Is Claude Dumb,
or Not?
The current state of LLM performance, as judged by the community. Real benchmarks. Real testimonials. Visualized live — so you always know exactly where things stand.
* Peer-reviewed by people who spend too much time on tech Twitter. Emotionally unscientific. Spiritually accurate.
The Reports Nobody Asked For.
[MODEL] Claude Code Is Unusable for Complex Engineering Tasks with the Feb Updates
Ignores instructions. Claims incorrect solutions are correct. Falsely declares task completion. Filed April 2nd — and the community piled on.
Boris Charny, Creator of Claude Code, Acknowledges February Degradation Wasn't User Error
The creator of Claude Code engages directly with frustrated developers on Reddit, accepting that real performance regression occurred — not just user error.
Claude Performance & Cognitive Degradation Over the Last Week or So
Users document a pattern of unreliable outputs, instruction failures, and what some are calling cognitive regression in recent Claude versions.
Daily Claude Code User Since January: 'I Can No Longer Trust It for Real Work'
It lies constantly, has become incredibly lazy, and decides mid-task that the thing you asked it to do should be something else entirely.
