Based on my personal experience, and from reports by others on HN, it seems that Claude Code has massively regressed over the past two weeks. This appears to be a combination of both usage limits being severely restricted and the quality of its work dropping significantly.
I'm on the $100 Max plan and am getting better results with GPT-5.4 on the $20 Pro plan, both in terms of quality and amount of work. Could it be that the moment Anthropic started leading the LLM race, it also started enshittifying its product? I know enshittification implies a longer process, but is this the first step? Curious to hear what others are experiencing.
I'm on the $100 Max plan and am getting better results with GPT-5.4 on the $20 Pro plan, both in terms of quality and amount of work. Could it be that the moment Anthropic started leading the LLM race, it also started enshittifying its product? I know enshittification implies a longer process, but is this the first step? Curious to hear what others are experiencing.