Hacker News Reader

Comment by coldtea | Hacker News Reader

...and it might not auto-magically create new jobs, something which some people repeat as if it is as inevitable as a law of nature.

Also, a large part (40-50%?) of the population might not be needed, or even desired, e.g. poor and older people that put a strain in the economy, so, they might be delegated to "b class" citizen status, and might end up living in some developing world like wasteland/slums in the next 50 years or so, while another big percentage has service jobs, and a 10-20% of luckier people live in closed communities.

It's not some sci-fi fantasy, it's already in development, and even the norm in lots of parts of the world, nothing (and surely not naive techno-optimism) prevents it from coming to the West.

View on HN

As of right now both views seem to just be idle speculation. I'm not an economist but every other time similar things have happened in industry the economy has balanced out. What I don't get is how what is happening now is fundamentally different to the point where the economy won't correct?

I could give you my take but I think CGP Grey really explains it best: https://www.youtube.com/watch?v=7Pq-S557XQU

>I'm not an economist but every other time similar things have happened in industry the economy has balanced out.

"The economy" did not "balance out". "The economy" is not an active agent. The economy was politically managed -- for instance, via the Fed's original mandate to encourage full employment. What prevents similar management taking place today is ideological opposition to it.

The Fed adding improving employment to natural zero to it's mandate post Great Recession alongside controlling inflation, seems to imply that the ideological opposition isn't that strong.

CGPGrey's "Humans Need Not Apply" video addresses precisely that question: https://www.youtube.com/watch?v=7Pq-S557XQU

EDIT: Bah. Too late.

Hehe I think you might have actually been first!

> What I don't get is how what is happening now is fundamentally different to the point where the economy won't correct?

Used to be, worse came to worst, you'd just go be a hired hand for someone: a farmhand, a servant, a soldier, a sailor, etc. It also used to be that four walls and a roof was a home and bread and what-you-can-grow was food. Our standards of living are different now, at least in developed nations. We also had social structures that included multi-generational communities. So the grandparents, older relatives, etc. would watch the children, help around the house, etc.

Also, people ate worse, got sick more, and died younger.

But haven't those other corrections been somewhat painful for many people? I think the fear here is that this is so much larger because instead of letting one person do the work of 10, leaving the other 9 to do something new, this is just replacing all 10 across numerous classes of work, leaving few to no new jobs. It's going to be a lot more painful for a lot more people than previous corrections. It could result in violence, starvation, and the collapse of the middle class, leaving the have-not's entirely at the mercy of the haves with no recourse to get out of that situation.

Really? AI-assisted automation has happened in the past?

If so, I missed that portion in my history classes.

Much worse! Completely, strategically dominant replacement of human and animal muscle with the steam engine.

Yet, without brains.

Still, arguably, a very primitive form of "artificial intelligence."

If robots and AI are eliminating an entire sector (starting with driving for instance), what other manual or low skill jobs are there that aren't already targeted for AI replacement? We don't need 2m more car mechanics or farm workers. This time it seem likely a certain class of job is going away.

Half the population is below average intelligence, and perhaps 25% (total guess) might be best suited to a manual or semi-skilled role. What are we going to do with them?

When similar things have happened previously, the once semi-skilled (eg) shipbuilder becomes a low wage driver or call centre worker. Many in an area didn't get those few roles, so became unemployed (and often depressed) for the rest of their careers. Some steel or shipbuilding towns haven't fully recovered even today.

Longer ago the farmhand could get any number of manual roles in the growing industries. Blacksmiths became car mechanics, fuel attendants and so on.

Also when it's only a number of regional cities affected, like with coal, steel or shipbuilding, it's much easier politically than when it's national.

Relatively low margin, fragmented service jobs. Waiters/waitresses, fitness instructors, barbers etc. Those jobs are often still better than the jobs being automated away (truck driving), just less economic to build such specialty robots, at least in the near term.

Except none of those types of jobs offer the stability of the original industrial jobs, and are generally geared more toward younger workers, who have longer. So it's a much worse situation for those workers as the jobs are higher risk.

