Google says self-improving AI isn't anywhere close

What happened

Google engineer Haydn Belfield says Google DeepMind has seen no measurable AI-driven acceleration of AI research despite substantial automation. AI writes more than 75 percent of production code, but every change receives human review before submission; Argon completes about half of RE-Bench's 74 ML engineering tasks. Google's General Capability Index shows steady Gemini capability growth without statistically significant acceleration, contrasting with researchers' warnings of a possible intelligence explosion.

The story so far

Google DeepMind has seen no measurable acceleration of AI research driven by AI itself despite substantial automation, according to Google engineer Haydn Belfield. Belfield said AI writes more than 75 percent of Google’s production code, but every change receives human review before submission. Google’s Argon system completes about half of RE-Bench’s 74 machine-learning engineering tasks. Google’s General Capability Index shows steady growth in Gemini’s capabilities without statistically significant acceleration. Those findings contrast with researchers’ warnings of a possible intelligence explosion: the reported automation and capability gains have not translated into measurable acceleration of AI research at Google DeepMind.

Keep reading

Trending now

1. 1Meta Releases Muse Spark Open With Mac and Cloud Support, Topping US Open-Weight Index
2. 2Google Launches Gemini 4 Argon Globally, Expands Paid Access as Benchmark Results Vary
3. 3OpenAI Updates GPT-6 Astra Pro in ChatGPT With Parallel Agents and Faster Responses
4. 4Meta’s New Business Terms Raise Dispute Over Website Crawling and AI Training Rights