Frustrating AI Behavior: Asking for a Clear Answer but Getting 'Directionally Right'

tokenbender · x · 2026-08-08

A developer expressed frustration over a typical communication pain point with current LLMs: when asking for a definitive answer, the model often responds with non-committal, fence-sitting phrases like "directionally you're right," which feels highly unhelpful.

Original post →

More from Fun

Fun channel →