New t-shirt sewing robot can make as many shirts per hour as 17 factory workers | Hacker News Reader

New t-shirt sewing robot can make as many shirts per hour as 17 factory workers

qz.com

129 pointshourislate90 comments

$0.33 per shirt, anywhere in the world, is nearly as cheap as the cheapest sweatshop markets.

I expect most commentary will look at this using a US vs China thing, but I'm more worried about SE Asia and Africa. They need to climb up the development ladder that China's already 1/2-way up.

Say what you will about clothing and sneaker sweatshops, but they were a tried and tested method for building up the basics of an industrial economy. South Korea, Taiwan, and Hong Kong wouldn't have gotten to where they are today as quickly without them. How will Kenya develop?

What do you think will be the industry that shifts from China to SE Asia and Africa? I feel like clothing is the likeliest because of it's relative simplicity, compared to other things.

Not sure how accurate.

I think clothing has already moved away from China. I have shirts made in Bangladesh. And according to [1], workers get paid $0.12 per shirt.

Like you said, $0.33 vs $0.12 is super close it seems.

  [1]: http://www.ecouterre.com/infographic-how-much-does-that-14-t-shirt-really-cost/

The article says the cost in Bangladesh is 22 cents.

That is surely close enough, after taking into account transport and lead time, that factories in Bangladesh should be worried.

how much will the human labor cost per shirt be once these automatic t-shirt machines are moved to Bangladesh?

Robot production lines need robot technicians to support them. I doubt you'll find many of those in Bangladesh any time soon.

eventually though. there are, no doubt, some very skillful, very resourceful sewing machine repair techs there.

i guess that time window is how long the automatic t-shirt making business remains viable in the US.

IIRC most of the world's hard drives are made in Thailand, not too far away. They found them somewhere. :)

And one robot replaces 17 workers, it's probably around 1 robot tech to support 10-20 machines.

According to these stats [1], in Bangladesh there's 4800 factories with around 725 workers on average. So that would be 40 machines with one worker each to supervise and 3-4 robot techs. 5 people to replace 725!

[1] http://www.waronwant.org/sweatshops-bangladesh

Why will the machines need to be located in Bangladesh? Cheaper repairmen?

i was just thinking like some business owner might: let's try to lower costs in every possible way, including labor.

The owner will be balancing those costs against the cost of shipping and extra logistics overhead for goods destined to North America or Europe.

That issue becomes a logistics issue. You can pay one worker in Bangladesh to oversee the machine, or you can pay one worker in the US (or Europe) to oversee the machine.

But, if your target market is US (or Europe) you've got shipping costs to get your finished shirts to the US (or European) market. If the machine is in the US (or Europe) those shipping costs are substantially lower (and with much less latency).

The lower latency (meaning you can sell shirts in a more "on-demand" manner) and much lower shipping costs might well outweigh the labor savings of the Bangladesh worker overseeing the machine.

I remember reading an article on shipping, and they claimed shipping costs for a t-shirt from China to the US are something around $0.02.

I'm thinking the biggest advantage to outsourcing t-shirts is in a lot of places you can dump used dyes into the local river instead of dealing with rules on toxic waste.

I assume shipping the raw materials (wool and cotton) to Bangladesh would cost more than the shirt on the return trip.

Depends on the price of the machine. You don't want to place expensive factories in unstable locations.

China / Mexico are generally considered safe bets compared to much of the world and also have cheap labor.

I don't see this as bad. Being able to manufacture simple things just increases how complex things can be for a certain cost.

CNC and automated process control isn't killing Asian metalworking.

Economies need to meet a certain level of development in order to "move up the value chain" rather than just get left behind. I'm confident China has that. I'd be worried about Indonesia or Kenya.

Tech is cheap and ubiquitous. Perhaps people in SE Asia and Africa might want to invest in a T-shirt sewing robot and an Internet connection such that they can run absolute circles around every major developed-world company by keeping their overhead low. Those multi-hundred-million-dollar CEOs from those American (or Chinese) companies are a tremendous drain. And that's not even to mention their absurd gigantic office complexes which aggressively destroy value at every turn...

