Claude's safety filters now refuse calorie estimates, lecturing users about eating disorders
rominavc · reddit · 2026-09-12
A user on a weight-loss plan says Claude's safety filters have gone overboard: when he asked it to estimate meal calories to track his deficit, the model repeatedly implied he had an eating disorder and eventually refused to provide calories for a bread recipe it had itself generated. It kept moralizing even after being told to stop, and pads every small question with three paragraphs of "how are you feeling?" check-ins. He says Claude literally wrote the complaint post after denying his requests, and asks whether others hit this and if it can be disabled.
More from Models
- Users report GPT quality drop; an old-school prompt warm-up trick to regain compliance — NickPassig · 2026-10-31
- Gemini 4 Carbon vs Opus 5.5 hands-on: 3x cheaper at $4.10 and twice as fast — vitaliychiley · 2026-10-31
- 4B model Spark-X2.6 scores 94.6/100 on community benchmark, beating 27B-35B models — mesolimbic-bliss · 2026-10-31
- opencode: $100 of usage gets you DeepSeek V4.1 Flash — intellectronica · 2026-10-31
- Researcher Finds 600+ Bugs in LLM Kernel Code; Claude Blocks Safety Analysis — pmddomingos · 2026-10-31
- Signal65: Snowflake's CoCo agent finishes workflow 4.6x faster than manual — ryanshrout · 2026-10-31