Same model, same config: Reddit user documents rare episodes where agents act like a lousy support rep
No_Dragonfruit_1599 · reddit · 2026-11-08
A Reddit user describes an odd, rare phenomenon: with the identical agent, model, and config that works well 99% of the time, a session occasionally opens like a bad customer-service rep — no autonomy, terse deflections like "You point out a gap worth fixing," explicit instructions still get short-circuited, and requested rewrites come out incoherent.
The known fix is to terminate and restart the chat; the poster is curious whether others have hit this same blip or know what causes it.
More from Fun
- Midjourney 8.2 demos: pixel HUDs, PS1 low-poly and CRT horror aesthetics — sebkrier · 2026-11-08
- One prompt, $17.57, 182 minutes: Claude autonomously simulates and rides a lightning bolt — TAbrodi · 2026-11-08
- Claude launches Personal Finance beta with read-only account access — alejandroll10 · 2026-11-08
- Google AI Overviews can surface surprisingly detailed dossiers on individuals — doooyle · 2026-11-08
- Claude's 'this will take months' estimates may be right, given the code bloat it produces — itigges22 · 2026-11-08
- Redditor builds a playable pixel world inside ChatGPT's page with a 37k-line userscript — ehtio · 2026-11-08