At some point, automation might be a cause for concern. But I really don't think this is that type of headline...
Now the question becomes, "is that a bad thing?" My answer is yes, because the forces that produce the negative feedback compound with each other. It becomes a "negative welfare contagion". It also is fundamentally unjust.
Here, let me pull an absurd scenarios out of nowhere too: Two of the now jobless workers are destined to become the next Beethoven and Musk, only able to follow their dreams being relieved of their crushing day job.
It is just absurd to diminish the positive effects of automation just because the negative effects combined with the (US-only) horrible state of state provided safety nets makes for better sob stories.
Thus, we'll need to find a way that combines providing value to most people, with a mechanism that actually results in methods that provide that value (e.g. machines) being developed and maintained.
This needs to be done in a global scale, otherwise those companies will just tax optimize all the income away.
In Econ 301 they teach you to never say never.
Whether this time is different than the previous rounds of automation or will have larger dislocative effects is the at the heart of the current discussion. We’ve literally been worried about robots replacing all workers since Fritz Lang made Metropolis in the 1920s. Maybe those concerns were early and maybe we’re repeating ourselves. It’s not really clear.
(Not to distract from your point, other works such as Rossam's universal Robots definitely covered the subject matter you refer to, I'm being pedantic about Metropolis largely because the subtle difference I'm emphasizing feels to me as what makes it _so much more accurate and predictive_ to the actual relationship between man, machine, and the owners of production.)
Oh, you mean like Uber and Deliveroo drivers?
In the last century, humans served machines. In this century, they will serve algorithms.
Lang's work is more relevant than ever. In many ways we are re-living the events of 100 years ago, only this time the US are in the driving seat, rather than Europe.
How is that any different than a waiter working for a restaurant or cafe? The old nature of things doesn't change if you replace a notepad with a web service.
This is the sort of insensitive and short-sighted remark that gives HN a bad name.
In one job, you work with a group of people, answering to people. In the other, you work alone, answering to a computer. The computer doesn't care if today you have a headache, your mother died, your child needs to see a doctor, or customer XYZ is an asshole: you will only be compensated precisely for how productive you are at any given time. The computer won't make you favours and won't help you any differently on any given day. You will not get to know the computer better over time, you won't become friends and go out together, or collaborate on a new business. The computer won't invite you to his wedding or introduce you to his children.
It's weird that we have to be reminded of this, but there is no humanity in the relationship with an algorithm.
Beyond that, it's hard to say. There is lots of work that needs to be done - picking up plastics anybody? - but the money is in the wrong places to pay for it, and at some point the flows in the system become so disrupted by wealth accumulation, that the only solutions is to press the reset button in any case.
...such interesting, interesting times.
I think the "Chinese" (or Austen Chamberlain [1]) had it right when "May you live in interesting times" was meant as a curse.
Why the huge discrepancy with the official rate? Well some of it's stay at home parents and early retirees, but a large chunk of that is prison and disabled which have both quietly ballooned over time.
These numbers look worse when you include women and older men, but here is the long term trend: https://fred.stlouisfed.org/series/LREM25MAUSA156S
https://fred.stlouisfed.org/graph/?g=lxWv
One does have to wonder how the statistics are affected by millions of working adult residents participating in the economy as illegal aliens and therefore not counted. The reduction in male labor participation also brings to mind what is clearly now one of the biggest splits in American conservatism: between people like Kevin D. Williamson[1] and Nicholas Eberstadt[2] who focus on a moral and cultural downward spiral, vs Steve Bannon [3] who credits the decline to free trade, and both legal (H1B) and illegal immigration.
[1] Read the last few paragraphs: https://www.nationalreview.com/magazine/2016/03/28/father-f-... [2] https://www.amazon.com/Men-Without-Work-Americas-Invisible/d... [3] https://www.youtube.com/watch?v=NAfm5L_DOLM
Now, if the current trend continues for 5-10 years then we might recover. But, based on historical trends 2018 is likely a peak right before a recession vs the new normal.
We are not seeing pervasive unemployment, we are seeing the opposite. A situation where there are too many jobs, and not enough workers to fill them
So econ 201 would say that there isn't any societial problem right now.
For perspective, a lot of households have seen a large increase in their wealth. Comparatively, the vast majority of food stamp recipients actively participate in the work force but still need government assistance. While it appears that there are many jobs available, it also appears that these jobs don't pay enough to allow the people who work them to live outside of a life of poverty.