If you were a 55 year old long haul trucker, you aren't going to be able to replace the income/benefits you had with a $7/hr (pre-tip) service sector job.

Many service sector jobs are not minimum wage and not geared towards young people (any more than truck driving). They do often have utilization issues, but so do truck drivers. Generally that's probably more of an argument for those folks to organize themselves than anything else. Truck drivers certainly did (The Teamsters).

>$7/hr (pre-tip)

waitstaff isnt paid that much hourly. Its generally on the order of 2-3$ per hour.

edit: since im being downvoted i assume someone disagrees - here is a source:

from nolo.com:

>Employers must pay tipped employees at least $3.23 an hour in 2016, $3.38 an hour in 2017, and $3.52 an hour in 2018.

http://www.nolo.com/legal-encyclopedia/michigan-laws-tipped-...

Confused Brit question - how? That's below min wage, which I thought was national law. I imagine it's a bit lower for youngsters.

> Under federal law and in most states, employers may pay tipped employees less than the minimum wage, as long as employees earn enough in tips to make up the difference. This is called a "tip credit."

Technically you can sue your employer if they wont pay you the difference if you dont end up making enough in tips to end up at or above minimum wage - in practice however, suing your employer is generally not a great way to stay employed

http://www.nolo.com/legal-encyclopedia/michigan-laws-tipped-...

Damn, that's harsh.

Only way I know round UK min wage is to be self-employed, or one of the fake "self-employed" like some delivery drivers have had to become. Surprised the IR haven't slammed the door on that little scam yet.

Wait staff have their own minimum wages. Iirc it's under $3/hour.

That said, it's not uncommon to get at least minimum wage.

7 is an average - as some states have a minimum up to $10/hr

https://en.wikipedia.org/wiki/Tipped_wage_in_the_United_Stat...

Aren't most of those already down at min wage? At least a proportion of waiting jobs will be going away too.

Can they absorb another 10m (or however many there are) truck and van drivers who probably made quite a bit more? I know it won't be overnight, but it seems to be rapidly becoming a political rather than economic question.

Like must we learn to run at 20-25% unemployment instead of <10? We'd probably have to learn to stop viewing them as scroungers too and adjust welfare somewhat.

Many are, but not all. And I suspect waiting jobs as ratio compared with eating out gross revenue is shrinking (though eating out is growing and is more labor intensive than a grocery store still). Part of that is labor organization, the truckers have unions. And unions were (and are) a big political issue.

> Relatively low margin, fragmented service jobs. Waiters/waitresses, fitness instructors, barbers etc. Those jobs are often still better than the jobs being automated away (truck driving)

No, they aren't. Many "low-margin, framented service jobs" of the type you list are often effectively sub-minimum wage (possible because they are usually independent sole proprietorships renting facilities from the entity that looks like their "employer") highly-saturated already, increasing the proportion of the population trying to work them will further crash wages.

so in the face of massive layoffs of entire sectors of the economy - you are suggesting that restaurants will be hiring /more/ people?

Who is eating at these restaurants that wasnt before? Why would they suddenly be hiring millions of people at restaurants?

Who is suddenly attending more fitness classes now that there is massive unemployment, to the extent that we need millions more fitness instructors?

Your suggestions dont make any sense whatsoever. If you had said Campbells soup company would be hiring you might be on to something, but in the face of millions losing jobs, luxuries like restaurants, fitness classes, etc certainly wont be seeing massive growth in demand....

>Half the population is below average intelligence

Assuming intelligence is normally distributed, which I guess is a commonly accepted thing :-D

It's not that it won't correct, it's that the correction offers no guarantee of being humane.

In the 20th century, the population of horses (in the US) started declining rapidly, from a high of over 20 million in 1920 to under 5 million in 1960. Industry decided the horse was unnecessary after thousands of years of history, thus they were slaughtered in short order.

The only difference between me and a horse from the market's view is that I do different jobs and consume human goods. To the extent that I am necessary it is because the market will need someone to move goods and services to, but left to its own devices it might well decide to sell weapons that I get killed with. Markets are honest, not smart.

