What evidence is there that gen AI hasn't peaked? They've already scraped most of the public Internet to get what we have right now, what else is there to feed it? The AI companies are also running out of time--VCs are only willing to throw money at them for so long, and given the rate of expenditure on AI so far outpaces pretty much every other major project in human history, they're going to want a return on investment sooner rather than later. If they were making significant progress on a model that could do the things you were saying, they would be talking about it so that they could buy time and funding from VCs. Instead, we're getting vague platitudes about "AGI" and meaningless AI sentience charts.