Ex-OpenAI Researcher: AI Agents May Soon Design Better Models Than Humans

A former OpenAI researcher has publicly stated that AI agents could soon surpass human researchers in designing new AI models — a claim about recursive self-improvement that was once considered speculative.