Citation on horse slaughtering in the mid-1900s that probably was relatively unrelated to the Great Depressions? Legitimately interested.

I don't have a detailed source on hand for the slaughtering operations specifically, but this is a good overview article of where the meat went:

https://priceonomics.com/when-americans-ate-horse-meat/

Just to be clear for everyone who doesn't click through. The car was one (and seemingly a very small one) amongst many reasons people ate horses, coupled with varying economic recessions, shortages of other meat, and changing consumer taste.

Yeah, over 40 years it would be easy to reduce the number of horses simply by attrition.

Could you explain what you mean by "the economy", and how it is of importance that it has "balanced out" ?

I mean, one could probably say that the economy has balanced out after the great depression, after WW2, after the 2008 crisis, and so on. Many people did not.

Well, if it takes some time for workers to shift sectors and if "creatively destructive" technological innovation on the other hand happens to accelerate, it's conceivable that at a certain point, the absolute majority of workers will be redundant. There won't be an equilibrium, because the workers are trying to chase an accelerating point while moving at a constant velocity.

There doesn't need to be anything "fundamentally different" except that we finally reached the exponential part of technological progress, which up until now seemed like a linear process. Just like population growth until 1800s, or economic growth until the 1820s. Those two were pretty much linear and not very different from zero until they weren't.

I don't actually subscribe to this rationale, though. I'm just putting it out like this because it's pretty much what everybody says and I recognize that it's quite plausible, given the two first hypotheses.

> There doesn't need to be anything "fundamentally different" except that we finally reached the exponential part of technological progress, which up until now seemed like a linear process.

Most processes involving people (and lots of others, besides) that look exponential at some point are really logistic, so even in that case, there's probably some equilibrium point somewhere in the future, though its possible that its many generations down the line from the point where the appearance of exponential change first occurs.

Good point. It's quite possible that there's a fundamental limit to technological achievement and that we just haven't reached it yet.

On the other hand, I'd like to say that it's also possible that some particular processes can follow exponential laws without any restriction. It might be somewhat hazy, but what if outside of a few measures (infant mortality, life expectancy, etc) economic progress is nothing more than a subjective measurement of how present consumers value past output?

I'm not saying anything new here, both Easterlin's paradox and the idea of a "hedonic treadmill" support the idea that more income doesn't necessarily equate into increased happiness. And specially in the case of economic innovation, which might not actually be technically interesting or require any "new physics" at all - just mashing some current technologies into a marketable product - it might never ever stop, and will just degenerate into an endless chase of ever changing fads. Past economies wouldn't be particular poor, just particularly quaint, and a repository of dead fads.

https://en.wikipedia.org/wiki/Easterlin_paradox

It's very far from idle speculation at this point; many prominent economists and economic researchers[1] have grappled with the issue. Both the Lump of Labor Fallacy and Luddite Fallacy may prove themselves to be fallacious in the new era. Gwern has a good post with links:

http://www.gwern.net/Mistakes?cx=009114923999563836576%3Adv0...

Here are a few prominent ones:

http://economics.mit.edu/files/5554 https://hbr.org/2015/06/the-great-decoupling http://www.economist.com/news/briefing/21594264-previous-tec... http://ftalphaville.ft.com/2014/08/28/1942571/david-autors-g... http://krugman.blogs.nytimes.com/2011/03/06/autor-autor/?_r=... http://www.asymptosis.com/are-machines-replacing-humans-or-a...

[1] Autor, Ford, Cowen, Krugman, Acemoglu, Brynjolfsson, McAfee just to name a few

I think some of these are good insights. There is no question that society will change a lot over this century, I'll be curious to see if it changes in a way that anyone predicted.

>Also, a large part (40-50%?) of the population might not be needed, or even desired, e.g. poor and older people that put a strain in the economy, so, they might be delegated to "b class" citizen status, and might end up living in some developing world like wasteland/slums in the next 50 years or so, while another big percentage has service jobs, and a 10-20% of luckier people live in closed communities.

I think this is very possible. And I suspect that the overwhelming majority of the b-citizens will be people not knowing how to write a software.

