Simon Willison reproduces Qwen 3.8 27B model-weight reading demo
11/04/2026 — 11/04, 01:19·1 sources (all)·1 reports (all)
Key takeaways
- Simon Willison says he reproduced an example from the paper "Sparks of Weight-Space Programming", showing that Qwen 3.8 27B running on a loc…
Background & analysis
On November 4, 2026, Simon Willison published a post saying he had reproduced one of the examples from the paper "Sparks of Weight-Space Programming" and could read the raw model weights of Qwen 3.8 27B running locally on a laptop. He noted that the paper, by Rikiya Takehi, Jayden Teoh, Alex Zhang, and Omar Khattab, had become widely discussed the previous day. It reported that frontier models, particularly GPT-6 Astra, can natively interpret and modify the raw weights of a machine learning model, given a text description consisting of floating point numbers. Willison said he had been playing with one of the paper's examples and was delighted to share that Qwen 3.8 27B could read model weights in that same format. That is where the matter currently stands. The available source does not give further detail on how Qwen 3.8 27B performed in this test or whether the paper has been formally published.
AI-generated from 1 reports · updated 5 hours ago
Related AI tools
- Heat index
- 86
- All sources
- 1
- All reports
- 1
- First seen
- 5 hours ago
Reports on this story headlines open the original
Simon Willison says he reproduced an example from the paper "Sparks of Weight-Space Programming" and found that Qwen 3.8 27B running on his laptop can also interpret model weights presented as a text dump of floating point numbers. The paper had reported that frontier models such as GPT-6 Astra can read and modify raw neural network weights.
Simon WillisonAI score 60
Other stories people are talking about
- Heat index 42User develops desktop invoicing app using Qwen 3.8 27B and coding agents1 sources
- Heat index 293Reddit rebuilds weight edits that lift Qwen3.5-0.8B to 48.1% on MATH-5005 sources
- Heat index 67Lessons Learned and Pitfalls in Integrating Qwen into Enterprise Ticketing Systems1 sources
- Heat index 140Local execution of Qwen3.8-Flash-Next paired with Claude Code for code review2 sources
- Heat index 43Qwen lead says Qwen 4.0 and 5.0 will stay open1 sources
How is heat calculated?About the methodHide
Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.
This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.
- Surge
- Discussion rising fast
- New
- First report within 6 hours
- Rising
- Still gathering discussion