First unmanned factory takes shape in Dongguan City (2015) | Hacker News Reader

First unmanned factory takes shape in Dongguan City (2015)

en.people.cn

107 pointsdoener61 comments

This article is wrong. The output per employee changed to 250%. So by reducing the staff to 10%, the total output is roughly 25% compared to what it was before.

http://monetarywatch.com/2017/01/chinese-factory-replaces-90...

edit: added original source courtesy of dmoy http://en.people.cn/n/2015/0715/c90000-8920747.html

Do you mean 25% higher?

If you read the original Chinese source, it's saying

Old: 650 workers * 8000pieces per person per month

New: 60 workers * 21000 pieces per person per month

It's a bit ambiguous but it makes it sound like total production output has dropped fourfold.

Even then, it's not telling the whole story without knowing the cost per worker and the value per unit. When you take costs into account, you can get a situation where lowering production will increase earnings.

Taking your numbers and plugging in some values like 75 per worker and 0.01 per unit would increase earnings by 250%.

and the cost of amortizing the new plant

Yea, I'm wondering if we can dig up a Chinese language source and see if it has more details. I'm guessing something got horribly lost in translation.

Or, possibly, the factory managers are purposefully not giving the whole story...

Previously they made 5.2 million pieces per month and now they make 1.26 million pieces per month.

Seems like a disaster to me!

If they reduced their workforce by 90%, so that the cost of labor is now a tenth of what it was, then they may have improved overall profit.

Of course, that doesn't include overhead, or anything else.

It it's an immediate decrease in output but they are planning on moving from 60 robots to 1000 robots in 2 years. so the 2 year expected capacity is 21,000,000 with a 5% defect rate vs. the 5,200,000 units with a 25% defect rate and wages/benefits for all those people.

All that matters is per unit cost. The rest is just a matter of scaling up.

That doesn't make sense from a business perspective. No one would invest heavily with output expected to drop to 25% existing levels. More likely is that the factory's total output is 250% of what it was before the employee cuts.

Tech Republic says:

> According to the People's Daily, production per person has increased from 8,000 pieces to 21,000 pieces. That's a 162.5% increase.

I'm highly suspicious of the factual accuracy of anyof these news reports.

It's scary how quickly all "data" in articles suddenly seems to have gotten unreliable.

I hope this leads to a resurgence in journalistic integrity. I would gladly pay for news I can trust.

I'd argue that such hasn't suddenly become unreliable. It was always unreliable, it's now far easier to fact check (and frequently do so in exchange with experts in a given field, as happens so often here). Now you have a million eyeballs on the bogus claims / data, and those eyeballs frequently have fast, easy access to vast research sources.

Today if the President says something bogus, it'll be fact checked in minutes by a thousand people and that fact checking will be syndicated globally nearly instantly on a dozen large media platforms. By comparison, consider the process, access and distribution that existed at the time of FDR (the things you can get away with when it takes a week or a month for distribution to occur).

We have a lot of historical examples from the last two centuries, to show that data accuracy and reporting in terms of commerce, politics, military, journalism etc. were not particularly extraordinary in the past. Agenda, laziness, bad research, bad reporting, misinterpretation, etc. are nothing new.

There is that, but also there's been increasing pressure to pump out more faster due to more competition, and people paying less/nothing. I think quality-per-article has probably gone way down in the last decade, but we do have a ton more articles.

They are also paying less salaries, so depending on their margins they might be making more profit that way, which is the actual thing that counts and not output volume.

Exactly! Thats what people don't get!

You can even now charge more because the quality better now!

The future will be like: better sell ten iPhones for 10.000 than 100 for 1000.

Why?

Less support,less headache,less struggle to organise the production etc.

dang changed the article so now this comment thread is a little confusing. The original article incorrectly claimed that total production increased 250%.

This should probably have a [2015] in the title and maybe point to at least the English language people's daily source:

http://en.people.cn/n/2015/0715/c90000-8920747.html

Thanks, we've done both. Url changed from http://www.zmescience.com/other/economics/china-factory-robo....

This story is from People's Daily two years ago.[1] The original article is more useful. The factory supposedly had 650 people doing final polishing on mobile phones. That task was automated.