I say it continually to those I love (close friends, family etc...) : unless very specific career choices, starting in life today and deciding that "having no clue about coding is fine", is at best a dangerous gamble, at worth a suicide.

I think Darwin kind of applies in our modern society but people do not notice it. For example, what led me to doing engineering studies was pure survival instinct.

I have a business profile - and am a business owner today - still heavily writing code, and intending to keep doing it.

Since I'm about 12, I knew that my stuff was starting companies, business strategy, product design, corporate finance and this kind of stuff. That's what I was toying with in my spare time. But a pragmatic analysis of the world led me to quickly realize that I'd be better off being proficient in technology stuff, to avoid being at a big disadvantage since all the stuff I'd be selling and designing were going to be technology-based.

At the time, I said in the way children speak : "I want to be able to build what I will be selling so that I do not have to depend on the good will of anyone". And this was the smartest move I ever made in my life. And this kind of thinking is what continually puts me ahead.

I think it would be interesting to find a way to measure the "survival instinct" of people, and see if there's a correlation with those who do "well" and those who find themselves surprised at every recession and industrial revolution.

You think millions of unskilled and semi-skilled workers can be retrained as software developers and that there'll be employment opportunities for these millions of new programmers?

I think it's much more likely that many existing programming jobs will also be automated away.

The concept of spending a life being employed is rather new. In the future the majority of people will be business owners... like in the old days.

And what I'm saying is that these businesses will likely be software-based, and that those being software-illiterate will have a life as comfortable as the one lived by illiterate people before them. Software is becoming the new paper, this is what I am saying.

Today, we on HN are like the Lords and the clergy who were the only ones able to read and write and hereby be involved in the important business of their times.

The history only repeat itself. And it will always.

I think, given modern progressive ideals. We will see shorter work weeks and multiple people with the same job. Instead of 2nd and 3rd shift, maybe we Monday-Tuesday-Wednesday shifts.

At least, I hope.

My sister does job-sharing. She works Mon-Wed, and her job partner works Wed-Fri (one day overlap for meetings and handover). This lets her spend more time with her family, something that's important to her. She gets healthcare coverage through her husband's job.

It works well, but you'd have to commit to a less-expensive lifestyle because you're not working 40+ hours a week. It has benefits to the company as well, as they get two people for the cost of a little more than one (and don't have to pay expensive benefits to either of them).

This will have a problem if you are no longer earning enough to be able to afford to live near to your job.

I still think that Basic Income at a level that will allow you to live somewhere inexpensive will be a better solution.

why not just an increase in the earned income tax credit?

Because you have to have a job to get eitc.

That would need some serious political intervention. For companies it is more efficient to have one employee work longer hours because it reduces the communication overhead.

I also don't see how such an arrangement will help, for example, taxi or truck drivers who will be completely replaced by autonomous cars in the near future.

It depends on what things become cheaper. I don't see how poor would endup in wasteland/slum scenario, unless the revolution is so disparate that it only benefits some people and completely excludes other people. (Remember, even if say programmers benefit from automation, non-programmers also reap the benefits of automation).

Wasteland/slum scenario seems unlikely; if 40-50% of all people did not live in a proper house[0], there would be a lot of empty houses.

(0) as opposed to a hut/shed

Yeah, if you have robot factories and practical batteries, the only thing holding back abundant energy becomes the political will to make abundant energy available.

Why wouldn't the poor end up in a slum, even if it's 50% of the population?

You seem to forget that "slums" were not always slums. They used to be reasonably nice places to live. (Some of them in the US used to be really nice, posh places to live.) All it takes to turn a nice neighborhood into a slum is a lack of income to keep the houses maintained properly. Human habitats fall apart pretty quickly when they're abandoned or neglected.

Because if significant amounts of the workforce are replaced there's an implicit assumption that a significant amount of things would be cheaper to provider, if the only cost of creating and supplying food is robot maintenance that means food should be significantly cheaper to produce than now.

Coming 2046: P2P food sharing.

