Expert Claims LLM Progress Has Stalled Except for Coding and Math

burkov · x · 2026-07-31

Countering the view that future AI will be mind-blowing, AI researcher Ruslan Burkov argues that LLMs are currently hitting a stagnation period similar to mature technologies like phones and cars.

He notes that the only areas where models are still substantially improving are coding and math, largely due to verifiable reward mechanisms. He pessimistically suggests that if true AI requires uninvented breakthroughs, it might remain uninvented for the next 50 years.

Original post →

More from AGI Musings

AGI Musings channel →