AI & Machine Learning
Jun 8, 2026
Boris Cherny discusses managing AI agents and the future of software development
Jun 8, 2026
AI Summary
Boris Cherny, head of Anthropic's Claude Code, reports a significant shift in software development, with AI agents now handling much of the coding work. He believes this advancement could lead to transformative changes in the industry, akin to the impact of the printing press, while also acknowledging the risks associated with AI's self-improvement capabilities.
- Boris Cherny, creator of Claude Code, has not handwritten code in eight months, instead managing numerous AI agents.
- At a recent tech conference, he noted that he can manage hundreds to tens of thousands of AI agents at once, a shift from previous methods where developers ran single instances.
- Cherny compared the impact of AI coding assistants to the printing press, suggesting they could lower barriers to software creation and drive significant innovation.
- Anthropic's recent blog post discussed the potential for AI systems to autonomously develop future models, hinting at recursive self-improvement.
- Since the beginning of the year, Claude Code has led to an eightfold increase in code production at Anthropic, with the AI also conducting its own security reviews.
- Cherny expressed excitement about the evolving role of builders in software development, noting that Claude is generating ideas and taking actions independently.
- He acknowledged the risks of recursive self-improvement in AI, indicating that it is a serious concern for the company.
ai agentsautonomous systemssoftware developmentself-improvementscalability