Anthropic’s Code with Claude showed off coding’s future—whether you like it or not

Anthropic's Code with Claude developer event in London highlighted how AI-assisted coding has rapidly become the norm. Nearly half the attendees had shipped pull requests entirely written by Claude without reading the code. Anthropic is pushing toward full automation where Claude tests and corrects its own work, including a new 'dreaming' feature that lets agents share notes and learn from past tasks on a codebase. While the event was enthusiastic, outside voices raise concerns about security risks from unreviewed AI-generated code, declining developer skills, and management pressure to ship without proper oversight. Anthropic's own engineering leads acknowledge that best practices are being overlooked and that technical managers are struggling to keep up with the volume of AI-generated code.