---
format: "aidr-story-markdown/v1"
id: "10c6be3ccf91e43bb49a450d6e7643f54c3a6137e8fd320a4a634af63d9d1dd8"
canonical_url: "https://aidr.today/10c6be3c?lang=en"
title: "Google says self-improving AI isn't anywhere close"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-10-14T19:18:44.000Z"
category: "Research"
topics: ["google","coding","llm"]
source_urls: ["https://the-decoder.com/google-says-self-improving-ai-isnt-anywhere-close/","https://news.ycombinator.com/item?id=50113841","https://www.businessinsider.com/sergey-brin-google-gemini-ai-argon-recursive-self-improvement-rsi-2026-10"]
summary: "Google engineer Haydn Belfield stated that there is no measurable acceleration of AI research driven by AI at DeepMind, with 88 percent of code changes still requiring human oversight."
---

# Google says self\-improving AI isn't anywhere close

> [Open the canonical story](<https://aidr.today/10c6be3c?lang=en>)

**Published:** 2026-10-14T19:18:44.000Z
**Category:** Research
**Topics:** google, coding, llm

## Summary

Google engineer Haydn Belfield stated that there is no measurable acceleration of AI research driven by AI at DeepMind, with 88 percent of code changes still requiring human oversight\.

## Sources

- [Story source](<https://the-decoder.com/google-says-self-improving-ai-isnt-anywhere-close/>)
- [Discussion](<https://news.ycombinator.com/item?id=50113841>)
- [Story source](<https://www.businessinsider.com/sergey-brin-google-gemini-ai-argon-recursive-self-improvement-rsi-2026-10>)