I don't think people do repetitive, automatable jobs because they can't do any other kind of job, but rather because our current market conditions make it a good choice for them.
And by market conditions I mean everything from wages to availability and cost of education to immigration policies and welfare benefits.
Once the dust has settled after major advances in automation I don't see why people wouldn't be able to work as managers, designers, programmers and engineers even if those were the only jobs left in the economy (but of course there are many more skilled jobs that will not be automated anytime soon). It's just a matter of training and expectations.
I don't think a world where people only need to work 20 hours a week, and where we have 5x the number of every kind of skilled specialists we have right now is a bad place to be. And it doesn't require ditching capitalism entirely, just regulating it differently (taxation, welfare, etc.).
It's only the transition to such a world that is worrying because it will happen faster than people can adjust, not the end result.
>The problem is you can't really run an economic system where only a very few percent of people take all the gains...
I could be wrong here, but I think the commenter, and the author of the article, are speaking on the bipolar nature of creative jobs like manager, designer, or programmer. The value of different designers and programmers to, say, Apple and Google, is kind of bipolar right now. The concern being that it will be more bipolar in the future. Right now we have the 5 or 6 figure a year sort of "warm body" programmer, and you are correct in asserting that anyone can do that job. But, we also have the 7 figure programmer, and only the very best can do that job.
I think their concern is what if the 7 figure Apple designer turns into a high 7 or 8 figure Apple designer, and the "warm body" designers get less and less over time? Kind of like athletes. A few 8 or 9 figure soccer and basketball players, and then all of the others. Most of whom will never be able to make a living at their sport professionally. (Consider: How many young girls did our athletics infrastructure chew through to get Simone Biles? I don't really know? But I'm guessing it was a lot.)
I think it's a fair concern. Certainly well within the purview of the sort of concerns we should be war gaming out at the policy making level.
All that said, yeah, your world of 20 hours a week sounds pretty sweet to me.
I don't see why programmers, chemists, researchers, engineers, etc. compensation would follow the way of athletes. Those specialists are paid for skilled use of their time, they are not paid for being famous. If you have more skills, you get paid more, that's ok. Those fancy 10x programmers don't make regular programmers less valuable though.
As for the entertainment industry - too many people agree to be underpaid for a chance to become famous one day. A career in those industries is not employment, it's running a personal brand business. No surprise that outcomes are so unequal.
While it may not be as extreme as athletes or entertainers, you already see tons of evidence of the huge discrepancy in programmer salaries right here on HN.
In every thread about programmer salaries, invariably you'll have a comment about "Senior/Staff developers making $300/$400k+", then some comments from folks saying "I've never heard/seen anything close to those numbers", then other comments about "Those numbers are totally doable at FAANG companies." Valued programmers at the best companies already make several multiples of income over 90%+ of the rest.
It has already happened. You absolutely did not see these astronomical programmer salaries 20 years ago (indeed, one reason they've become so astronomical in the past 5 years is equity comp tied to companies that have tripled in value). The reason that top programmers at FAANG companies are able to make so much money is the "automation" of the Internet where you have, for example, a situation where only TWO companies take in nearly 60% of all online digital ad dollars (and that percentage has only recently been falling because Amazon is taking some share).
The concentration in market share that the Internet allows also allows for a corresponding concentration in programmer salaries/talent.
And yet despite the monopolies and despite the superstar compensation for top programmers at Google, mediocre programmer salaries did not suffer. Because everyone still needs mediocre programmers. Because they still provide a lot of value.
Fighting off monopolistic tendencies and the disproportionate influence of monopolists and oligopolists – both economic and political – is already one of the greatest challenges of capitalism.
Automation makes this problem worse because our political institutions are corrupt and incapable of fixing the rise of monopolies, not because automation is bad in itself. And unlike brand recognition, lobbying, and other barriers that monopolists put up to keep their moat safe, automation is generally available to everyone, so it doesn't make sense to penalize it.
Inequality is a product of market regulations, not automation. Put a 90% tax on income over $1M, make 20 hour work week / 12 weeks vacation the new standard, introduce a land value tax, make it impossible to hide wealth offshore, and find a good way to distribute all the extra revenue so that people don't starve but still have just enough incentive to work (in a world that needs much less work because of automation). That'll go a long way towards fixing excessive inequality in a heavily automated world. Half of this can be done to great positive effect even today.
I don't know, maybe that's "a whole new system" that the author theorized that we need. But really it's just turning a bunch of knobs on the existing system. We haven't tried even that much, and everyone's already spelling doom.
> Inequality is a product of market regulations, not automation.
Inequality arises naturally out of many different processes. You can think of the economy right now as being Pareto optimal or efficient; any reallocation of capital will necessarily make one person worse off than another. It is an example of the power law distribution.
> Put a 90% tax on income over $1M, make 20 hour work week / 12 weeks vacation the new standard, introduce a land value tax, make it impossible to hide wealth offshore, and find a good way to distribute all the extra revenue so that people don't starve but still have just enough incentive to work (in a world that needs much less work because of automation). That'll go a long way towards fixing excessive inequality in a heavily automated world. Half of this can be done to great positive effect even today.
That sounds to me like the way to hell paved with good intentions. If you increase the tax rate to 90% over $1M, those individuals will simply move their assets elsewhere. What's stopping them from investing in real estate in Japan (where you can still get good deals, because the dollar is still relatively strong, even though it soon won't be)? Wealth redistribution sounds nice in theory, where you can feel good about yourself because hey, you're saving the world right? Except you're making things worse. All those wealthy individuals you want to heavily tax will stop hiring here, and suddenly you have a population of very angry people that cannot work anymore. The illusion of a living wage is better than no wage at all.