> Because if significant amounts of the workforce are replaced there's an implicit assumption that a significant amount of things would be cheaper to provider

That's a wild assumption. I would predict that it's more likely all that additional value will be captured by already-rich business owners and shareholders, not consumers.

> I would predict that it's more likely all that additional value will be captured by already-rich business owners and shareholders, not consumers.

If you are concerned about that, lobby to create laws that make it easy for consumers to invest their money into this kind of undertaking.

What money are the unemployed going to invest?

There is a lot between "insanely rich" and "no money".

History indicates otherwise, no?

I don't think food and manufactured 'stuff' has ever been cheaper relative to median wages than it is now.

Jobs have rapidly been disappearing for 250 years, consistently resulting in equally rapidly increasing living standards.

You have to (1) demonstrate that you understand how this very counterintuitive historical fact happened and (2) give a plausible theory for why it's different now, for me to even begin taking talk like this seriously.

For "250 years" jobs have been replaced (first with factory jobs --Ford, Taylorism, etc.-- and then with office and service sector jobs), they haven't been actually disappearing.

Now and in the future, with software, robotic automation, and even more so AI, jobs are actually disappearing for the first time ("software is eating the world" so to speak).

Examples from the era of the steam engine and the advent of electricity when faced with what's possible with advanced technology like AI, modern automation and 21st century internet and software are not really applicable.

That said, it's not that mankind cannot manage this. But it cannot manage it in the ways it had in the past -- this will only lead to the marginalization of large parts of society. It can only manage it with political action: 4 and 3 day work weeks being mandatory, universal income schemes and other such measures.

The jobs have always disappeared first, then new ones have appeared. People have always not known where the new jobs would come from, and they have always appeared.

Even right now, when software is eating the world, unemployment in the US is low and decreasing. The disaster is a bit like that my new diet, it always is happening soon :)

>The jobs have always disappeared first, then new ones have appeared. People have always not known where the new jobs would come from, and they have always appeared.

That's less like actual history, and more like a retelling of it as some inevitable repeated story.

If by people, you mean the ones whose jobs/skills became deprecated, then yes, they didn't know where new jobs will come from, and tons of them didn't get new jobs anyway.

But the industrialists in general knew they will have needs for tons of workers in the new factories -- and in fact the government then (in England, etc) passed laws that made sure they "encouraged" (forced) millions of farmers and such to urbanize and become factory workers.

The difference is that now we know of no such need.

>Even right now, when software is eating the world, unemployment in the US is low and decreasing.

Mostly by cheating: counting tons of low quality, low pay, part time (not as a choice) jobs, not counting people who have worked in the past X months, not counting tons of people outside certain narrow criteria that still don't translate to actual jobs, etc :)

E.g.:

The unemployment rate dropped because people unable to find jobs ceased looking and are no longer counted as being in the labor force. If you are unemployed but not considered part of the labor force, you are not included when unemployment is measured. The BLS says that in May there were 1.7 million Americans who “wanted and were available for work,” but “were not counted as unemployed because they had not searched for work in the 4 weeks preceding the survey.” In other words, the unemployment rate is a useless measure of unemployment, just as the consumer price index no longer measures inflation. What were once useful statistical measures have been converted into good news propaganda.

http://www.counterpunch.org/2016/06/06/employment-lies/

> nothing (and surely not naive techno-optimism) prevents it from coming to the West.

Well, if you stop dividing people into "West" and not West, and consider the whole world without artificial borders as just humanity, it's already the case.

I was making a point for those that feel that they are the "West", and that "it can't happen here".

That said, the world is not without borders, artificial or not, so it's not wise to see it as such in determining the future. It might be wise to want to abolish those borders (and that remains to be proven -- empires are even less democratic than nation states, and civil wars are just as bloody, if not more, than wars between states), but not to live in la-la-land where they don't exist.

> some people repeat as if it is as inevitable as a law of nature.

It's just happened every other time something like this has happened. For tens of thousands of years.

> wasteland/slums in the next 50 years or so

Standards of living are rising globally. Even if comparative wealth drops vs other humans, most people will still see significant improvement in their lifetime.

>It's just happened every other time something like this has happened. For tens of thousands of years.

