This is our monthly place to discuss relevant topics that have not appeared in recent posts.
Confirmation that cave paintings on Sulawesi are at least contemporaneous, and likely older, than the earliest paintings in Europe is a major discovery.
The most parsimonious explanation is that early Homo sapiens coming out of Africa already possessed artistic abilities and the self-consciousness and creativity that implies.
It’s hardly a new thought, but on Friday Larry Page said in an interview (http://www.ft.com/intl/cms/s/2/3173f19e-5fbc-11e4-8c27-00144feabdc0.html#axzz3Hpz5nKIG):
“Wouldn’t the world be a happier place if 90 per cent of the people with jobs put their feet up instead and left the robots to do the work?”
And Robin gave a talk the other day on “I Robot, You Unemployed”.
Page should know as well as anyone how close we are to automated systems that can really replace the majority of jobs.
Assuming that starts happening in the near future, what do people here think is the best policy response?
Despite all the hue and cry over “unemployment” , I think few people really want to work (at other people’s jobs, anyway) , but they certainly want the income from work.
I suspect the high productivity resulting from automation will create gigantic returns on capital, so those who own even a little capital today may be well off once this starts happening.
But most people live paycheck-to-paycheck, and a large fraction (majority?) are in debt. What of them? I can imagine huge social disruption, even revolution, if some kind of provision for these people isn’t made.
You could implement something like a basic income/negative tax. You could also implement policies that lead to shorter workweeks (higher pay per hour) as long as far as globalization allows (globalization should become less and less of an issue in the future). You could try to create as many new jobs as possible that cannot be automated (because it would be too expensive or because people don’t want them to be automated). You could also mix these options as much as you want to.
I think that when the time comes where artificial intelligence can do everything a human can do we should give them civil rights and that would mean they get the right to be paid a salary and would thus prevent complete automation of the economy.
Re shorter workweeks, I’m not sure that policies that make human labor even less efficient (compared to machine labor) improve the situation – if anything that would seem to make things worse.
A basic income policy seems more likely to help, if it were politically acceptable. But would most people end up with that as their _only_ income? And what would be the effects on work incentive during the transition to 100% machine labor (esp. for low-paid workers)?
How do AI rights “prevent complete automation”?
“Re shorter workweeks, policies that make human labor even less efficient (compared to machine labor) would seem to make things worse, not better.”
The idea is that there are jobs that are easily automated and jobs that are incredibly difficult and expensive to automate with not much in between. For the difficult-to-automate jobs even a four-fold increase in hourly wage shouldn’t make automation feasible and the goal is to increase employment in those jobs by spreading them out over more employees.
“A basic income policy seems more likely to help, if it were politically acceptable. But would most people end up with that as their _only_ income?”
That really depends on how many people you can keep employed in non-automated jobs.
“How do AI rights “prevent complete automation”?”
With rights they become a lot more expensive to employ and difficult to control (even if someone can employ them and make a profit there’s no guarantee their manufacturer will be the one making that profit).
Re shorter workweeks, policies that make human labor even less efficient (compared to machine labor) would seem to make things worse, not better.
What’s bad is the unequal distribution of the available labor, not the replacement of labor by machines.
I hear equalitarians arguing for equal distribution of good things all the time, but this is the first time I’ve heard one argue for equal distribution of bad things.
Few people want to work (at least on other people’s projects) – it’s money that people want.
If we can free most people from needing to labor to survive, without seriously harming other people, I count that as a win, not a loss.
I don’t so much disagree with your last paragraph in principle as I question it’s practicability. The working minority will not be likely to tolerate giving the masses of loafers a decent standard of living. A sliding scale of wages and hours seems the only socially acceptable solution.
I think it matters a great deal how much it costs to support those with no capital. If it is sufficiently small, I don’t think it’s a problem.
Of course that implicitly accepts a lot of wealth inequality; which I don’t think is a problem as long as the absolute standard of living is high enough. (Wealth inequality is pretty much inevitable no matter what you do.)
I can also imagine demands for limits on child bearing among those receiving subsidies (for fear of a Malthusian doom scenario).
Let us hope that Robin is correct; if so this problem won’t actually come up for quite a while.
You seem ambivalent:
“If we can free most people from needing to labor to survive, without seriously harming other people, I count that as a win, not a loss.”
“Let us hope that Robin is correct; if so this problem won’t actually come up for quite a while.”
Yes; radical change is risky – things can go wrong. I’d much prefer this happens gradually with lots of time to adjust.
