>>2683620the problem with your premise is underrating the limits of the technology.
when Concorde or the TU-144 left the ground, you would've been here posting about how any anti-supersonic transport leftists were just liberals. and, indeed, the class-character of the anti-SST movement was suspect but… they were also right about the economics and the physics of the thing. concorde is by many measures one of the worst investment decisions in human history. and unlike AI, it actually delivered on the technological promises it made instead of overhyping and underdelivering!
like SSTs, AI also finds that its real world use case doesn't match up to the amount of investment poured into it. the tiny profit British Airways made operating concordes which they were handed for basically free never came remotely close to repaying its development costs, and (like 99% of AI firms) Air France didn't even make an
operational profit. Then, within the life-cycle of the airplane, the internet came along and made even that niche unnecessary by severely cutting the amount of times you've got to urgently dash off to America or London. To save the rich and famous about 6 hours a day and to give France and Britain (mostly Britain in my experience) a little bit of national pride at an engineering marvel, £1.3bn in 1960s-70s money was spent. It is fairly unambiguous that £1.3bn could've been better spent on basically any other project. The US somehow did even worse, spending $1bn without even getting a plane out of it.
If, in 30 years, we find that the primary use case for LLMs is erotic roleplay, 75%-accurate text summaries, and writing e-mails that don't matter, will that be worth the $1.6tn of capital investment poured into it? is it "retarded liberal doomerism" to suggest that $1.6tn of housing, $1.6bn of transport infrastructure, $1.6tn of NASA spending (they currently get 24.4bn/year!! with a
b!), hell, $1.6tn (=~ 10 years) of food stamps, would have been a smarter bet with a higher return on investment?