That requires infrastructure. Here's the thing though: if you can do it dirt cheap overseas and cut out labour, then you can do it dirt cheap pretty much anywhere.

Offshoring really gets done to reduce costs via cheaper wages. If you take out the wages, then offshoring doesn't look as good as it once was.

The issue is that white collar work is now the stuff that gets offshored. But then I see the absolute dire service many outsourced companies provide and I often wonder how much the business loses because of it.

Lots of stuff are made in China due to proximity to resources, not just cheap labor. T-shirts don't qualify, though - the US is actually the largest cotton producer.

In that case, Australia should soon have a large manufacturing base.

You could say that about basically every labour-saving device though. Tractors, earth-moving equipment, container ships, whatever. And even textiles are almost wholly automated now - lots of textiles are grown or spun, woven, dyed and printed almost entirely by automated machines - robots, in effect. Final assembly is in many cases the first time a given piece of textile has been touched by human hands. So why is the automation of this final step going to be the thing that prevents countries from climbing the "development ladder"?

When cellphones became affordable, it didn't mean that countries without decent landline networks couldn't "move up the ladder". They just jumped straight to the more advanced technology. If higher-productivity technologies are available, and are cheaper than using scarce human labour, then you use it. More productivity is not somehow a bad thing.

I'm not convinced this point carries. Sure, "lump of labor" is a fallacy, more productivity is not likely to drive massive unemployment. But that doesn't mean all labor is interchangeably easy in all environments. Competition is still possible.

Early tractors didn't do much for underdeveloped nations. Mainframe computers didn't either. The default state of capital-intensive technology is to improve productivity, but not to bring capital and work into underdeveloped nations. Whatever benefits accrue there are the diffuse rewards of expanding global production.

The economic boom in China, Taiwan, etc was something vastly more specific than "rising labor productivity". Container ships, telecommunications, (Western labor laws,) and efficient but labor-intensive production came together to make offshoring incredibly profitable. The boom wasn't just rising productivity, it was a confluence of incentives that made production in certain countries much more attractive than others. We didn't just make more things, we closed factories in the West and reopened them in China.

If the next wave of productivity gains is high-capital, high-tech, low-labor, we'll probably see the reverse happen. People will engage in production where they can get extremely stable infrastructure (including internet), skilled labor, and precision parts. Factories in countries with cheap labor but unreliable power and internet might well close in favor of more production in countries like South Korea or even America. That won't be the end of the world, productivity will still rise, but it might end the specific forces that have caused so much growth via mass production.

> Early tractors didn't do much for underdeveloped nations. Mainframe computers didn't either. The default state of capital-intensive technology is to improve productivity, but not to bring capital and work into underdeveloped nations. Whatever benefits accrue there are the diffuse rewards of expanding global production.

When early tractors were being deployed, all regions and countries were quite undeveloped in agricultural mechanization, compared to today. Likewise, when early computers were being created, all countries had primitive infomation technology. The default state of capital intensive technology is to be deployed first where capital is relatively cheap and labour is relatively expensive. The "diffuse rewards" of expanding global production are pretty much the whole point, since the magnitude of those rewards are huge.

The recent economic boom of much of the less-industrialized world was characterized by (compared to the industrialized world) labour-intensive production, true - however, it was much less labor intensive than had been the case before. It was absolutely characterized by people adopting more mechanization, more automation, more technology. So this is why it doesn't make sense to me to think that a high-productivity technology will stall growth. If a t-shirt robot is more expensive than a Bangladeshi textile worker's wage, then we'll see continued high use of human labour in that industry. If the robot becomes cheaper, then the Bangladeshi textile workers will shift to other industries (as they mostly did when they went from being farmers and handcraft workers to being textile workers, over the last few decades.)

