This wave of Notion's Claude malfunction is interpreted as model degradation, which is quite typical—API timeouts and the model becoming less intelligent are two different issues, but people's first reaction is always 'AI is failing,' and the anxiety over closed-source black boxes is indeed profound.

Notion announced on June 7 that Anthropic Claude Opus 4.7/4.8 experienced performance degradation leading to increased AI failure rates, and related models were disabled in the model selector, sparking many interpretations of model deterioration. Notion's product manager clarified that it was only an API failure and connection timeout, a short-term fluctuation in cloud services, which was restored within a few hours. The incident exposed the semantic conflict over "performance" in the AI era, as well as the biases of closed-source anxiety and blaming output fluctuations on the model becoming dumber.