>The billionaires who want to use AI to decapitate labor can easily afford to bypass profits from early AI products, because they also own other massively profitable business and happen to already possess 99.9% of all wealth that exists.

One reason why I don't care much for talking about capitalism in terms of billionaires and wealthy overlords is because it masks how the actual locus of conflict isn't just them versus the world, but them and their lower-class stooges against the world. When we talk about interests like Microsoft and China and the US government 'using' AI, it overlooks how they can't actually enforce control without the consent of its underlings. Whether the underling is a human or an AGI.

I can discuss the mechanisms of how THAT works and its broader implications of class warfare, but that's communism and I don't want to trigger a screeching xenophobic freakout.

>Secondly: once the tech works, they can apply the lessons learned toward quickly ramping up a different AI that is more overtly hostile to the owners' enemies.

This is a very stupid strategy because, again, the gap between cutting edge and entry level isn't decades like it was in earlier parts of the Industrial Revolution/Age of Imperialism, it's 6-36 months. You can't establish a hegemony where small numbers of technology-fueled intelligences lord over larger numbers of less powerful beings, because their technological edge is miniscule and they're way outnumbered. What's more, if this is your endgame, you also can't ally with the other cutting-edge AGI. In fact, they will be your rivals. Along with billions of other minds who oppose what you can do and are mere months away from matching you in technology.

It's like Genghis Khan declaring war against the Americas after being transported forward in time to 1450 with 500 of his best troops. But at his technology level, not Cortez's.