A new paper describes why "AI" is very close to a hard mathematical limit to it's ability to get better. Must watch.
Maybe AI bros are calling for limits to protect them from open source competitors that will inevitably catch up now that they are hitting a wall.
www.youtube.com/watch?v=6xQ8...
youtube.com
MIT Just Found a Hard Limit in LLM Scaling, and Money Won't Fix It
YouTube video by Dan Grib