[1] http://en.people.cn/n/2015/0715/c90000-8920747.html

In most factories in China, humans work like if they where robots, one job might be to go with a part from one place to another place, witch could literally be replaced by a transport band. Some workers are extremely skilled and fast though, replacing them with robots would probably cost much more then their salary. It can be hard to see the economic sense in deploying robots when the robots does things five times slower, but when the robot works, you can do something else. Even if you can clean the dishes faster then a dishwasher, it does save you time! And once something has been automated, it's much easier to scale.

Dishes are a bad analogy. I can wash a load of dishes by hand in 15 minutes. My washer takes 90 minutes. From a time POV, it could be worth it to pay fleshy robots to do the work.

Dishes are really a bad analogy, yes: http://www.treehugger.com/kitchen-design/built-in-dishwasher...

How about hygiene standards and temperature? The machine can have a major hedge on that.

They can get hotter, but they are also happy to take some crusted on food and bake it on to the plate... making the QC factor a bit lower.

These should be American robots.

They probably will be made by Kuka...

Anyone here questioning what the value is aren't realizing that scaling goes up so so much easier without humans getting tired, getting distracted. It's much easier to coordinate a million robots than a million humans.

Too bad the article does not discuss the time it would take to break even on that investment.

I feel like this is always missing. Robots don't scale down, which is a real issue.

They kind of can - you can lower the duty cycle and batchify things. You can also outsource to them, and/or otherwise share them with other producers/production lines. I personally believe future manufacturing will be significantly more dynamic with respect to assembling/disassembling production line unit operations via electronic contract (re-)negotiation.

At http://8-food.com/ we are essentially placing miniature (fresh food preparation) factories in retail spaces, and so I've been reading a lot of industry literature recently. I've found very interesting crossovers with former work on a major Bitcoin exchange - eg. multi-asset/service/settlement path capable transaction markup and negotiation systems - http://ifex-project.org/ In the same way factories use JIT (just in time) concepts to optimize inventory and supply-chain logistics, we similarly optimize for delivery of ingredients and packaging, and thus reduce storage bulk overheads and achieve form-factor minimization. Operations research truly is operations generic, and quaintly fascinating coming from a programming background.

You can do anything with unlimited funds, the existence of such technologies or the dream of such technologies doesn't really mean anything. The comparison should be against alternatives.

Human labor is still far more scalable than any technology can provide. The reason the tech industry so strongly supports minimum wages and other increases in cost to labor is because it makes their products more competitive against humans.

A fair point, though I think it's fair to say that due to the very nature of HN and startups a lot of people here choose think in the 'what is just around the corner' and 'what is coming' domains, shunning obstinate currency (perhaps subconsciously as done/uninteresting/unchallenging).

Defect rate dropped by 80%, I wonder how much of the productivity boost is related to that.

Off tangent question - did anyone here fill out the email subscription form that takes up the entire screen when you go this article? I keep getting told they "work" even though I think they're terrible.

They work on the average person. Mostly due to them putting no resistance to filling out a form.

I wonder what they'll do when there's nobody left to buy the products. Maybe they can sell them to the robots themselves?

Why would there be nobody left to buy the products? And what does it have to do with replacing workers? I imagine that if they are producing something, they are doing so because of a demand. Once the demand is gone, they most likely will stop producing.

demand for consumer products is driven by consumer confidence, i.e. by surplus in worker wages after cost of living is accounted for.

if the workers are having trouble getting a decent wage, where does that surplus come from? what happens to consumer confidence, and aggregate demand if people are losing their jobs?

If we can figure out a decent economic system, then there won't be problems giving people a decent wage. Sure, if we keep doing things the way we are now then nobody will get a decent wage, but that's a very fixable problem.

It is not a fixable thing.

You don't fix thousands of years of human evolution overnight. Or one generation.

It will take several thousand years to fix that.

Are you actually arguing that it will take several thousands of years to change our economic model? I find this argument hard to take seriously

My guess is five to six thousand years before humans move away from the "winner takes it all" scheme ( or alpha behaviour)

In twenty years we pretty much will know who is right.

