We arent really close to AGI

We're not really close to AGI. People are very focused on the technical benchmarks but just wait until the resistance starts hitting back around Legality. The US for instance is super litigious. Wait until the lawsuits start pouring in for malpractice. Wait until the big problems with generated code make the news. This whole thing could shift overnight and the media will have a feeding frenzy over it. Assume somehow none of that happens either. Even AGI now doesn't mean nearly as much as people think - yes it would be big but still have a long way to go and then there's robotics needing to catch up which we're quite far from.

The truth is we don't know and it's very easy to draw up some plausible sounding scenario but look at Yearly predictions for any year you've been alive and compare them to what happened in the end. Even when things are directionally correct, in the details they are much much different. It's just too early to tell. Fortunately there's a lot of time and yes, its advancing faster than society but there's still a lot of manual effort behind the LLMS and I'll just say, maybe I'm naive but I've been working in AI since 2005. I'm working here now. At my age I really doubt I'll live to see any of the crazy stuff that's commonly predicted now. Maybe some version of it but not comprehensively. Way too early to tell