In a striking demonstration of technological advancement, Boris Cherny, the creator and head of Claude Code at Anthropic, shared insights on a consequential shift in software development. Rather than manually writing code, he now oversees numerous AI agents that autonomously manage coding tasks, marking a significant change in how programming works today. Cherny noted, "This morning I was managing maybe a few hundred. Some days it’s…thousands, or tens of thousands."
The transition from a single instance of Claude Code run by developers to a vast network of AI subagents signifies a leap in productivity. Cherny explained how the current setup enables a level of automation that was once unimaginable. "Fast forward to today, it looks very different. You have a Claude Code, but it has subagents that are other Claudes," he elaborated. Instead of relying on direct prompts, AI is now initiating its own prompts, which could foster a more efficient and innovative coding process.
Cherny compared this shift to the impact of the printing press, which transformed literature production in the 15th century. He believes AI coding assistants like Claude Code are lowering barriers to software creation and could spark a new wave of innovation. "This massive speedup in coding will be as consequential as the printing press," he stated. The implications of this technology are extensive, with the potential to reshape industries and inspire new forms of creativity.
Recently, Anthropic published a blog titled “When AI Builds Itself,” discussing the use of AI systems in developing future models. This recursive self-improvement could enable AI to autonomously design and enhance its successors. Cherny noted that Claude Code's capabilities have already led to an eightfold increase in the code produced by the company since the start of the year. "I think this might be the first product that actually just takes off," he remarked.
The autonomy of Claude Code is also noteworthy. It can now conduct its own security reviews and generate ideas for future projects. Cherny expressed his excitement, stating, "We’re starting to get to the point where it has ideas. It’s looking at GitHub, it’s looking at [X], it’s figuring out ‘what should I build next?’"
For Cherny, this evolution allows him to concentrate on more strategic and impactful work as a builder. The role of developers is changing dramatically as AI takes on a greater share of the workload. However, this transition brings notable risks. When asked about the dangers of recursive self-improvement in AI, Cherny acknowledged the concern: "It’s one of the big risks for AI." Anthropic is aware of the challenges ahead and is taking necessary precautions to navigate this new frontier.
As AI systems become more integrated into software development, traditional coding methods are being transformed. The rapid acceleration in efficiency suggests that the future of programming will not only be faster but also more innovative, as machines take on roles once reserved for human developers. The full extent of this transformation is yet to be realized, but the potential for disruption is clear.
The stories that move AI & crypto markets — before the market reacts.
Free. 7am ET. Five stories. 62,400 readers.
