Researcher: anti-distillation claims rest on weak evidence, CoT distillation may be unstoppable
adam_karvonen · x · 2026-11-22
Adam Karvonen pushes back on an anti-distillation blog post, arguing its evidence that distillation can actually be prevented is weak for typical product use cases — and that similar posts put little effort into tuning the no-CoT-distillation baseline. With modest improvements to distillation training (e.g. synthesizing a CoT for a given output), distillation may be practically impossible to stop.
Related event: Researcher Questions OpenAI's Anti-Distillation Claims(4 posts)→
More from Models
- GPT-6.2 Astra writing success rate falls to 85.9% from 91.2% — arena · 2026-11-22
- GPT-6.2 Astra gains +50 in coding, drops -54 in chat on Agent Arena — arena · 2026-11-22
- Apple testing daily limits on advanced Siri AI requests in iOS 27.3 beta, per leak — alvelda · 2026-11-22
- Muse plays GTA6, suggesting guardrails don't cover game playing — Dave_Maynor · 2026-11-22
- Paying Google Workspace users locked out of Gemini over custom-domain account restrictions — film_girl · 2026-11-22
- Lean Community Verifies 17M Lines of Code, Confirms Claude-Assisted Riemann Hypothesis Proof — Polymarket · 2026-11-22