Nothing like this has happened before.

For one, people in 17th century didn't live much differently than people in 500 A.D. or 500 B.C -- the huge majority lived in the country, milked their own milk, work with manual tools, etc.

The process we're discussing started mostly with the industrial revolution. With the advent of the steam engine and crude early factories, for the first 2 centuries of the Industrial revolution, as factories and factory assembly lines still needed tons of manual labor (see Taylorism).

Then, as automation advanced, around the mid-20th century, and there was not the same demand for factory workers anymore (a lot of production lines became more efficient and less manual), people were given office and service sector jobs -- because that was another underdeveloped sector.

Today, software and the internet makes lots of office jobs and procedures obsolete, advanced robotic automation makes tons of factory jobs deprecated, and with advances in smart software, business networks, etc, lots of service jobs are not needed anymore either. Advanced AI software and hardware will only make this worse.

And there's no next step to go besides either creating useless busywork or reducing work-hours/days (so work provided meets the actual work needs). And even that will still leave tons of people not needed to work anyways.

I'm not sure how anyone in a technology forum can believe technology can't get us to a point where a much smaller number or very few needs to work anymore, because "it never happened in the past".

That's exactly the idea behind innovation, to not get the same thing you had in the past...

>Standards of living are rising globally. Even if comparative wealth drops vs other humans, most people will still see significant improvement in their lifetime.

We're not talking about whether millions rural chinese person improved their "standards of living" in the 20th and early 21st century by getting work in a factory, but whether they'd still be needed in 2025 or 2030.

http://www.bbc.com/news/technology-36376966

You vastly underestimate the technology differences between 500 b.c. and 1600ad. The printing press for example. Windmills/waterwheels for another. Did the automation of grain mills result in mass unemployability?

> office and service sector jobs -- because that was another underdeveloped sector.

Or perhaps automation allowed more people to pursue that type of work? Was working in an air conditioned office an option for your average person in 500bc? Doubt it. Instead it sounds like we created a whole new type of work in response to a society wide change in technology.

> software and the internet makes lots of office jobs and procedures obsolete

Some, but fewer than you would think, and it makes many other jobs accessible. Often technology advances are unskilled people + tech/process improvements displacing skilled artisans. The factory worker jobs were an example of technology displacing skilled workers. Computers and the internet allow for unskilled/untrained people to enter fields that they would be unsuited for alternatively. How many retail bankers could do an amortization table without the computer's assistance?

> I'm not sure how anyone in a technology forum can believe...

Precisely the opposite. Technology eliminating work has happened several times in the past. Just never with disaster. We always seem to find new things to work on. The greater efficiency we have the more our horizon expands about what we can pursue as 'work'. If you were a good actor in 500BC, what would that buy you? Some local notoriety? Would it be 'work' that you get paid for in your local town? Probably not. Now it means you get a private jet. New areas will expand to fill the gaps.

Even among the industries being automated, the results are often not that dire. There is a milder version of Jevon's paradox at work here. Once you can produce something with great efficiency and low cost, people will demand more of it. Sometimes demand will even increase faster than efficiency and you'll have net employment growth.

>You vastly underestimate the technology differences between 500 b.c. and 1600ad. The printing press for example. Windmills/waterwheels for another. Did the automation of grain mills result in mass unemployability?

Not many worked in grain mills in the first place for that to be much of an issue. Something like the fishing sector becoming deprecated in the 16th century B.C. would more faithfully describe the kind of impact modern automation technologies can have.

Besides there were still tons of work opportunities in ALL kinds of manual and menial labor fields as the standards of living were very low and urbanization was just starting.

>* We always seem to find new things to work on.*

From a list of limited options -- not from some magic infinite bag of sectors.

>Precisely the opposite. Technology eliminating work has happened several times in the past. Just never with disaster.

You conflate technology eliminating SOME work, with technology that can replace all kinds of jobs -- including lots of mental labor with AI.

You also conflate technologies that were small multipliers of productivity, with technologies that allow 1 person to do the work of hundreds of thousands, or even millions.

Send them to Mars!

Reply on news.ycombinator.com