> the resources (building materials) and the energy to power the robots or whatever, will be more expensive and tightly controlled than ever.
I am also concerned about this possibility, but come at it from a more near-term problem.
I think there is a massive danger area with energy prices specifically, in the immediate run-up to AI being able to economically replace human labour.
Consider a hypothetical AI which, on performance metrics, is good enough, but is also too expensive to actually use — running it exceeds the cost of any human. The corollary is that whatever that threshold is, under the assumption of rational economics, no human can ever earn more than whatever it costs to run that AI. As time goes on, if the hardware of software improves, the threshold comes down.
Consider what the world looks like if the energy required to run a human-level AI at human-level speed costs the same as the $200/month that OpenAI charges for access to ChatGPT Pro (we don't need to consider what energy costs per kWh for this, prices may change radically as we reach this point).
Conditional on this AI actually being good enough at everything (really good enough, not just "we've run out of easily tested metrics to optimise"), then this becomes the maximum that a human can earn.
If a human is earning this much per month, can they themselves afford energy to keep their lights on, their phone charged, their refrigerator running?
Domestic PV systems (or even wind/hydro if you're lucky enough to be somewhere where that's possible) will help defend against this; personal gasoline/diesel won't, the fuel will be subject to the same price issues.
> Power and wealth simply wont allow everything to be accessible to everyone. The idea that people would be able to build enormous mansions (or personal aircraft carriers or spaceships) just sounds rather absurd, no offense, but come on.
While I get your point, I think a lot of the people in charge can't really imagine this kind of transformation. Even when they themselves are trying to sell the idea. Consider what Musk and Zuckerberg say about Mars and superintelligence respectively — either they don't actually believe the words leaving their mouths (and Musk has certainly been accused of this with Mars), or they have negligible imagination as to the consequences of the world they're trying to create (which IMO definitely describes Musk).
At the same time, "absurd"?
I grew up with a C64 where video games were still quite often text adventures, not real-time nearly-photographic 3D.
We had 6 digit phone numbers, calling the next town along needed an area code and cost more; the idea we'd have video calls that only cost about 1USD per minute was sci-fi when I was young, while the actual reality today is that video calls being free to anyone on the planet isn't even a differentiating factor between providers.
I just about remember dot-matrix printers, now I've got a 3D printer that's faster than going to the shops when I want one specific item.
Universal translation was a contrivance to make watching SciFi easier, not something in your pocket that works slightly better for images than audio, and even then because speech recognition in natural environments turned out to be harder than OCR in natural environments.
I'm not saying any of this will be easy, I don't know when it will be good enough to be economical — people have known how to make flying cars since 1936*, but they've been persistently too expensive to bother. AGI being theoretically possible doesn't mean we ourselves are both smart enough and long-lived enough as an advanced industrialised species to actually create it.
* https://en.wikipedia.org/wiki/Autogiro_Company_of_America_AC...