Qwen 3.8 27B reads and edits model weights on a laptop
Simon Willison says a local Qwen 3.8 27B reproduced a paper example previously shown with GPT-6 Astra.
46
AI model
Qwen 3.8 27B
- Mentions · 7d0
- Mentions · 30d3
- All mentions5
- First seenSep 8, 2026
- Last mentioned8d ago
Timeline
5 articles · newest firstCoding-agent memory files show no consistent security effect
Malicious CLAUDE.md-style files shifted secure completions by -8 to +6 points, while disclosure behavior differed by agent.
58
Stronger models defer maze turns only when humans hold the advantage
Reliable solo solvers almost never deferred without a human route advantage, but deferred on about 80% of initial decisions when the human path was shorter.
36
Local AI Model Modifies Windows Credential Dumper to Bypass EDR Detection
A local uncensored Qwen model revised an LSASS dumper until two lab EDRs no longer detected it.
64
I tested 10 model/harness combinations on the same Three.js task
A developer benchmarked 10 model/harness combinations on a Three.js task; Qwen 3.8 27B on OpenCode scored 95.64% fastest at 8m48s.
30
Appears with
- Command A+ai model · 1
- Anthropicorganization · 1
- OpenCodeproduct · 1
- Sonnet 5.5ai model · 1
- GPT 5.6 Solai model · 1
- Nvidiavendor · 1
- Elasticvendor · 1
- Anthropicvendor · 1
- DeepSeekvendor · 1
- GPT-6 Astraai model · 1
- Claude Opus 4.8ai model · 1
- DeepSeek V4.1 Flashai model · 1
- Microsoftvendor · 1
- Astra 6.0ai model · 1
- Claude Opus 5ai model · 1
- DeepSeek v4 Flash 0731ai model · 1
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.