I think also how, after Feb 2025’s Paris AI safety summit, it looks like the world has given up on AI safety and has decided to splinter into many different sovereign models all doing their own thing. Coz the leading colonials (OpenAI etc) suck and can’t be trusted. But that they’ll fundamentally end up doing similar things to what postcolonial leaders did. Inherit the basic structure of oppression, install ‘local’ leaders, and call it a win.