Developers are downgrading their AI on purpose — and it’s working

Anthropic confirmed deliberate modifications to Claude Code's instructions on the Opus 5.5 model to prevent users from recreating existing video games and infringing intellectual property. However, developers report the model now refuses additional tasks like reverse-engineering and software analysis that it previously handled. The restrictions have disrupted paid client projects for some developers, leading them to switch to older Claude models. Anthropic's Addy Osmani acknowledged the changes and stated the company is iterating to balance policy compliance with developer needs.

Summary and key points written by AI from the headline and a short excerpt, so they can be incomplete or wrong. The New Stack did the reporting; read the original for the full story.