When robots learn to draw silly pictures (or make the economics of doing so untenable), then my world will be shaken asunder:
This has been your fear-for-the-future of today.
I’ll be worried when they figure out how to automate government.
We’ve really needed to realize we are in a largely post-scarcity society for a while now. The solution to the problem presented in the video is simple: Humans don’t need to work. There’s no magical need to have people have jobs, as long as they are fed, housed, happy, and not causing trouble to society as a whole.
Uncouple people’s welfare from their occupation, and the problem goes away. It’s a major shift in thinking, but it’s not impossible.
It’s figuring out HOW to uncouple welfare from occupation that will cause issues. There will suddenly be a divide between those who are able to work and those who aren’t. To most people it will seem fair that those who are working, i.e. those who are contributing to the functioning of the machinery of society, getting more of the abundance that permeates society than those who aren’t. This creates an instant divide, rather like the one that currently exists between the rich and the poor only more so because those not working will be superfluous to society.
So how do you deal with the majority of a society no longer needing to work for 8-16 hours a day in order to survive? What will they spend their time on? Mac Reynolds wrote several novels about it back in the 50s and 60s.
Will we invest these people in the arts? No, the robots will be doing that.
Will we invest them in the sciences? Nope, robots.
So they’re basically sitting around, doing nothing, for the benefit of society. If you can figure out how that doesn’t end in disaster, you’re smarter than I am
The difficulty is that as long as some things that people want are scarce, there’s still a need for currency with which to buy those things, and a need for ways in which to earn that currency. Adapting to a more automated society could in principle be done by revaluing labour (people working a 4-hour work day instead of today’s 8-hour day, just as we no longer work the original “dawn to dusk” working day), but in practice that’s not going to work.
The problem that the video pointed out, that I hadn’t noticed before, is that quite a lot of people will be intrinsically unemployable if AI takes off to the degree that the video predicts. This means that even in an economy that still has scarcity, there will be a large number of people who have to be permanently supported by society rather than being able to earn incomes of their own. In a society that still has competition over scarce resources (like income), that’s not going to be an easy sell. It’s a hard enough sell up here when it’s supposed to be temporary.
Regarding whether or not AI can get that good without becoming full-blown human-equivalent… I’m on the fence. On one hand, there are a number of powerful and flexible approaches to making expert systems that can solve problems once you nail down the problem that they’re trying to solve, but describing the problem and the criteria for evaluating solutions in a way that a machine can understand is still something that requires quite a lot of work. An analogy that came up earlier was formal proofs of correctness for software: sure, you can do them, but asking a computer “prove the game Quake works” requires a question that’s about as complex as the game’s source code (as opposed to easier questions like “prove Quake has no memory leaks”). So, the range of jobs that are threatened is still murky (though definitely enough to be a problem).
AI that’s independent enough and creative enough to answer questions like that on its own is its own major problem, but that’s another discussion.
Admitted some things are still scarce, but the essentials aren’t: Food, housing, clothes, medical care are all automatable at the basic level. Define the rest as ‘luxuries’, and start thinking. 😉
My proposal would be to give *everyone* a stipend equal to some basic level of the basics. You want a luxury, you have to buy it, either by not using all of your stipend or by finding some other way to generate the income.
I know, it would be a hard sell, but in the long term I think something like that is probably inevitable – either that or a luddite war, where we destroy all our own tech.
My point is that it’ll be a “hard sell” to the point of civil war.
Take a look at the liberal/conservative arguments up here (or the democrat/republican arguments in the US). We have people very strongly objecting to “their” money being spent to support other people as it is – trying to have a substantial fraction of the population permanently supported will prompt open revolt.
The expectations for what “luxuries” and “non-luxuries” are will also shift. This has already happened repeatedly: cars were a luxury 100 years ago, and computers were a luxury 20 years ago. I’d expect the “minimum acceptable standard of living” to require goods and services that are scarce, and so require significant resource investment from the still-employed to provide. You’d either have to provide them, provoking revolt amongst the employed, or you’d have to make the unemployable second-class citizens (with a standard of living that the employed would consider unacceptably low), which would provoke revolt amongst the unemployed.
It’s a lose/lose scenario.
I’d be overjoyed if that didn’t happen and we end up with a smooth transition, but I really doubt that’ll be the case. I agree that a change of this sort is inevitable (barring strong AI or other game-changers) – what we appear to disagree about is the amount of social turmoil that will result.
Nah, we don’t really disagree. We just started with different ideas of what this conversation was about. 😉 I was focusing on the technical problem, not the social one of getting the technical one accepted. (Which I agree is going to be a major issue – and something we really should be working on. Which was kinda the point of my first post.)
I can see how this would be a problem to humans who want to work hard and breed freely, because in the future humans will become pets to robots. Their every need will be taken care of by robots who are their betters in every way except in ‘being human’.
This is my main issue with the Matrix. Why make up some ridiculous and physics-defying idea of using humans as a source of energy when it would’ve made perfect sense if humanity had been kept safe and alive because that’s the goal the robots had been programmed to achieve?
The Matrix would’ve made more sense if they were using the humans’ brains as a distributed computing network to compensate for damage inflicted during the war. And the system was trying to get the humans to come up with a solution to the permanent clouding thing.
Other than the effects, it’s really a rather blah movie. Precisely because of things like this.
That’s pretty much exactly what the humans were supposed to be used as in earlier versions of the script; you can still find references to it in the supplementary material, reportedly.
I doubt that human brains would’ve been worth the effort. Sure, they could’ve gotten some computing power out of them, but enough to be worth all the infrastructure to keep those humans alive? Enough to cover the computing power needed to run the Matrix in the first place? Seems to me as if that would’ve run into the same issue as the harvesting of body heat: They would’ve had to put in more than they could possibly have gotten out.
That depends on what you are getting out of the human brains. If ‘intelligence’ is just a matter of throwing computing power at the problem, then yeah. If there is something in the human brain that allows sentience/intelligence that can’t be duplicated in silicon, then humans become an irreplaceable resource; something you can’t do without.
Of course, there’s no good reason for keeping the *rest* of the human bodies around, or even keeping them in condition to move…
©2011-2015 Aaron Williams | Powered by WordPress with ComicPress
| Subscribe: RSS
| Back to Top ↑
Bad Behavior has blocked 7761 access attempts in the last 7 days.