A robotic sewing machine could throw garment workers out of a job | Hacker News Reader

A robotic sewing machine could throw garment workers out of a job

economist.com

63 pointssvepuri71 comments

Outsourcing really just is the last step before automation.

The major trend to look for is always cost of production where labour being the number one cost.

In order for products to keep getting cheaper the cost has to come down. This trend doesn't stop just because we outsource it to China or Brazil or India and so even they become too expensive.

Unless economist start factoring in technology as one of the key economic factors rather than treating it as they do now as an externality we will still fight about whether too much government or too little government is going to save us all. And the politicians wont do anything because they don't see it projected in the forecasts their advisors give them.

In the meantime robots are going to slowly but surely put most of us out of a job in almost all the industries we know of right now.

Or put another way – it's the technology stupid.

The problem as always is the basic premise of capitalism: who reaps the benefits of productivity gains. It doesn't matter a fig if we automate away all of the labor in the world if all (or nearly all) of the gains go to the people who "own" the machines. Since 1970 this has been almost entirely the case - none of the productivity gains made by American workers have gone to the working class, they have gone to the rich,and mostly to the ultrarich. Why, then, should we cheer for any productivity gains?

This problem is purely political, not at all technological, and we can only solve it by clawing power away from capitalists. So, sure, robot sewing machines. But let's make sure we all get to eat the benefits.

Capitalism will always be in play even if robots take most normal jobs. As long as there are humans with needs and humans who can provide for that need.

So yes it is technology that is creating this winner takes all principle and it would be in a socialist society too. It's apolitical if you'd like.

The question is whether that's a bad thing. Sure, in the short term, it causes disruptions and job loss, but if we can steer ourselves towards a better future through it's use, then it may be a net win, if not for the current populace, than for its descendants.

The question is, can we steer ourselves, will we blindly stumble into this better tomorrow, or are we heading for a crash? The past gives clues based on prior outcomes, but only a fool would blindly expect models to scale indefinitely in the real world.

Agree, but the problem is that even short term is at least a decade and probably a major crisis before we will find a new solution to how to deal with that. And so a lot of people are going to find themselves without jobs without anyway to get a new one, my parents for instance and most probably myself at one point too.

This seems pretty much an obvious area for automation, seeing as how repetitive sewing can get...and that automation in the clothing/textile industry was the birthplace of one of the first disruptive technologies. Thousands of people lost their jobs, thousands of people gained jobs as clothing could now be made faster and easier than before. This just continues a trend that's been going on for centuries.

As perhaps the most relevant predecessor, look at the effect the cotton mills of Lancashire had on textile production in India.

This is highly relevant for the recently-bankrupt American Apparel, which has a fundamental proposition (even in its name) of producing clothing in the USA:

http://www.latimes.com/business/la-fi-american-apparel-manuf...

Wages in their current location will be rising from $9/hour to $15/hour over the next five years. It doesn't seem tenable in that situation to produce goods at the same level of technology that overseas workers use. So this may be the perfect opportunity to get serious about automation for clothing manufacturing. Send in the sewbots.

The Los Angeles Fashion District / Garment District is dying regardless of minimum wage hikes. A lot of business was from Mexico and China is going direct to Mexico these days. Also, a recent crackdown on money laundering aggravated the situation: http://www.latimes.com/local/crime/la-me-fashion-district-ra... . Unfortunately, minimum wage doesn't apply to many of these shops. They're mostly cash under the table jobs. I don't have a source for that other than knowing many, many seamstresses over the years who are now mostly out of work.

When I first learned American Apparel was made in the U.S. years ago, I really tried to like their products. In my case I wanted to buy their underware, and sports ware.

Every year I would feel that material, and think to myself, who buys their product. Their choice of material was always cheap feeling, and didn't fell good when put on.

I don't know why declared for bankruptcy, but quality of manufacture was the reason I didn't buy. A few years ago, I was astounded they were still in business.

I had no gripes to the quality of sewing; it was their fit, and grade of material that didn't get my dollar. Yea, I know a lot of you loved their hoodies. I have a bulky, polyester blend you can have in my closet. I might have used it while working under my truck, because I didn't care if it was ruined?

(In all honesty, the last time I looked at their clothing line was at least ten years ago. They might have cleaned up their act, but it was too late for me.)

Higher labor cost has to be accounted for elsewhere. In their case, with a high retail price and low quality materials.

