Researchers Slam DeepMind for Skipping Control Evals on Carbon Model
jammastergirish · x · 2026-11-17
A researcher argues Google DeepMind's handling of its new model Carbon is indefensible: despite marketing Carbon as "our strongest model" with "significant gains across coding and agentic tasks," GDM ran no control evals on it and instead reused Argon's monitor-evasion results, reasoning that Carbon "did not show meaningful new capabilities."
The poster draws a Boeing analogy — we would not accept a planemaker claiming its new aircraft is barely faster or bigger and therefore needs no fresh safety review. The dispute centers on whether capability claims should trigger matching safety evaluation obligations.
More from Models
- 'Humans Also Read Linearly and Causally, So It's Not a Transformer Issue' — yacineMTB · 2026-11-17
- Fudan & OpenMOSS map the trade-off: long-context hybrids learn more, extrapolate less reliably — jiqizhixin · 2026-11-17
- Models favor their own APIs in code: Opus calls Claude API in 97% of implementations — santiviquez · 2026-11-17
- Search behavior gap: Opus searched in 10/60 answers, GPT models searched every time — santiviquez · 2026-11-17
- 79% of 2,880 LLM API recommendations went to just four big-cloud services — santiviquez · 2026-11-17
- Hy4 Flash hands-on: DeepSeek-V4.1-level quality at ~1/15 the cost, 60% faster — IlovedamemesUwU · 2026-11-17