and the automation machine manufacturer and support companies.<p>Additionally the "Made in good old USA" sticker or wherever you happen to be brings up images of a local using timeless traditional methods making the garment, apart from a very few very expensive items that image is obviously a joke but it certainly works for sales and marketing to manipulate peoples perception and many will pay more for the sticker.
>savings of over $180,000 over 5 years<p>I don't know how thin the margins in this industry are, but this doesn't strike me as a substantial cost-saving move, especially if this is how much the US would collectively save.
This is a moral dichotomy that's been around for ages and I have yet to see a satisfactory answer on: are sweatshops evil because the workers are grossly underpaid and work in terrible conditions, or are they good because they're better than nothing and people will apparently starve to death otherwise?<p>Who should bear the responsibility of that decision? The workers? The governments of the workers? The governments of the consumers? Or the consumers?
> This isn't an entirely bad news story: the South Asian garment industry is dangerous and underpaid<p>Yet they are choosing it over the alternatives. unemployment, sex industry....<p>How about we let them chose their lives rather than somehow twisting that firing them is a good thing.<p>These robots will kill people, just in boring ways like increasing child hood mortality.<p>But it's going to happen, lets snap this band-aide off.
Can someone point to any actual sewing robots that would make this question not a hypothetical? I just searched Youtube and these videos came up:<p><a href="https://www.youtube.com/watch?v=-iudSnw--Y8" rel="nofollow">https://www.youtube.com/watch?v=-iudSnw--Y8</a><p>Automated rows of buttonholes and overlocking simple edges? Done in the late 1980s.<p><a href="https://www.youtube.com/watch?v=Uk3kNnHZdl0" rel="nofollow">https://www.youtube.com/watch?v=Uk3kNnHZdl0</a><p>How much slower can you go?<p><a href="https://www.youtube.com/watch?v=TBLEx-WaF9A" rel="nofollow">https://www.youtube.com/watch?v=TBLEx-WaF9A</a><p>Pocket sewing. Again, done in the late 1980s.<p><a href="https://www.youtube.com/watch?v=cSPb44jxA6M" rel="nofollow">https://www.youtube.com/watch?v=cSPb44jxA6M</a><p>Notice the video quality? Most of those machines use regular manufacturing automation techniques and seem to be designed in the early 1990s. They are not adaptable to different patterns/hems/collars/pockets.<p><a href="https://www.youtube.com/watch?v=1GUhlfjqQmQ" rel="nofollow">https://www.youtube.com/watch?v=1GUhlfjqQmQ</a><p>This one is actually not bad.<p>There was a lot of development in automated sewing in the 1980s that was abandoned as a result of American and European apparel manufacturing collapsing and all the work being outsourced to Asian countries and Mexico.<p>I have not seen any robots that can do overlock or coverstitch of shirts, dresses or tights at full speed. Until you have that, Asian apparel factories are not going anywhere.<p>It is still not clear how sophisticated and cheap cut and sew automation has to get in order to make it cost effective as a capital expenditure. For example textiles was already a very automated field in the 1980s but US companies found out the hard way that capital expenditures for increasing automation was not worth it: <a href="http://www.berkshirehathaway.com/letters/1985.html" rel="nofollow">http://www.berkshirehathaway.com/letters/1985.html</a><p>The one area of apparel manufacturing that does look like it will be fully automated (by definition) is complete garment knitting. The kinds of apparel it is useful for is sportswear, tights (thank you Lululemon for convincing women that tights are pants, sometimes it makes walking down the street that much more enjoyable), and certain kinds of dresses and sweaters.
Google deepmind and other such ai type software is going to make doctors, architects, lawyers, CFA, engineers, software engineers irrelevant as well. And I feel is a lot closer than robots that have human dexterity and can do all physical things humans can do.