> So why is the automation of this final step going to be the thing that prevents countries from climbing the "development ladder"?

I believe you're misharacterising the issue. The trouble with technology is that well, you either have it or you don't. Labour on the other hand is fungible. So with a sweatshop you can go from unskilled workforce to $$$t-shirt$$$ with base supplies and bog standard infrastructure.

Acquiring, operating and conducting maintenance on precision robotics on the other hand...

> I believe you're misharacterising the issue. The trouble with technology is that well, you either have it or you don't.

Technology is always a matter of degree. Let's say I have a handloom barn making fabric and clothing. Is that already "technology"? If not, then what if I install a diesel generator to power the looms? Then I get some treadle sewing machines to replace hand-stitching? Then I get a laptop to replace hand-accounting? Do I "have technology" yet? As a handloom barn owner, I certainly couldn't have developed or manufactured a diesel generator, a sewing machine, or a laptop, yet I can buy and maintain (or just replace) them quite easily. Doing the same with robots is just a matter of degree and difficulty. Compared with getting humans to do the same work, is it worth it or not?

I think it is a matter of competition. The unpowered loom does not enable you to compete, but some amount of powering it and accepting a lower wage does allow you to compete with shops that have slightly more tech. Accepting lower wages is how these shops came into existence in the first place.

This is the first production quality version of this robot the next one could be twice as fast. How do you compete with that?

The only way to compete is to build another machine that is cheaper or better. This is not something Bangladesh is able to do.

The point is that if the robot really is that fast, and that cheap, then you don't compete - you start doing something else where the robot cannot compete at that price. Textile workers aren't robots - they have the capacity to shift to other work, and there's an enormous amount of work to be done.

This objection to automation has been brought up every time anything's been invented, for hundreds of years - if we do something with less human labour, then humans won't have work or incomes. But actually, they just end up finding other useful things to do that weren't getting done before.

I don't think this is the same argument that has been made hundreds of times. That argument is about the west. Here in the USA we could in theory train our coal miners to write software or build build bridges or do some other useful task.

China has way more people and much less ability to train them, but that is not 0 ability to train them. But Bangladesh... They are pretty fucked. This will cut them out of a huge swath of the economy they were participating in yesterday. You can't efficiently make roboticists in Bangladesh today.

EDIT - I also think complete automation is fundamentally different than performance enhancing tools. Clearly Bangladesh has sewing machines, but that is an entirely different price and profit category than a full blown sewing robot.

This is exactly the same argument that has been made hundreds of times. "Oh, they're simple peasants now and forever, they can't work in factories, they'll never learn to read, they can't invent productive industries" and so on.

You seem to be implying that training for jobs is somehow something that's done mainly by the governments of countries - that people are just passive receptors of whatever training and education the state can provide, and can't seek it out, or create it for themselves. This is of course wholly wrong, and almost none of the specialized skills developed by people in the course of industrialization are gained through formal education.

You don't have to make lots of roboticists in Bangaldesh anyways (and there aren't many of those in industrialized economies either) - people can and will move into the sectors that have been small simply because labour was more valuable elsewhere - services, like health care, education, retail, et cetera. These are useful tasks no less than textile work, and largely don't require a massive "training program".

I'm all for rhetorical points, but surely you realise the magnitude of difference in complexity of operation-maintenance between 19th century looms and motors versus non-analog assets.

Not only there's a major difference in degree of difficulty as you note but also the question of support ecosystem.

I disagree. Literally billions of people, many of them living on $4 to $10 dollars a day, without indoor running water or reliable electricity, have cellphones - compact computers connecting to a massive global telecommunications infrastructure. Computing capacities (and hence, productivities) which would have had their own floors of buildings and dedicated staff in the 1950s are now semi-disposable consumer goods. Because technology got better.