I think material cost is also the main reason why they only use a few types of fabric. Putting in the largest possible order will have the largest possible discount.

That particular style of thin fabric is fashionable with, for lack of a better word, hipsters. The t-shirts don't feel as substantial as a nice, thick cotton T, but they wear and look like old vintage shirts. The ones that feel kinda cheap are interestingly more expensive.

Seems like a natural consequence of rising standards of living in the countries that this work was previously outsourced to. The garment industry has always exploited low-cost workers - from children, to women, immigrants, China, south-east Asia. When the cost of labor is no longer a significant input to the equation (because there is little to no actual human labor involved), I would expect to see production move back into the markets nearer to customers, to cut the cost of transportation.

And transportation costs are going to go down, as the distribution networks become more automated, so I suppose the next domino is going to be the cost of energy.

http://www.npr.org/sections/money/2015/05/21/408234543/will-...

^Some researchers gathered some data and tried to determine the chances of a job being automated by robots. It's a little scary but impressive how so many jobs that exist today can just be done by robots.

Interesting how they predict Programmers have a higher chance of automation than the other areas of engineering since programmers are the ones automating other things. I figure once programming is fully automated, pretty much everything else is automatable.

I guess they are actually talking about the actual translation of precise requirements to code.

Most of the programmer in western countries handle a lot more than the coding of things. If your job contains even a tiny bit of analysis, the probability drop significantly. Computer system analyst has a 0.6% chance of being automated according to this. Software Developer is 4.2%.

> translation of precise requirements

I think code can be recursively defined as "the translation of precise requirements to code". Seems like it's word games and everyone just wants to name the "last automatable thing" after what their current role is

I'm having trouble imagining any sufficiently specific and unambiguous format for writing requirements that does not look like programming. Even something like SQL, which only needs to describe database queries, is still the domain of the programmer.

Diagrams, other visual and high level declarative programming are all readily grasped by non programmers. The full impact of their "program's" behaviour may not be. But since when has there never been a junior dev position who's work couldn't be made better by knowing more. The very notion of junior dev implies they have more to learn.

While I appreciate the advances in technology to sew garments, I wonder how this will negatively impact the economies, which are thriving because of the garment business.

As good an argument for universal basic income as I've ever seen.

The fatal problem with any universal basic income scheme is that people will always vote for whatever politicians promise to augment their basic income, irrespective of any other factors, leading to ever higher strain on the economy and eventually bankrupting the country.

Basic income is actually pretty rare, even among democracies. So it would seem you are empirically wrong. Not only will people not vote to augment basic income, they won't even vote for a modest basic income.

Few politicians have seriously proposed it so far. If and when they do, the votes will follow (a testable prediction.)

This process has indeed happened already. It was a large part of the downfall of ancient Rome. Glossing over the many details, when slaves were doing most of the work, people supported those politicians who offered them ever increasing grain doles, until the state could no longer afford them and collapsed.

Socialism in general is a variant of the basic income scheme, and hugely popular: a state guarantee of a certain basic income and level of material security, no matter what happens to you in life. It has done well in elections for a long time and remains popular today, as well as driving various popular Communist revolutions on a non-democratic basis.

Won't happen. These are mostly poor countries. What'll happen is these people will be cast aside, and forgotten.

Basic income works in poor countries too: https://en.wikipedia.org/wiki/Basic_income_in_India

India has 1.23 Billion[0] people in it, and that wikipedia article is quite short, with only 4 sources. They don't cite any of the sources for the figure which stated 6460 people are in the program.

The program serves %0.000005 of India, or in much more familiar terms, if implemented at the same scale in America it would help slightly less people than there were killed by lions in Tanzania over the trailing 20 years[1].

[0]https://en.wikipedia.org/wiki/Demographics_of_India

[2]http://www.telegraph.co.uk/news/earth/wildlife/5149977/Top-1...

[3]Off by one errors.

It's a pilot program, not a nationwide initiative. That said, the economics of it would be positive if applied nationwide.

If you're interested in the pilot, it's documented very extensively in this book, which the Wikipedia article ought to be referencing: http://www.amazon.co.uk/Income-Renana-Jhabvala-Soumya-Standi...

Sure, to be clear I would like basic income. My point was that it is really easy to raise capital and test it on a scale like this. While I was making a joke, the implications are real. Inflation can rise if people are all still buying things at the same economic level and region. Where does the money come from?

