I too believe that there will be a future paradigm vastly more efficient than LLMs + RLVR.

This is part of the reason that I believe "AI Governance" would actually shorten timelines! Instead of spending vast amounts of money training bigger models AGI labs would be forced to innovate due to the compute limit...

Elon on why he became an accelerationist: https://x.com/SawyerMerritt/status/1935809018066608510

Sorry for your loss.

I believe that if we stop AI research then everyone alive today will be dead by 2150 with near certainty. LEV without AGI does not seem realistic at all…