I'm a little more cynical. I think the AI companies are making a fortune selling the fiction that human workers can be replaced.

Megacorps invest for the short-term gains (wage suppression via layoffs that are spun to be a positive thing) and avoids short-term losses (holdouts may lose investors because AI is hip).

There is also a growing cadre of these people who, I think, truly and deeply believe Peter Thiel's hype about making a godlike super AI. Silicon Valley has been due for a religious awakening for years, and a techno-god that you can, theoretically, entreat with personally from the safety of your browser window? It's tailor-made. At the very least, I feel like a lot of them have fallen into Pascal's Wager about the whole thing, and that's enough for them to devote time and energy and the limitless money they can beg or borrow to this.

FINALLY someone gets it without me having to explain it.

And its clearly not working

Imo it's all for the government surveillance state. They already have terabytes upon terabytes of data, they just can't really use it. But an AI that could sift through all that shit in 5 minutes? Sounds great.

But I suppose both paths lead to AI companies getting bailed out.

My concern is a mass depopulation when there are sufficient robo-slaves to take up the slack of maintaining society and producing technological progress. Although it'll likely be temporary. I'm fairly sure humans when digitized will grow in number or similarly when we start space exploration and colonization in earnest. Without FTL some faction could always just go somewhere else if they want to do society another way.

They are so fixated on this that they are totally blind to that even if this were so it would remove the whole foundation of business: having customers able to pay, at all.