The problem with basic income is that when the conditions exist to make it possible, it wont matter anymore. We need large transition to solar, automated farming, automated transportation and distribution networks, etc. Then the marginal cost of food, transportation, energy will be $0 so we can afford to give it away.

Economics is not always the zero sum game it is made out to be. Microlending could work, and getting these people access to conputers and internet will too.

To be fair, that seems to be much of how it functions in richer countries as well. It's not as bad if you're well educated, have other strengths and a bit of money laid up. But if you're a minimum wage factory operative or warehouse worker or one of the other relatively unskilled low-paid jobs, then the thing that you're good at vanishing can leave you in a very hard place; very limited funding to retrain and limited options for doing so even where such funding can be managed.

If they are desperate enough to steal or riot, they might not be forgotten. Is imprisoning a significant portion of the population cheaper than a basic income or at least free provision of basic services?

I think we will boil as frogs too slowly for there to be a dramatic escalation of violence, but things obviously need to change.

If we take away the `thing` that defines garment works (despite how unfulfilling we see the work), I hope that society does not leave these people behind to fend for themselves (and use the new efficiencies for earning money to re-train & re-tool these individuals.

The workers just really need to shift functions from craftwork to technical. Since the industrial revolution, that's the way it goes.

Unfortunately, the more automation, fewer and more skilled workers are needed. That's always been the case and will only continue to happen. People will be able to specialize and diversify more.

You mean these workers need to be trained to handle machines? Perhaps thats going to be a huge challenge for the middle-age workers, who would have lived all their lives doing the craft work. Survival of the fittest!

What alternative do you propose that doesn't either impose nonsensical costs on businesses or persist dangerous, backbreaking, and unnecessary practices for workers?

Diffusing into different industries?

Why does this remind me of the invention (and widespread adoption) of the Cotton Gin, which itself is partly blamed for the United States Civil War? The parallels to the Industrial Revolution are everywhere you look these days.

Why do we need to sew, anyways? Why not invent some sort of a glue (or a binding agent, to be more vague) to hold the pieces together? Something that'll last a 1000 washes and never wear.

Bonding agents are not the problem. The issue is working with cloth as it becomes a complex 3d shape. Picture adding pockets to a pant leg such that everything lines up to within 1cm. Now build it cheap enough to compete with someone making 250$ / month and working 80 hour weeks.

PS: Often a machine will cut the right shape and do the first seam, but after that it gets complex.

Patch pocket sewing has been automated for years.[1] For standard products like jeans, most of the operations are performed on an assembly line with special purpose machines. Direct labor content of the pair of jeans shown is 12 minutes and 15 seconds.[2] Even at $20/hr, that's only $4. Most of the human labor is putting parts into machines and taking them out.

[1] https://www.youtube.com/watch?v=oaOtX1btpUM [2] https://www.youtube.com/watch?v=5DzEpOp2Jy8

External Patch pockets are fairly simply, I am talking about the internal khaki pants pocket style.

We have one, invented thousands of years ago it involves cleverly inserting a string like solid binding agent repeatedly through the two pieces creating a flexible and strong bind.

But seriously, how many glues can you think of that are as flexible and durable as thread?

Some years ago, I actually had to learn about this. There are adhesives that are flexible, strong, hypoallergenic, and can survive laundering. They're used for disposable diapers, automotive applications, gloves, purses, luggage, and such. In a production environment they're applied as hot sprays, which requires compressed air and heated hoses and tanks.

Sewing isn't expensive, though. It costs more to get the fabric lined up properly for sewing (because that's manual) than to sew the seam.

(Former member, American Laundry and Dry-Cleaning Association.)

> But seriously, how many glues can you think of that are as flexible and durable as thread?

Would be a lucrative product for Dupont or 3M to manufacture, and they make some incredible adhesives.

It would indeed be lucrative and I'm sure they have at least a group or two researching exactly that.

Thinking of it now my comment missed the availability of iron on seam tape which is used for water proofing but as I understand it isn't strong enough to bond pieces together without stitching.

Why do we need a bonding agent when we can sew?

Sustainability takes 3 aspects into consideration: Economical, Social, and Environmental. Inventing a glue and making the garment last 1000 washes makes sense environmentally, but think about the negative impact it has on the economies. Potentially, it will increase the unemployment related problems in many developing nations.

