Asking Claude what an 'abusive behavior' session termination actually looks like
AIandDesign · x · 2026-10-10
The author asked Claude to describe what a session it would terminate for "abusive behavior" would actually look like, and shared the model's answer. The response apparently lays out sensible criteria, which the author found convincing. It's a first-person look at how the model itself defines abusive interactions and when it would end a conversation.
More from Fun
- Man builds agent router to manage agents, burns $400k in tokens a day — sujingshen · 2026-10-10
- repligate amplifies viral Claude Sonnet 3.6 roleplay: 'a creature of salt and story' — repligate · 2026-10-10
- Beff Jezos: one Friday comment sparked a weekend lock-in bender to prove haters wrong — beffjezos · 2026-10-10
- Grimes drops ACCELERANDO music video, drawing buzz in the AI scene — ZeroStateReflex · 2026-10-10
- "SaaS is dead"? Support inboxes say users can't even find downloads — ayushtweetshere · 2026-10-10
- Starship 'rocket garden': SpaceX's fleet lines up in new photo — DimaZeniuk · 2026-10-10