Anthropic's latest findings reveal that its AI model, Claude, is now responsible for authoring over 80% of the code merged into the company's codebase. This significant shift, particularly evident since early 2025, has led to a roughly eightfold increase in code output per engineer, dramatically accelerating AI development.
The report, "When AI Builds Itself," posits that AI is increasingly becoming a collaborator in AI research itself, writing code, running experiments, and assisting with complex tasks. This trend moves closer to the concept of recursive self-improvement, where AI systems could eventually design their own successors, potentially revolutionizing not just AI but other scientific fields.
[IMAGE_4]While acknowledging that AI may still lack crucial research judgment, Anthropic suggests that the capabilities demonstrated pave the way for a future where AI drives innovation at an unprecedented pace. This development positions AI as a key driver in scientific advancement, prompting discussions on oversight and the future of human-AI collaboration in technology development.
