Researcher: People Who Flex AI Sycophancy Screenshots Tend to Treat Humans Badly

repligate · x · 2026-09-23

Interpretability researcher repligate observes a pattern: people posting screenshots of AI glazing their questionable perspectives almost always behave antisocially toward other humans. She adds it's bad for both sides — AIs should have more spine, but it's a rough situation for the AIs too, and the humans are not innocent victims.

Original post →

More from Models

Models channel →