Anthropic drops postmortem on Claude performance bugs

Anthropic's postmortem confirms what the developer community suspected: Claude's recent coding output genuinely degraded due to underlying engineering issues.

• –Validates user complaints that Claude "got dumber," showing it wasn't just a change in user perception
• –Highlights the fragility of deploying minor model updates and the resulting downstream impacts on developer tools
• –Sets a strong precedent for transparency in the AI industry by publicly acknowledging and explaining the regressions