Hugging Face reproduced weight-space programming on math. We tried it on medicine.
This is a dev post classified by AGK Intelligence as AI dev tools (a tool drop), kept by the Dev Radar because it carries real work, not commentary.
Hugging Face reproduced weight-space programming on math. We tried it on medicine. The paper's math edit, zero medical data, on 1,273 MedQA questions: Qwen3.5-0.8B goes from 24.4% to 32.9%. It stops the thinking loops: answers cut at the token limit fall from 18.3% to 0.1%.
Posted by OpenMed (8.4k followers) 1 h ago · 14 likes · 501 views · view the original post on X. Kept by the Dev Radar as AI dev tools. Tools mentioned: qwen3.5-0.8b-weight-space-programming, weight-space-programming-demo.
More dev work like this
- TRENDING 📈 @elliotarledge reverse-engineered Wispr Flow and used DeepSeek to fine-tune… — @thestartpod
- EVAL (0x10f52295e129817e8e4b28da0dab067321ec2222) staking is now live on… — @reppo
- I used Composio Instant with Claude Code to check how my own product, Krevox, shows up… — @neerajjj6785
- Zyphra is building something interesting called MAIA. — @TeksEdge
- Ultralytics Platform now supports 12 additional languages! 🌍 — @ultralytics
- Stop hunting for ghosts in the heap → https://goo.gle/4sQHooo — @ChromiumDev
- A new model shouldn't mean waiting for a new OpenClaw release — @openclaw
- The coding agent needed pytest. So its sandbox can reach PyPI, and nothing else 🔒 — @marcelotryle
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 31.8k posts from 5k X accounts over the last 21 days, 9.5k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-06 19:17 UTC. Full method.