3 ms·
As a swiss I'm quite excited about this. Even if I'm pessimistic about a successful vote. I want to mention a side-point which I haven't encountered so far: In
by iliis 13y ago
As a swiss I'm quite excited about this. Even if I'm pessimistic about a successful vote.
I want to mention a side-point which I haven't encountered so far:
In the 'problem of lazyness' (i.e. "who will do the boring work if you are paid anyway?") one answer that often comes up is automation. Menial/low-skill jobs will be done by robots.
It's a common image that in the future robots will clean our toilets and grill our burgers while the humans have ample time to enjoy life, build rockets and do fulfilling work.[1] However, if AI research has told us one thing it's that this view is quite wrong - the opposite is true. Things humans think of as 'easy' like walking around or doing laundry are actually very difficult engineering problems and 'hard' stuff like playing chess or doing rocket science are easier in an absolut sense.
The reason behind is simple: Evolution had millions of years to perfect walking on legs and interacting with the environment. Our bodies almost literally have an walkTo(Place) API, so the only thing 'we' (our conscious self) have to do is to call it. Abstract thinking is something new and our brains have to do it manually. It takes you less time to run over debris than to multiply large numbers and so we have a distorted picture of what is hard in an absolute sense.
The picture is true for humans but it falls apart when we want to automate it. We are still decades away of beating humans in soccer but Jeopardy is a solved problem.
I'm not convinced that 'low'-skill jobs are the first to fall into the hands of machines. [2]
---
[1] Ignore for the moment the fact that unclogging toilets can also be fulfilling.
[2] This is of course a broad argument. We still have no clue what Intelligence or Creativity actually is - It may still turn out to be very complicated. But I nevertheless think there's a good chance that we will figure out hard AI before you can buy a robot which washes your socks, mows the lawn and fixes your plumbing.
- polarix 13y agoSo, if the low-skill jobs stay around, but nobody wants to do them, their price increases, because nobody is forced to do them to stay alive. Then we have a greater incentive to automate and deprecated the immorally unfulfilling jobs.
- iliis 13y agoYes, the economic argument looks solid. Good old supply and demand. Problem is, we may not be able to automate these jobs so easily.
- PeterisP 13y agoWhile we're not able to automate them, the labor costs for them will rise - and that's a very good thing. I've looked at business process automation analysis in a few industries, and I've seen a TON of mind-numbing things that we actually would able to automate, but it's not worth to try since a bunch of low-wage people is cheaper. In some cases it's not even a question of "how to automate" - simply purchasing well-known machines would do that, but people were cheaper. It would be oh so great if that would change - there are a lot of jobs that IMHO are below humans as such, which are so stupid and soul-crushing shouldn't ever be done by a sentient being, but we as society force them to do so because they want to eat. People can handle shit-jobs - they "turn off", or socialize with coworkers and power through it in order to get their paycheck and come back to their "real life" afterwards - but in the long run, people shouldn't be forced to do that, not anyone, not ever.
- marcosdumay 13y ago> However, if AI research has told us one thing it's that this view is quite wrong - the opposite is true. AI research has told us a lot of things in the last decades. We are in a point where we have a quite good idea on how hard it is to automate most tasks, and robots that clean toilets, build rockets and drive cars aren't very far away. But robots that decide that a rocket should be built or a toiled should be cleaned are.