Right, well even without AGI (no two people agree on whether it's coming within 5 years, 30, or 100), finely-tuned LLMs can disrupt the economy fast if the bottlenecks get taken care of. The big one is the robot-economy. This is popularly placed further off in timescales, but it does not require AGI at all. We already have humanoid robots on the market for the price of a small car, they're just dumb. Once we scale up solar and battery production, and then manufacturing, it's coming for menial labor jobs. They already have all the pieces, it's a foregone conclusion. What we don't know how to do is to create a real "intelligence", and here the evangelists will wax about the algorithms and the nature of intelligence, but at the end of the day it takes more than scaling up an LLM to constitute an AGI. The bet is that AI-assisted research will lead to breakthrough in a trivial amount of time.

With white-collar jobs the threat of AI feels more abstract and localized, and you still get talk about "creating new jobs", but when robots start coming off the assembly line people will demand UBI so fast it will make your head spin. Either that or they'll try to set fire to them or block them with unions, etc. Hard to say when because another effort like the CHIPS act could expedite things.