Asking Claude to save its reasoning for a grading tool got the user's session flagged
Orochhe_Marou · reddit · 2026-10-10
A researcher running multi-agent workflow experiments had Claude save its reasoning traces to improve a coding-effort estimation tool — and got their session flagged by Anthropic's guardrails, with Rewind unable to recover it.
They warn others not to ask for saved reasoning output in chat, and argue that extracting model reasoning as training data isn't fundamentally different from Anthropic previously training on human-generated books, articles, and chats — criticizing the company for framing it as a moral issue while doing the same thing.
More from Models
- ChatGPT web can actually listen to audio, but the same model can't via API or Codex — Possible_Statement84 · 2026-10-23
- MiniMax M3.1 Flash free for 3 days on 0G, with 1M context and 512K output — MiniMax_AI · 2026-10-23
- TLA+ expert corrects Grok: TLC only exhausts finite state spaces, not all behaviors — tianyin_xu · 2026-10-23
- Open LLM cost calculator shows Arabic tokenization and 6x price gaps across models — Moham-saed · 2026-10-23
- OpenAI blocks manual reset with 13% usage left: 'no reset needed right now' — AccomplishedSugar490 · 2026-10-23
- Meta's 816B Muse Spark Open scores 39, losing to 180B Qwen3.8-Flash-Next at 40 — BTA_Labs · 2026-10-23