With 4.7 worse than 4.6, and 4.8 worse than 4.7, and how they’re vibe coding claude code which is so full of bugs I had to stop using it a month ago, and now rust bun publicly refusing to do human code review.
Can their members of technical staff dig themselves out of the slop spiral.
Notion just removed Anthropic models.
Opus 4.7 and 4.8 degraded, causing higher failure rates in Notion AI.
The harshest model review is not a benchmark. It is your partner removing you from the model picker.
This matches my experience too: Opus 4.8 has felt weirdly worse than Opus 4.6 lately.