It will be run by a group of relatively few elites that are involved in power struggles for control of the AI and the rest of us that do stuff like pick fruit (assuming it's cheaper than having an AI do it) and serve as cannon fodder in the AI armies.
I find people who believe that it will lead to some kind of Utopia horribly naive and they have no understanding of human nature.
The lucky ones will be shining their shoes, giving them massages, tutoring their kids, cooking their meals, and giving them manicures.
Many will prefer the human touch for things like that, even if they could be done by machines, and having human servants has always been a status symbol.
This is what the non-wealthy among us have to look forward to.
You have used technology right?
You know it is all mostly shit that breaks down, crashes, and gets replaced and thrown out on an 18 month cycle?
What I think lots of people get wrong is forgetting just how crappy consumer tech is, and industrial/military isn't far behind. We may get AI working well in 5-25 years, but I doubt deployments will be any less buggy.
Repairing & provisioning the billions of bots will be most of the "tech" jobs. Think like the IT dept at your company that hands out new laptops, times a million.
For the first, you are right - certainly in that timescale. For the later, I assume we will need to justify our continued existence.
It is extremely well defined. The best example I can point you to is from the 1990's by Russell & Norvig in their seminal work "Artificial Intelligence" which coalesces a half century of research. They spend the first 100 pages clearly defining the types of domains and vocabulary needed to define the space of AI: it is an exhaustive analysis with citations going back decades, covering early thinkers like Norbert Wiener and Von Neumann. Your claim indicates you haven't really studied the field. But everyone jumping into tensorflow on a raspberry pi has probably had zero exposure the to decades of research put into AI.
Your second claim about a giant mind that might dwarf all of humanity, is, well, grandiose sci-fi. Much like Dyson Spheres or Time Travel.
...simply because even a reasonably intelligent AI would have the ability to improve itself at a rate orders of magnitude faster than biological evolution.
Actually, there is a larger and larger gap between educated people with office jobs who earn a good salary to do "stuff that hadn't been worried about before"... and the rest of the barely educated people who barely earn anything and get to do all the manual work (taking care of the children, cleaning homes, offices, cooking food, delivering food, harvesting delicate fruit, etc.). The gap is bigger and bigger, despite or thanks to the progress of humanity.
In a world where less of the manual people are asked to do manual work for some reason, these uneducated people won't transition to other jobs. Best case scenario is they get a minimal universal income to stay calm at home and generate clicks on social media to feed AI some new data. Worst case scenario is they will want to work, and be angry at society for making them feel useless.
It's cyclical, and while I will concede that people lose their jobs because of automation, zooming out to the overall jobs market (and figures, which probably don't tell the whole story) things haven't really changed much.
On an individual level though, people lose their jobs and livelihood, and have to settle for a lower paying, dumber job.
That said, a lot of jobs replaced by automation are already low-paying, dumb jobs. Why stand 8 hours / day doing a repetitive manual job if a thing can do it just as well? Why expose yourself to the worst the internet comes up with when you can train an AI to detect CP?
I don't believe we'll see jobs becoming largely obsolete in our lifetime, but I would like to fantasize about a future where work is a choice and everybody can live comfortably (quality housing, food and spending money) even without a job. Not going to happen as long as capitalism and the never ending pursuit of profit and profit increase are still a thing though. But, we're seeing some hopeful developments in that regard, with things like UBI and the like.
(Aside: I don't believe UBI will work in the end, if everyone has a baseline income, nobody does. UBI only works if it's coupled with assurances of housing and that it's enough to cover all basic needs, which right now depend largely on the area you live in. I couldn't get an affordable rental place on my own income some years ago, because first there was a 15+ year waiting list, and later I was earning too much so I was thrown out to the 'free market', and there's a pretty big gap between the most expensive social housing and the cheapest free market housing. That gap is filled up with 'student' houses, think houses split up into multiple units / rooms. But nobody should be forced to live like that if they're out of college and have a paid full-time job.
I think you are assuming an auction like situation for stuff like limited housing but it doesn't have to be like that. You can build more.
I believe the argument for UBI is based on economic utility: $1000 a month is worth more to a poor person than a rich person.
I want UBI to work. But I haven't heard of a good strategy to mitigate what I just described.
what we don't generally have is an optimal dispersion of capital (and power) to fulfill many of those wishes. concentration of capital and power also concentrates labor allocation and decision-making (something capitalism in ideal form opposes), so instead of drawing from the ingenuity of hundreds of millions, we're trending toward that of hundreds of thousands. it doesn't matter how clever those hundreds of thousands are, they'll never out-imagine hundreds of millions (or billions worldwide).
ubi is predicated on the idea that lots of people will have nothing to do soon, but that's only effectuated in a dystopian world where capital and power are extremely concentrated. rather than altruism, ubi is rather a critical step toward fascist corporatism, acting to stave off potential opposition by trickling down a tiny bone to the masses. the populace assuming mass joblessness are falling into the rhetorical trap of the subjugating corporatists, not accessing a benevolent key to economic and social freedom.
Right now few bright folks are interested in revolutionizing affordable housing, public education, or payday lending because the money's not in it. They make a lot more money advertising to the affluent class or owning a marketplace that gives them vacation rentals or cheap transportation.
If we had UBI, prices would eventually equilibrate to eat all the new disposable income, but the gold rush in the process might build a bunch more halfway-decent housing.
As far as real cash payouts go, there's value to them, but not universally. In order to have the maximum impact they need to be limited to people with little to no other income.
AI can also help augment the work of “high-paying, smart jobs”.
Meaning that instead of hiring ten lawyers, you can hire one lawyer and a license to lawyer.ai which will empower the one lawyer to do the work of ten.
If you have automated cars tomorrow on the streets, most of the truck and taxi drivers are not ready. Heck, most of the oil countries in the middle-east are not ready for a post-oil economy.
Population capacities and thus overpopulation aren't fixed things meaning overpopulation is relative to conditions. It would be absurd to call algae overpopulated if they thrive just fine at concentrations in one epoch with different population conditions just because it would cause red tide in another epoch.
If you live in the USA and Europe, you are the recipient of generations of global technological dominance. It's hard to understand the possibility of being on the other side of that.
AI is all nice and rosy but how do we know it's "right"? How can a machine be indifferent to human influences? How do we know it's not biased? Does it discriminate? It's a not a human, it can't "think" like one, it's not capable of rational and fluid thought and argument. It's a dumb machine feeding off data, unless you give it all the data it's bound to be inaccurate at some point.
AI is given too much credit for a little bit of intelligence (if you can call it that, I think that's how we perceive it), however as society changes so the AI has to. Does AI understand brexit yet? Does it influence policy decisions? Can it understand how people will react? Can it help during a pandemic? Can it talk and negotiate?
Brexit was dumb and so is the UK.
Also:
> it's not capable of rational and fluid thought and argument
applies to most humans, and
> unless you give it all the data it's bound to be inaccurate at some point
applies to all humans.