The 'Astra rewires Qwen' story gets a strong pushback
Follow-up to Tuesday Nov 3: AI that edits AI: Astra rewires Qwen, and Anthropic says self-building AI is closeYesterday's top token was about AI that edits AI. A paper-shaped blog post by Rikiya Takehi, Jayden Teoh, Alex Zhang and Omar Khattab showed GPT-6 Astra, an OpenAI model, reading and changing the raw weights of a small Qwen model. Weights are the numbers that make up a model. The edit touched 0.016% of the weights of Qwen3.5-0.8B, and its MATH-500 score went from 7.9% to 48.1%.
Now the pushback. A post on honnibal.dev titled 'The Astra can read and write weights to improve Qwen post is misleading' got 990 points on Hacker News. Our sources give only the title, not the full argument, so read it before you decide.
At the same time, people are testing the idea at home. Simon Willison says Qwen 3.8 27B on his laptop can interpret model weights too. On Reddit, one user had Claude change 8 neurons of Gemma 4 12B to fix one habit in its Italian output. Another used Anthropic's 'Jacobian lens' to look inside Qwen when it gets stuck saying 'Wait'.
AI that improves AI is one of the biggest safety questions there is, and a new arXiv paper, RSI-Dojo, even trains agents to automate model training. That is exactly why big claims need careful checking, not hype in either direction. Today, the checking is happening in public, and that is healthy.