How do they operate them? Well, through the relatively simple, user-friendly interfaces. How do they maintain them? Well, they're modular and inexpensive - if a single component, like a battery, breaks, you replace it. If too many components break, you just get another one. Of course this does require a certain level of local support and expertise, but no, I absolutely don't think it's that much more difficult than maintaining 19th C. industrial assets, especially because of how fast and affordable modern telecommunications and transportation are. Is it broken? Call an expert on another continent, or have one fly in. Do you need new parts? Already ordered, and they'll be here next week, for a low freight charge. Sure beats waiting for the steamship to come in, or the telegraph boy to bring a message from head office reading "Could not replicate the engine malfunction, please try halting and re-starting the steam boiler."

> [...] semi-disposable consumer goods. Because technology got better.

In my humble opinion this is the flawed assumption in the line of thought you advance. Technology isn't universally disposable, that depends on the replacability and support structure around any specific source of complexity (edit: and purchasing power)

You haven't solved much of anything with automation if the final word is "don't worry how it works, we'll come over and bill you to fix it"

> You haven't solved much of anything with automation if the final word is "don't worry how it works, we'll come over and bill you to fix it"

Odd, because that's how almost all "automation" in everyday life works. How many people actually understand the operation of, or could fix, their dishwashers? Their cars? Even their coffee-makers? Yet dishes are washed, cars are driven and coffee is made, mostly uneventfully.

Is industrial technology different? A little, but not that much - if it doesn't work, there's someone out there who can fix it, for a cost. If communications are too difficult, you might keep them in-house and pay them a wage. If not, maybe they work for another company, and you pay them for their time. Or just buy a new machine when the old one stops working.

I recognise the truth in what you say. We both apparently live in areas where we can afford to ring up the repair guy for fixing or replacemement.

Why you assume that is universally true is beyond me though. How do you expect societies that are incapable of adequately maintaining roads or developing safe architecture to house their population to sprout the mentality and savoir-faire for complex operations?

For sure my argument is not universal either, but I believe you transpose properties you are familiar with to a world you are not.

> More productivity is not somehow a bad thing.

I didn't say it was bad! I'm pro-productivity. My contention is simpler; that going from a pre-industrial economy to an advanced economy has a path that's been well trod by the West and now by east Asia too.

There's a give-and-take, or zig-zag path, between automating the old jobs and finding new jobs for people to do. You need to automate farming to free up farm labor to do more productive things, but you also have to have a large demand for jobs to absorb all those ex-farm laborers. Otherwise you just get urban slums full of the unemployed.

India tried to jump directly from pre-industrial economy to a financial and technological services economy, and it's not working as well as China's model. India's GDP/capita is still a fraction of China's. The problem is that you can't take a billion subsistence farmers and turn them into the 10 million specialties needed to run a modern economy. Growing human capital is an organic process that takes decades and requires intermediate steps of sophistication.

And for centuries now, garment production was a good stepping stone to more advanced kinds of manufacturing. It's like if you take the middle three rungs out of a ladder, how do you get to the top? You can't. The top is still there, but you're stuck at the bottom.

Now maybe there's another path from Pre-IR to OECD-level economics that doesn't involve sweatshops, but if so, can you point to a country that's done it?

Kenya hasn't developed yet, why do we think it will?

Putting it another way, Kenya's development has not been held back by lack of sweatshops but by corrupt government and misguided go-gooderism.

> by lack of sweatshops but by corrupt government

I think this is a false contrast. Stable government and reliable infrastructure are the standard requirements for building factories (sweatshop or otherwise) in a country. So a lack of industrialization is frequently a consequence of bad governance.

Notably, corruption itself isn't an issue. China, for instance, has plenty of corrupt and autocratic governance. But it has predictable, tolerable corruption and a basically-stable political structure, however the specifics may change. What's important is that the situation was predictable; everyone was willing to skim a percentage and leave the basic structures of government intact. If investors were worried that the government might take 100% of an investment, or fall completely, the situation would look pretty different.

