Claude is suddenly stupid, there is nothing you can do

· Blog Kronis · Nov. 14, 2026, 9:01 p.m.
Summary
The author expresses frustration over a perceived decline in the performance of the AI model Claude, attributing it to possible A/B tests or server-side changes by the developers. They detail specific instances where Claude failed to execute tasks effectively, contrasting these issues with earlier successes. The piece critiques the lack of accountability and transparency from AI providers, advocating for better regulations to protect users from service degradation and arbitrary bans. Ultimately, the author calls for more consistent AI performance and accountability from providers while lamenting the current state of affairs in AI services.