The conclusion of the article, to me, is that the concept of a central bank and a State is obsolete. The future is rich city states with zero regulation, Pareto optimally efficient economies, surrounded by wastelands of those left behind. This is what the Singularity entails.
> You can think of the economy right now as being Pareto optimal or efficient
That's true but I'm not sure how that's relevant. The economy is Pareto efficient because it's literally impossible to make a significant change without making someone worse off, not because the current state is somehow good. This doesn't mean we can't make tradeoffs that will be better than the status quo (just not Pareto-better).
> It is an example of the power law distribution
People can tolerate such distributions as long as the vast majority who gets the short end of the stick has it pretty well off. That is the current situation. With automation we will see a higher concentration of wealth in the hands of fewer people. Whether it's natural or not is inconsequential if this puts most people into misery. They will not stand for it, and the society will collapse, but that will not happen because no one wants that, not even the rich, so the government will find some way to redistribute the tax revenue to make sure everyone can live comfortably enough that they don't need to riot. There's really no way around this redistribution if automation will indeed end up a serious problem.
> If you increase the tax rate to 90% over $1M, those individuals will simply move their assets elsewhere. What's stopping them from investing in real estate in Japan (where you can still get good deals, because the dollar is still relatively strong, even though it soon won't be)?
I am talking about a world where automation has eaten most jobs but the economy keeps producing a similarly high GDP (because all the same jobs are being done, just more efficiently), except the wealth is much more concentrated than today and needs to be distributed to a higher degree than today so that people don't starve. My back of the napkin suggestions are for that world.
In that world, other countries face the exact same problem because automation is not bound by country borders, and their governments will face the same pressures. The situation in different countries will not be perfectly equal of course, but neither are countries equal today (by tax rates or wages or environmental regulations), and we're able to manage that with visas, customs, tariffs and other barriers on international trade.
Even today the ultra rich and big corporations avoid taxes by parking money abroad. It's already a problem, it needs fixing regardless of automation and the taxation changes it will likely bring.
There is nothing special about how we happen to be doing things right now. All the tax rates, work hours, work culture, trade barriers or lack thereof – all that and more will change on a dime if we actually have a reason to change that. And with automation we very well might. Even 40 years ago the US had 70% marginal tax rates in the top bracket, and it's not like they were facing mass unemployment or any other crisis.
Yes, thank you, that is exactly my point. The nature of creative work (the kind that is least likely to be automated) is that the output can be highly variable (think Botticelli's paintings vs 99% of everyone else), and thus it is impossible with that type of work to create a broad base of jobs where lots of people (percentage-wise) are well compensated.
Another way to think of it is not just "repetitive vs non-repetitive" jobs, but "variable vs. well defined output" jobs. It may be currently very difficult to become a radiologist, but at the end of the day a diagnosis is either accurate or it isn't, and those types of "right-or-wrong" jobs are the ones most amenable to automation. So what you're left with are non-automatable jobs where the output can be highly variable, and in those highly-variable output jobs all the gains go to the very best.
Creativity in arts and creativity in engineering and other skilled labour are vastly different.
You don't need to be a Botticelli programmer to provide value to your employer because being famous or unique is not how programmers provide value. Programmers are paid for solving business problems. The variability of that is not going to dramatically increase due to automation.
Even today you can tell that mediocre programmers are earning good wages while mediocre painters / sculptors / actors are starving. It is not because they have been automated.
> Even today you can tell that mediocre programmers are earning good wages
Today mediocre programmers still provide value, that's because there are not yet good enough tools that would make these developers obsolete. These tools are coming within the next 10 years.
Speaking as a competent but unexceptional programmer, I find this an alarming claim. Have you any links to worthwhile articles/research on this topic?
If I really am going to be thrown on the developer scrapheap, I guess I should start thinking now about how I'm going to keep a roof over my kids' heads.
I'll believe it when I see it. So far none of the website builders, low code platforms, rapid app development tools, etc. made a dent, and not for lack of trying.
With a best of the best system, you've got to work all waking hours to have a chance of being the best.
So rather than 20h each, we should see some people working a lot, and then a lot of people realizing they have no chance of making it and dropping out of the system completely.
We're at a crossroads. There's different roads we can go down. There's different ways this whole thing can develop. What we look like when we come through this might be pretty ugly? Or not? We don't really know right now?
We just have to hope that society's stewards are being responsible and considering these very real mechanisms when settling on plans for our long term future. These forces are not going anywhere. Technology is with us to stay, the only question is how we live with it.
And that article only highlights my point. When it comes to AI, the very best solutions will take virtually all of the value. If your autonomous car program only fails .0000001% but another autonomous car program fails even less often at .00000005%, eventually the first program is going to be essentially worthless. Of course programs are not always exactly comparable along a single metric, but my point should be clear.
I don't understand why that would be the case. Unless the autonomous car program is free, there will always be room for a lower-quality, cheaper alternative.
Indeed, this is a good question. Along with 7 figures, what about assets vs liabilities: I'm curious how many programmers think about this part of the equation over their career? Except for shares, as you write, I wonder how many are investing those assets into new ideas - to, in turn, build more valuable assets to grow their wealth?
And by market conditions I mean everything from wages to availability and cost of education to immigration policies and welfare benefits.
Once the dust has settled after major advances in automation I don't see why people wouldn't be able to work as managers, designers, programmers and engineers even if those were the only jobs left in the economy (but of course there are many more skilled jobs that will not be automated anytime soon). It's just a matter of training and expectations.
I don't think a world where people only need to work 20 hours a week, and where we have 5x the number of every kind of skilled specialists we have right now is a bad place to be. And it doesn't require ditching capitalism entirely, just regulating it differently (taxation, welfare, etc.).
It's only the transition to such a world that is worrying because it will happen faster than people can adjust, not the end result.