(As if what I’d prefer matters.)
You have a strangely narrow idea of “wealth” and “cost”. If a working minority had to support a non-working majority than that would always cost that minority a lot in terms of the precious resource that’s called “leisure time” (what good is money if you do not have time to spend it?) I’m also curious why you consider working say 16 hours a week as a cost people would like to avoid, while they would be completely okay with not being allowed to have children (a policy which you couldn’t even justify on the basis of work ethic since it would be a society with a severe and chronic lack of employment so luck rather than work ethic would determine who gets to have a job and who doesn’t). There will also always be a status component attached to employment, especially if it means having a lot more income than non-working people, that gives unemployment a real psychological cost.
I agree with Stephen Diamond that the socially acceptable solution is a sliding scale of less hours worked as a percentage of lifetime to push more people into jobs that are very difficult to automate (or that society refuses to automate). I also think that 100% automation would not be a good idea: we’d be handing over our fates to the machines, if something happened to the machines or if they turn against us and we survive we’d be back in the stone age unless we maintained a reserve of subsidized human scientists doctors and engineers.
“I suspect the high productivity resulting from automation will create gigantic returns on capital”
Not necessarily. Just because a robot is cheaper than a human doesn’t mean the returns are gigantic.
If I were mega rich, I’d just donate 1% of my income to prevent a revolution. Pay poor people to eat fast food and watch netflix etc.
Bread and circuses. Won’t work out any better this time ’round.
Many of those poor would be very intelligent and ambitious people. I doubt bread and circuses would indefinitely protect you against a movement that seeks power of its own and draws its support from the poor.
My solution was meant as a response to the speculation that people who live “paycheck to paycheck” might revolt.
If you think there could be a different revolution, where a much smaller minority of intelligent and ambitious people revolt, it might warrant a different solution.
I heard smart and ambitious people like to play MTG though…
Forget the U.S. underclass. Think of the Third World, which shows our future. Do you really suppose that the masses of unemployed workers and impoverished peasants worldwide lack ambition and intelligence?
Ambition and intelligence are probably irrelevant.
North Korea exists. The USSR happened. Etc…
You can also automate guard labor. There are realistic technological solutions to keep all humans under control 100% of the time.
It’s whatever you want to do with your pile of resources. I’m just saying there’s at least 1 way to prevent a revolution due to economic exclusion of the masses.
I do like the thought of having a giant black castle on top of a mountain with hundreds of 10 ft tall robots patrolling.
My new blog entry, “Univocality, the highest stage of clarity” ( http://disputedissues.blogspot.com/2014/11/univocality-highest-stage-of-clarity.html ), is largely a polemic against Robin’s old post, “Against disclaimers.” ( http://www.overcomingbias.com/2008/06/against-disclai.html )
So you mean, like, “Eschew Obfuscation,” right?
Is Elon Musk talking complete nonsense?
“The risk of something seriously dangerous happening is in the five year timeframe. 10 years at most.” Seems bettable, if anyone can offer verifiable markers of such dangers.
There’s a rising new paradigm for physics that’s now got ‘Many World Interpretation’ (MWI) of quantum mechanics in trouble! If this new paradigm is correct, then the Bohm (pilot-wave) theory is correct, and MWI is out!
The big idea is that space-time is a super-fluid. Although this idea is not new, over the last decade, a number of physics theories (in cosmology and quantum gravity) based on the idea, have reached the point where they are fitting together coherently enough to say that a tipping point has been reached, and space-time as a super-fluid is an idea whose time has come.
What are the advantages of the space-time as super-fluid idea?
*Potentially offers explanations for dark-energy, dark matter and big bang
*Offers a way to unify gravity with other forces (quantum gravity)
*Black holes – theory of gravastars – avoids Singularities, properties of black holes explained as ‘phase-changes’ in the space-time fluid.
*Fluid dynamics – strong analogies to quantum effects , offers concrete model for the Bohm pilot-wave model.
I want to provide a link to more details about this last item , because its really quite striking – all the quantum effects such as wave inference etc. seem to naturally fall out of fluid dynamics, look here:
Now I’m not saying that this stuff is about revolutionize physics just yet, but its all reached the point where other physics theories are starting to come under a little pressure!
In the context of the ‘Less Wrong’ sequences and Yudkowsky’s claim that ‘Many Worlds is a slam dunk’, well the fact of the matter is that MWI is now coming under pressure from space-time as a super-fluid idea, and Bohm pilot wave is making a comeback!
… be a charity angel.