Anthropic Is Testing User Patience
Anthropic is throttling usage again - heavily, and seemingly randomly. Whilst I’ve bestowed accolades upon Anthropic’s Claude model family for its coding prowess, the company’s model throttling is infuriating. This is why I cancel my Anthropic subscription regularly and use other premium models in its stead, most notably GPT 5 (OpenAI) and Gemini 2.5 Pro (Google, GOOGL).
My guess is that Anthropic’s dollar renewal rate is probably in the 50% range, which is terrible, and why I believe the company will eventually be sold, most likely to Google.
Developers that pay Anthropic for firehose access no doubt have a better user experience than me, but I’d prefer to go the opensource route were I to write code 12 hours day, 6 days per week. DeepSeek and other opensource providers build models that have outstanding performance.
