I want to piggyback off this to talk about the inevitable Uber comparisons, because not only is the mismatch between investment and returns several orders of magnitude greater, but there's also a difference in kind. Uber's model was to undercut the taxi industry and establish a dependence within their niche before increasing revenues. It's the classic enshitttification cycle. But the AI plan, at least as advertised, isn't to undercut a specific industry as much as it is to undercut literally the entire white-collar labor force. There are several problems with this, starting with the fact that the technology isn't actually able to replace the target in the way it would need to. More significantly, however, is that labor doesn't work like taxis. If labor can't get work it shuts down the entire economy because they lose their income and can't actually consume any of the things the market offers. Also labor tends to get mad and break out the pitchforks and molotovs if things get too bad, and "restructuring the economy to no longer provide you the means to sustain your family" seems like the kind of situation that definitionally makes things too bad. In either event the point is that even if this tech is somehow as revolutionary as advertised then there's not really any winning for the company.