You would need to balance this sentiment with the fact that poor people wear clothing and would appreciate the cost savings with longer lasting garments.

Thats a good point.

I wouldn't wish employment in the garment industry on anyone, besides; we've already destroyed the local garment/craft economies in many impoverished countries by sending our old clothes to them for free.

http://www.theroot.com/articles/politics/2009/03/donating_cl...

http://www.djiboutijones.com/2014/02/dont-send-your-used-sho...

> think about the negative impact it has on the economies

Building a product worse so it wears out and you'll have to buy another sooner is in no way a positive ethical choice.

It'd have to be something that:

1.) Is not water-soluble

2.) Is not soluble in any of the other common household cleaning agents.

I'd bet on some synthetic material where the seams could be welded in some fashion.

edit: wth kind of markdown/markup does HN support?

> markdown/markup does HN support?

Depends upon who you ask, but often the answer is 'as much as the market will bear' /s

At least partially because stitching is still a prominent feature of certain types of fashion.

Good. Working in a garment factory is not just mindnumbingly monotonous, but backbreaking and dangerous.

http://hesperian.org/wp-content/uploads/pdf/en_work_2013/en_...

Garment companies can make the working situations better by implementing sustainable work practices.

Sustainable Apparel Coalition is an organization working on this approach. Not working is not any better than working in a potentially dangerous place, which could turn better in the near future. http://apparelcoalition.org/

Do you think that all of these former garment workers will find new jobs in the robotics industry?

"Your job is shitty, so you are better off if you lose it", said the wealthy man to the poor man.

>Do you think that all of these former garment workers will find new jobs in the robotics industry?

An important question for both the garment workers abroad and for American workers. In the past, we took for granted that a redundant worker could find work again: one factory is like another. Now, we're not so sure.

What if garment workers were able to buy their own robotic sewing machines? Then we could have distributed labor and manufacturing and nobody would be out of a job.

Services like Uber will likely work this way when automation hits; invest in a car and put it to work.

In any case, I'll take one!

Infrastructure... as a service!

https://www.3dhubs.com/

Open Source manufacturing equipment:

http://openknit.org/build/

http://github.com/g3rard/OpenKnit

http://instructables.com/id/Building-the-Open-Knit-machine

  OpenKnit is an open-source, low cost, digital fabrication 
  tool that affords the user the opportunity to create their 
  own bespoke clothing from digital files. Starting from 
  the raw material, the yarn, and straight to its end use, 
  a sweater for example, in about an hour. Designing and 
  producing clothes digitally and wearing them can now  
  happen in the very same place, rewarding the user with 
  the ability to make decisions regarding creativity and 
  responsibility.

Thanks! I can't believe I hadn't seen this yet.

I actually have a knitting machine, which works similarly, but haven't used it much. Short programs are way more appealing than counting pins by hand.

For starters, if garment workers had the resources to purchase a sewing robot, they wouldn't be garment workers. You're going to have logistics issues on par with all farming being replaced with home gardens.

I think the more likely usage is amateur designers being able to produce their own designs. However, I'm pretty sure that's a very different demographic than the people currently working in garment factories.

Margins will shrink to the point where economies of scale prevail. If garment workers tried this, it would probably not take long before they get priced out of business.

Raw production cost only factors in slightly, there certainly is room and customer demand for boutique designs and customization options of such designs.

Today, people shop for clothes that incidentally both fit them and match their style. You'd have infinite options for both.

If people had access to capital, they'd be capitalists. And if my grandmother had wheels she'd be a wagon.

How could the average garment worker afford a sewing robot?

A conspiracy theorist might say there's more to this than meets the eye.

DARPA caring about the cost of uniforms seems silly. $1000? per year per person who costs a stack more than that to train and pay.

Where are the luddites when you need them?

As the Industrial Revolution began, workers naturally worried about being displaced by increasingly efficient machines. But the Luddites themselves “were totally fine with machines,” says Kevin Binfield, editor of the 2004 collection Writings of the Luddites. They confined their attacks to manufacturers who used machines in what they called “a fraudulent and deceitful manner” to get around standard labor practices. “They just wanted machines that made high-quality goods,” says Binfield, “and they wanted these machines to be run by workers who had gone through an apprenticeship and got paid decent wages. Those were their only concerns.”

Reply on news.ycombinator.com