[TeMPOraL]

You seem to be ignoring the current wave of automation, and focusing on the one that started with the industrial revolution. Automation used to provide muscle, and thus replace manufacturing. Nowadays it starts to provide brains, which makes it replace services. As it continues, where will the human labour go?

The number of jobs was highly correlated to the size of population so far, but we already see the rise of bullshit jobs, that serve each other in closed loops, providing no real benefit except of wasting resources to give some people something to do. It's not as obvious as ordering people one day to dig up a hole and the next day to fill it up again, but e.g. a lifecycle of a leaflet - from commissioning, designing and printing it, to delivery, giving it to people and having it end up in a thrash can, seems really similar.

[maratd]

> Nowadays it starts to provide brains, which makes it replace services. As it continues, where will the human labour go?

Is there actual evidence of this? As in, actual statistics rather than grandiose claims of AI destroying the human race?

Yes, you might get replaced by a piece of software and you may need to retrain to a different industry, but this has always been true.

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

> The number of jobs was highly correlated to the size of population so far, but we already see the rise of bullshit jobs

This doesn't make any sense. At some point, someone is handing over money and expects a specific amount of work be done, which he finds to be a good exchange. It's far easier not to hire someone, than to hire them. So if you don't need work done ... you don't hire anyone.

Why would someone pay someone else money to perform "bullshit"? You only have instances of this when it is mandated by regulation, not in a free market.

[TeMPOraL]

> Is there actual evidence of this? As in, actual statistics rather than grandiose claims of AI destroying the human race?

I'm not talking about AI, I'm talking about replacing people in service jobs who were needed for some particular function provided more by their brains than by brute strength.

For instance, consider self-checkout machines replacing casheers in shops. It's already happening, it makes shops employ less people. Not zero, because automation is not a binary process, but less. (Also, n=1, I used to do inventory in a shop as an external contractor, because they couldn't count up everything in one night with the staff they had; that ended pretty much overnight when they got themselves automated inventory trackers with barcode scanners, which reduced the number of people needed by half.)

So is automation replacing people in services? Yes, you can see this everywhere around you. Do those people end up jobless? Not yet.

> Why would someone pay someone else money to perform "bullshit"? You only have instances of this when it is mandated by regulation, not in a free market.

Because it's not obvious bullshit, per leaflet example I described. The nonsense nature of work done only becomes apparent when you look at the whole chain of services provided. At each link, someone is paying someone else for a job they need to get paid by someone else entirely. And as long as you get paid, why would you refuse the job, or care what happens with the artifact you've built later? So there you have, some chains of services are direct mind-to-thrashcan work. Others run against each other, in a zero-sum-game, each cancelling out the results of another one. They exist, because at each step someone is paid to do something that the other party needs.

[maratd]

> For instance, consider self-checkout machines replacing casheers in shops.

In my experience, half the machines are out of order. The other half are working, but they need a person to stand there full-time to help the customers with items that don't scan and make sure that people don't just walk out of the store without paying. You also need someone to maintain those machines. If I had to guess, I'd say they don't really save the store any money since they still end up employing the same number of people. I'm also starting to see less of them than before, suggesting they were more of a fad.

> Others run against each other, in a zero-sum-game, each cancelling out the results of another one.

I think I understand where you're going with this. A competitor engages in a behavior that results in a benefit so long as nobody else is doing it (ie leafleting), but once everyone is doing it, it benefits nobody, but you still have to keep doing it because everyone else is?

That's not leafleting. That's marketing as a whole.

I suppose you can say that marketing/advertising is a bullshit job (and you wouldn't be the first), but that's just a fundamental outcome of human nature and capitalism. It's always been around and doesn't have anything to do with technology.

If anything, technology gave us things like AdWords, etc., automated ways of marketing. I imagine there are fewer people employed in advertising/marketing today than before largely because of those technological changes.

[TeMPOraL]

Yes, I count most of marketing industry as a big pile of zero-sum games. As for what it has to do with technology, I noticed people replaced out of service jobs by machines (either directly or through effectivity gains) to migrate towards jobs in ad industry. So yesterday's shop clerks are today's Social Media Marketers and/or graphics designers. In this way I see marketing as an ever growing sponge sucking people in and giving them bullshit jobs.

[Futurebot]

Yes, there is evidence, but you don't need AI, per se. Just automation in its various forms (robots, software, etc.):

http://www.economist.com/news/briefing/21594264-previous-tec...

http://www.npr.org/2011/11/03/141949820/how-technology-is-el...

http://www.nytimes.com/2011/03/05/science/05legal.html?_r=0

Automation is moving up the skill / knowledge chain.

[maratd]

That's not evidence. Just more hand waving. Those are ARTICLES. Where is the DATA?

[Futurebot]

The articles reference studies and books which have the data. See work by Ford, Autor, Acemoglu, Autor, Brynjolfsson. If you want to read just one thing which has plenty of data, read the book "Race against the machine" - they talk all about the studies used to come to their conclusions:

http://raceagainstthemachine.com/

[maratd]

Thank you for the reference, I'll give it a read!