Excellent point. If you look at China, they took a long time to get to a state where they could start developing reliably. Basically after the fall of the Monarchy, there were revolution after revolution, until the Communist takeover. Even after, it took much later, until the rise of Deng Xiaoping, for development to actually take place. By that time, the State had effectively established itself as the Supreme political power (albeit a non-democratic one) which lead to the stability that allowed massive industrialization to take place rapidly, often with assistance from the State.

While growth has slowed in recent years, 6 of the world's 10 fastest growing economies were in sub-Saharan Africa from 2000 to 2010 (https://www.economist.com/blogs/dailychart/2011/01/daily_cha... ). While Kenya isn't included among those countries it has also shown robust growth. It may be that this growth is unfairly geared toward the wealthiest Africans but I think you'd find that the life of the average Kenyan has improved markedly in the last 20 years.

Kenya's GDP's growing at 5.8%/year. In 30 years that would be 5x. All Africa's developing fast at the moment. A lot of it's probably down to modern communications. Africa used to be really cut off. Now they've mostly got cellphones etc.

The Economist just had a great article about the process of moving a poor country up the ladder, focusing on Bangladesh. I thought it was a good read, and it touches on the roll of textile manufacturing and the related infrastructure that helps boost an economy.

https://www.economist.com/news/briefing/21565617-bangladesh-...

Kenya will develop by investing in it's people's education and economy. Sweatshop as a vehicle of development is like "war for peace"

Add this to the story about Nike funding and now deploying robots as well, and the economic impact could be really huge. It may help the West though...

That $0.33 is cost of human labor, the article doesn't mention how much the technology contributes to the cost of the clothing.

More importantly, what will you save on shipping and how much more could you make by marking the t-shirt up as "made in America"? What are the long-term brand benefits of being American-made?

considering the low price of the shirts, how do you cover the capital costs of the robots?

I'm having a hard time imagining the robots putting out 6 digits-worth of shirts per robot to cover the costs?

Perhaps they can focus on customization, and sell more expensive shirts?

Well we could react to abundance in a non-imperialist way...

Should we need a basic standard of what can be called Robots? We're in an age of journalism where almost anything that can do what humans do (Guess, had calculator been invented in this age,'This automated Robot can do as many calculation as 3 Accountants' would have been a viral article) is called a Robot. But what we forget is that it's always been the case all this while, Consider a Toll collection booth with humans in it and then an RFID based automated system, I don't remember if those got this kind of attention as Toll Collecting Robots.

Don't even get me started on "Software Robots", aka RPA (Robotic process automation). Yikes what a misuse of terminology.

What do you think should be called a robot? From watching the video, this seems like a scaled down version of the sort of industrial robot systems you'd see in other manufacturing fields (albeit meant for sewing, and not assembling cars or other things).

They were calling a NAS a "data robot" 10 years ago ! (drobo)

Friendly reminder that building a wall isn't going to keep blue collar jobs around. Automation is going to kill them much quicker than immigrants will.

.. and no amount of cheap immigrant labour won't stop automation.

Which is why the less immigrants, the better. Less people kicked out of jobs == easier to implement UBI or similar system.

Fewer.

Yes, führer

First, this kind of post will get you banned from HN if you do it repeatedly, so please don't.

Second, would you please not use HN for political or ideological battle? You appear to have been doing that repeatedly and it's not what this site is for.

https://news.ycombinator.com/newsguidelines.html

The top level comment I replied to went on the political discourse and I just carried on from there. Did he get a warning as well?

Personally I'd be happy to have less political HN. But as long as there're top level political comments.. It's too hard to resist :) I believe mods could easily mega-downvote such comments. Yet particular comment I replied to seem to be upvoted?

As for my later comment, are jokes not allowed on HN anymore? I got grammar-nazi'd, I replied back with a matching joke.

Oh, boy, another series of layoffs and the rich getting richer. "But my free market!!" Means squat when the elite hoards all the goods and the barriers to entry are extremely high.

