Developers Shipping AI-Written Code at Scale

Code with Claude, Anthropic’s two-day developer event, kicked off on May 19, 2026 in London, the same day as Google’s I/O in Palo Alto. The event revealed striking adoption metrics: almost half the developers in the audience raised their hands when asked if they had shipped a pull request in the last week that was completely written by Claude. When asked a follow-up question whether they had shipped a pull request completely written by Claude without reading the code at all, most of the hands that were raised earlier stayed up.

Claude Now Writes Most of Its Own Code

According to Jeremy Hadfield, an engineer at Anthropic, most software at Anthropic is now written by Claude. The tool’s reach extends particularly to Claude Code itself—Claude has written most of the code in Claude Code.

Boris Cherny, who heads Claude Code, described a fundamental shift in the development paradigm: “The default is now ‘I’m going to have Claude prompt itself’ rather than ‘I’m going to prompt Claude.‘”

Recent Model Updates Boost Capabilities

Claude 4.6 and Claude 4.7 were released in February and April 2026 respectively, improving Claude Code’s capabilities. A new feature in Claude Managed Agents called “dreaming” allows Claude agents to write notes to themselves and save useful information about specific tasks for other agents to use.

Enterprise Adoption Accelerates

Companies including Spotify and Delivery Hero have reshaped their software development teams around Claude Code. Startups Lovable, Base44, and Monday.com have similarly reorganized their engineering practices around the tool.

Expert Engineers Still Essential

Katelyn Lesse, Claude engineering lead at Anthropic, stated that Claude is probably as good as a midlevel engineer at writing code but that expert engineers are still needed to design systems and troubleshoot harder problems.

Lesse also reported that some technical managers at Anthropic are exhausted by keeping up with all the code their teams now produce.

Looking Ahead to Self-Building AI

Angela Jiang, Claude product lead at Anthropic, stated that the absolute end state they are trying to get to is Claude being able to build itself.

Source: MIT Technology Review