I'd rather buy shares of companies then hoping for a sudden change in human behaviour.

We can change human behavior. We do it all the time. We just need to incentive it differently like we do with other systems. In 20 years we won't know "who is right", we'll just know that we went with system X instead of system Y.

Funny because there have been societies which have not been that way in the past. Also there's pretty much 0 evidence that that's a component of evolutionary psychology vs just the social norms in our given society.

>if the workers are having trouble getting a decent wage, where does that surplus come from? what happens to consumer confidence, and aggregate demand if people are losing their jobs?

Probably not a lot. Automation headlines are probably going to take a back seat to Trump administration headlines for a good long while. Just from feeling, it seems to be happening fast enough to be a real concern, but not fast enough to cause panic, draw attention, and quick action.

My guess is that if robots put enough people out of work then the ability to pay will impact the demand

Basic income is the only future, truck drivers will be all out of work by 2020 for example

Read "The True Believer" and see if you come to the same conclusion.

"It is usually those whose poverty is relatively recent, the “new poor,” who throb with the ferment of frustration. The memory of better things is as fire in their veins. They are the disinherited and dispossessed who respond to every rising mass movement."

In 3 years all truck drivers will be out of work? We don't even have fully autonomous trucking released now (beginning the early adopter part of the technology adopter cycle)... so that seems a bit far fetched.

Trucks can drive a well studied route, self driving cars are much more complicated

The road is not predictable. A deer running across, a car cutting you off, rain, snow, a tire blow out, any number of random events can happen. It just occurs less often on a freeway than the city. You still have the solve the same problems, however, if you don't want a human in the car.

>Basic income is the only future, truck drivers will be all out of work by 2020 for example

This has become the refrain in the comment section for every article dealing with automation I've seen, of late. Some of the articles, I suspect, are being posted by basic income proponents.

Even accepting the premise that a basic income would be needed in the case of near-universal automation, there are plenty of us who are yet to be convinced that we're heading in that direction.

I don't think most of the people talking about basic income are the proponents. I'm a strong believer in that you should be paid the market rate for your work, so if you want to make good money, you must make smart choices about your education and career. That makes the whole system more efficient. I do however think that basic income is inevitable, because of the automation on the scale we've never seen before. Inevitable, because the other option is to let the unemployed starve. So not really an option.

I'd like to think that letting the unemployed starve is not an option, but it is. People can be extremely cruel when they don't have to take responsibility for the outcomes of government policies.

They lean on "those people should have worked harder/planned better"

It wasn't that long ago that poor people who needed critical medical care were kicked to the curb to die on the street because they couldn't pay.

Ok, you're right, it is an option in third-world countries. I don't see it happening in the EU/US/AU/CA or even China (even considering their past human rights violations). But if it does happen on a large enough scale, hungry riots will just take over the means of the food production (mostly robots and other machinery these days).

> It wasn't that long ago that poor people who needed critical medical care were kicked to the curb to die on the street because they couldn't pay.

Healthcare is a separate issue. What you're describing is still happening all over, including the US. I've been looking at the US health insurance premiums, and I have no idea how an average person can afford something like that.

Idk if it's right, I'd like to see other options. And those proponents you speak off usually aren't talking about themself either. They want it for others. The people who would be receiving it are against it

In most of human history, we didn't have basic income. Instead, societies either let the working classes wither away to reduce in size naturally, which came with the risk of revolution as people would resist to accept their fate, or sent them to war, which had the side-effect of increasing national cohesion and possibly result in the gains of territory or resources as spoils.

Basic income is an expensive band-aid for overpopulation where far more productive, albeit less compassionate solutions exist.

The difference is, modern wars are either extremely deadly or not at all.

That solution won't work anymore.

How about going to space instead?

Basic income is a solution that assigns value to human life outside of its ability to build widgets in exchange for pay. An economic system which values productivity over all and which totally devalues human life is disgustingly cruel and inhumane. I would not want to live in such a world, even with a guarantee that I would be among the productive elites and not the unproductive masses.

There must be some solution that doesn't involve letting more than half of us starve to death or die in meaningless conflicts.

Reply on news.ycombinator.com