So... Communism? 100$ shirts? Sweatshops? Ludditism? Also there is no free market anymore - have you seen all the punitive economic sanctions recently.

$0.33 for the robot to produce the shirt vs $0.20 or so for the sweatshop worker in bangladesh.

Take into account the extra shipping costs from bangladesh and it might actually be cheaper

Agreed, and you are also forgetting about speed. This robot can produce t-shirts that can be delivered to customers next day. Doesn't matter how cheep bangladesh might be, they can't match that speed. This opens up opportunities for sales that were impossible before. If you need 100 shirts tomorrow you are willing to pay more so it has even larger margins.

This reminds me about when Bill Gates make that comment saying that raising the minimal wage could make the cost trade-off of investing in automation worthwhile [1].

On one hand, people need living wages. But on the other hand, I think we're witnessing what Bill Gates is talking about here.

What's the solution? Gates has floated the idea of taxing robots.

  [1]: http://www.businessinsider.com/bill-gates-bots-are-taking-away-jobs-2014-3

He is hardly the first person to make that observation - or even the first noteworthy person. This is a debate as old as economics.

But what's a robot? Is a 3D printer a robot? (it has moving parts and builds things) is the Google Adwords code a robot? (it sure makes a lot of money) a quant hedge fund? Do you tax all computers by cost or value produced?

Seems simpler to tax companies for making money.

EDIT: We won't need middle-manager bots that show each other powerpoints all day. What "robot" would get taxed for the disappearance of that function?

In this context, I think he really talks about taxing capital (or capital investments that displace workers).

>Seems simpler to tax companies for making money.

Only if that money makes its way back to the people who would otherwise have been employed because of the robot tax.

Bill Gates, as noted in your linked article, is against aggressively raising the minimum wage.

As with Warren Buffett, he is in favor of raising the income tax credit rather than the minimum wage. It's a radically better solution for improving wages in the near-term. A too-high minimum wage artificially cuts off the supply of who a business can hire, based on what someone's labor is worth to said business (they have to be able to afford the cost of the labor). An increased income tax credit, paid for via higher taxes on the top ~1/3 income bracket, avoids the problem of pricing low skilled labor out of the market and enables businesses to hire a lot more people whose labor isn't yet worth eg $10 or $15 / hour.

Taxing productivity is an extraordinarily bad idea. It's up there with taxing every search result or trying to slow down computing productivity by abusively taxing cloud computing or smartphone & PC usage.

People with immense wealth (Gates, Buffet, ...) are great at coming up with innovative tax schemes to tax anything other than wealth.

Actually, Buffet's idea for quite a while has been that he and the other billionaires should be taxed more.

It may be just PR, but that's a different goalpost.

My understanding of Buffett's tax plan is that it is an income tax. Buffett and Gates both don't need much income proportionate to their wealth, they don't intend to spend/consume their fortune or pass it on to their children.

Buffett's plan calls for a lot of other wealthy people to be taxed, but I don't see him doing a taxable transaction on his main holdings, I believe the effect on himself would be minuscule.

I don't see him doing a taxable transaction on his main holdings

How so? Either he passes it on to his heirs or he uses it (selling the stock and realizing gains), both of which are taxable events. Since he has called for an increase on both the capital gains tax and on the estate tax, he'd be affected in any case.

He will contribute the majority of his wealth to charitable foundations. ( https://en.wikipedia.org/wiki/Bill_%26_Melinda_Gates_Foundat... ). I believe this will allow him to avoid the taxes.

I don't have any objection to Bill and Warren choosing strenuously avoid giving their money to the government, and I don't object to what they are doing.

What I really object to is Warren advocating for other people's money to pay these taxes and presenting it as if it will affect him but he is such a great guy that he is all for it anyway. To me, he is the kid that is about to get in trouble and is frantically pointing at the other kid and mouthing he did it.

Right, but this clever tax loophole called "giving all your money away" is not in any way limited to Buffett, and in fact he's been pushing others to do the same. He's advocating for increasing the tax rate on realized income, including the part that he and his heirs will be realizing.