That paper also gets into exactly how the size of the various socioeconomic groups are changing. This is their change in size from 1979 to 2014, as a percent of the total population:
- Rich: 0.1% -> 1.8%
- Upper Middle Class: 12.9% -> 29.4%
- Middle Class: 38.8% -> 32%
- Lower Middle Class: 23.9% -> 17.1%
- Poor or Near-Poor: 24.3% -> 19.8%
Statistics like this are certainly subject to biased interpretation and 'massaging'. If one is curious about the source, wiki has a section on the political stance of the Urban Institute [2]. Though the paper itself is very transparent in their methodology and extremely readable. I found it all eye opening to the point that it literally changed my worldview. Everybody is moving up at an incredibly rapid pace. I mean keep in mind this is only over 35 years! We are doing something very right from an economic point of view.
[1] - https://www.urban.org/research/publication/growing-size-and-...
[2] - https://en.wikipedia.org/wiki/Urban_Institute#Political_stan...
Labor participation on the other hand isn't [1]
[1] https://www.statista.com/statistics/191734/us-civilian-labor...
They’re also not stagnant https://tradingeconomics.com/united-states/wage-growth
https://www.rba.gov.au/publications/rdp/2018/pdf/rdp2018-10....
I think you misunderstood. You are describing negative societal value. Long-term unemployment can ruin not only your life but the lives of your family members.
Econ 201's solution is to punt on the issue and start doing calculus, because the economics orthodoxy was developed by heartless market neoliberals and computer scientists, and they can't really rationalize it in human terms. You literally have to go to grad school to un-learn all this shit.
Eh? I've shopped at Uniqlo in SF. Compared to buying clothes at hipster boutiques, it was shockingly affordable. I'd say whatever they are doing is unquestionably offering a direct benefit to the consumer.
They almost certainly will; that's just ECON101. I'm sure H&M and Zara aren't just sitting around admiring themselves in the mirror.
Clothing is a weird market but there is definitely an elastic low-end and Uniqlo is playing in that space. There's no reason to believe the rules of basic economics are not functioning.
With that said, what do you think happens when a second company in the same sector starts to realize the same cost savings, and realizes they can move more units by using the cost savings to lower prices and ultimately return even more to shareholders?
Again, you're completely right. Companies don't lower prices out of the goodness of their cold, black, capitalist hearts. Yet, might there be other factors at work here?
Note that these two are entirely compatible. Prices routinely go down and customers receive benefit but they do not notice.
For example, I wrote about research on effect of barcode scanner here: https://news.ycombinator.com/item?id=11414435
Can you name a single instance where prices did _not_ go down due to automation? You're saying process won't go down tomorrow, but of course that's not how any of this works.
Automation is correlated with cost savings for consumers because usually the competition can figure out the same thing after seeing another player do it. It's a mistake to think they are causally related
Prices also go down if it will lead to more profit. If you can sell 2.1 times as many widgets selling at half the profit then of course you will.
Well sure, but in what world is that going to happen in here? They didn't invent time travel; they automated a factory. This can be replicated. They may also lower prices simply to undercut competitors if they think the increase in volume will work out to more profits overall.
The point I was clumsily trying to make is that it doesn't really matter. There are forces in the US economy that lead to strong feedback loops of negative welfare for individuals and the families of individuals who come on economic hard times. So while your shirt gets $2 cheaper, someone out there is getting addicted to Klonopin while their wife buys bread on food stamps, and their kid grows up to be a gangbanger, who kills another kid at the age of 15, because he had anger issues and they couldn't afford therapy.
I'm being deliberately hyperbolic, but I hope it helps illustrate the point. In the USA it's shockingly easy to end up in a vicious cycle of poverty that hurts everyone, not just the person with no money. Net welfare gain could well be negative from the situation described above.
It could be argued that since automation drives the need for these arrangements, it ought to be taxed accordingly to pay for the measures that mitigate them.
It ends with replacing humans with crystals. (so a combination of paperclip problem and automatization)
You can't imagine having a family when working on Uniqlo.
It should better be automated for the rest of the world.
Yes, consumers are also a stakeholder, but your typical consumer will receive very little or no direct benefit from this change. Eventually, yes, the profits will be reinvested, but the impact will be so diluted as to be a noticeable. Some small fraction of it will probably also go to tax evasion, being expatriated or sequestered by wealthy upper management.
Meanwhile, a whole lot of people are suddenly out of work, imposing significant negative externalities on them and their families. This could conceivably ruin a marriage, or turn a good parent into a raging alcoholic. It may well be that the negative gain to society (in the form of laying off so many employees) has severe follow-on effects, and in terms of net welfare this could do more harm than good.
The effects of leaving the workers out of the profit-sharing is more than just a matter of equity. Without redistribution of the welfare created by automation, it's perfectly conceivable that the net welfare gain is negative due to the negative second-order effects outweighing the positive second-order effects.
This is the stuff they don't teach you in Econ 101, but it's how things really work.