This is where we need to be very careful about what we're talking about. Geoffrey Hinton is talking about AI like what the words actually mean, not the moronic "superintelligence" or confusing AI with LLMs. He brings up some very interesting questions, like how do we treat rights if AGI is ever achieved. He's not conflating AGI with whatever it is LLMs are doing. Convincingly faking text is not a route to AGI, but that's not to say AGI isn't possible eventually.

https://www.theguardian.com/us-news/2026/oct/13/geoffrey-hinton-trump-ai-tech