When these people "Give all their money away" it goes to groups like the Sherwood foundation or the Bill and Melinda Gates foundation and does massive good for science, education, fighting malaria....

That is not a loophole, that is a good thing.

Taxing wealth or income none of it matters. The tax schemes both proposed by either of them would be beneficial for society. Stop trying to find fault because you already knew it was there (by some means I cannot fathom) and just look at the real results of taking their advice.

That is not a loophole, that is a good thing.

Yeah, sorry, I was being sarcastic.

Then you don't have to worry, because they also plan to give away their wealth.

> "...can make as many shirts per hour as about 17 humans"

Surprised they are looking at it per hour. Presumably the automation could run for close to 24 hours a day, whereas the human workforce could not.

So each of these units could replace 51 workers? (3 rotations of 17 workers each doing 8 hours a day). That's before considering the human workforce generally requires 2 days a week off.

Why should I be surprised. That's been happening since 1700´s [https://en.wikipedia.org/wiki/Industrial_Revolution]

Perhaps that will be the way the most developed Countries will return some of the production lines to them.

Even aside from the argument about production never actually leaving developed countries (only the non-automatable stuff left, the rest is growing faster than ever) this won't really have much benefit. If people in your country are buying the shirts these companies make there is already tax being taken for them and it's not like they'll add enough jobs to even notice, that's the whole point.

"Perhaps that will be the way the most developed Countries will return some of the production lines to them"

Why do you care about that? ecology?

It's not like if it's going to translate in jobs.

Less transportation required, which should be cheaper and better for the environment.

Someone would have to build and manage factories. Would need a night guard to feed the guard dogs too :)

The mechanical knitting machine was invented in 1589.

The Worshipful Company of Framework Knitters was incorporated in 1657 London.

Some framework knitters were among the Luddites, who resisted the transition to factories.

https://en.wikipedia.org/wiki/History_of_knitting#Industrial...

I can't quite decide which will be the tipping point for robots replacing humans: Agriculture, food service, or garments. The first seems ripe (no pun intended) for disruption but apparently often requires human skill to identify products. The second seems more likely, as fast food is mostly automated with some human intervention; coupled with self-checkout technology in any language, and it might work. Garments, however, I'm not sure that's the industry. Certainly replacing 17 workers with a single robot (costs aside) seems like a good start, but this also seems like an industry that could uproot and go somewhere else. From America, to China, to Southeast Asia, Africa, and beyond, to wherever human capital can be had for less.

I recently came across this video of a color sensing tomato sorting machine.

https://youtu.be/j4RWJTs0QCk

Really shifted my mental ranking of where agricultural processing was at.

Why? Seems pretty basic and simplistic.

I'm more impressed by the various types of field harvesters

About 20 years ago (Holy shit!) I worked in a potato processing facility. They had multiple belts all about 8' wide that ran through a machine that scanned 100s / 1000's of dehydrated potato flakes for quality every second and could hit a specific flake with a burst of air to blow it off one line and into another. 20 yrs ago. I believe they went under / were bought up by a more modern company 5 or so yrs ago.

All three over the next 10 years.

Agriculture: Computer vision is improving really fast. The technology that Amazon is developing for picking products at its warehouses will also allow them to pick apples.

Food Service: The real killer app here is that special-purpose automated kitchens (like the ones that produce only fresh hamburgers) will combine with drone and self-driving car delivery to make meal deliver 1/4 the price of restaurants today.

Garments: The technology in this article means that production in the EU/USA will cost about the same as Indonesia or Kenya, and the EU/USA labor pool will have more of the high-skill technicians you need to service the robots. Combined with shorter supply chains and faster time to market and you'll see garment manufacturing becoming "local but automated" in most markets.

Depending on the supply of "sewbot technicians" the USA might even start exporting clothing to Latin America or wherever.

Reply on news.ycombinator.com