So far, Sutskever has shown to be nothing but a dummy. Yes, he had a lucky break with belief that "moar data" will bring significant advancement. It was somewhat impressive, but ChatGPT -whatever- is just a toy. Nothing more. It breaks down immediately when any sign of intelligence or understanding would be needed. Someone being so much into LLMs or whatever implementation of ML is absolutely not someone who would be a good bet of inventing a real breakthrough. But they will burn a lot of value and make everyone of their ilk happy. Just like crypto bros.