I've been literally saying this for years. Yang had the right idea, and every year it becomes more obvious.

Most industries will be automated within 10-20 years, or so transformed as to be unrecognizable. I'm not just talking about stuff that can be done by AI art tools or a future version of ChatGPT (which is in itself a large chunk of the economy). There's also logistics (self-driving trucks, trains, taxis, planes, and boats), there's food service (burger-flipping robots are already a thing), there's groceries (robot stockers and self-checkouts), there's hospitality (Japan already has automated hotels), there's construction (we already see robots at construction sites), etc.

Within a couple decades I see no reason why these jobs would still be commonplace. Compare the world of the 1980s and 1990s to the world of today. Computers in the 80s and 90s are comparable to AI today, and in 2040-2050 I see no reason why we wouldn't be living in a completely different world.

It's true that some jobs will simply transform - programmers might become prompt engineers, for example. But many jobs will be eliminated completely, and I don't buy the argument that people will just find new things to do. At a certain point, people will be automated out of the economy - to borrow an analogy from CGP Grey, the invention of the stagecoach may have been great for horses... but the invention of the car was less great.

I firmly believe that UBI is the only way forward, long-term. COVID already worked as a trial for it, and we're seeing the economy contract in part because it stopped (in addition to other things, e.g. the Fed).