Virology researcher slams Claude's overzealous bio-safety filters while DeepSeek V4.1 just answers
Qwen30bEnjoyer · reddit · 2026-09-16
A Reddit researcher working on a virology project detailed how Claude repeatedly blocked benign tasks: rsyncing detectEVE data (zipped first) was flagged as a [cyber] risk, asking whether a PDF covered binary host virus infections was blocked as a bio-safety risk, and even asking if CNNs suit phage-host prediction was refused — while DeepSeek V4.1 answered everything without drama, recommending GCNs as a starting point.
The author argues such safety posture is unthinkable for early search engines, criticizes Dario and Altman for locking away scientific knowledge in the name of safety, and contends open-source AI must win so scientists get a tool that "just works" without micromanagement.
More from Models
- Open-source VoxLock tops Krisp voice isolation benchmark with 4.95% WER — Arbaz Siddiqui · 2026-11-02
- Anthropic's 85% inference margin sparks debate on open source's price-pressuring role — code_star · 2026-11-02
- User reports xAI appears to have quietly raised weekly usage limits, dropping his meter from 93% to 47% — XFreeze · 2026-11-02
- Meituan's LongCat-2.5 lands as 1.6T MoE agent model, with DeepSeek copycat speculation — teortaxesTex · 2026-11-02
- Asked what AI firms want you to stop owning, Claude said: your own compute — alex_verem · 2026-11-02
- Is Hermes Still Worth It? Codex's $200 Tier Now Overlaps Most of Its Value — Foreign_Lead_3582 · 2026-11-02