[throwaway4good]

OpenAIs models are so powerful and so close to AGI that the global balance of power would flip if autocratic regimes like China were to gain API access to it.

That’s what I am getting from the geopolitical safety researching marketing department.

Anyway as far as I can tell OpenAI has been turning access on and off for a while. And oddly enough now have decided to turn it off completely just as plenty of credible alternatives are emerging.

[malakai521]

Close to AGI? It's a different type of search engine (hard to say if better or worse) and has nothing to do with intelligence

[throwaway4good]

I was trying to be tongue in the cheek - my point is emphasizing dangers and restricting access can act as a form of marketing.