AI CEOs are making some dramatic predictions these days, but they seem to know what they’re talking about.
Mike Krieger, Chief Product Officer at Anthropic, recently shared a striking glimpse into how the AI company builds its products today. He said that Dario Amodei’s prediction from March last year of 90% of coding being done by AI in 3-6 months, and 100% in twelve months seems to have come true at his own company.
“(I’ll) maybe give a vignette from what it’s like right now inside Anthropic on our lab team,” Krieger said. “We are regularly producing 2,000 to 3,000 line pull requests with each other, especially on the lab team, where we’re moving extremely quickly. Claude is being written by Claude. Claude products and Claude Code are being entirely written by Claude.”
The scale of these pull requests—massive code contributions that would typically require days of human effort—speaks to the transformation underway at the AI safety company. But Krieger’s most remarkable claim came next, when he referenced a prediction made by Anthropic CEO Dario Amodei roughly a year ago.
“When Dario was on stage a year-ish ago and was like, ‘Oh, by the end of the year, 90% of code is gonna be written by Claude,’ people thought it was crazy, but we were already seeing that trajectory,” Krieger explained. “Right now for most products at Anthropic, it’s effectively 100%.”
The implications of Anthropic achieving 100% AI-generated code are profound, both for the company and the broader software development industry. If one of the world’s leading AI labs is building its own AI systems entirely with AI assistance, it suggests we may be approaching an inflection point where AI development becomes largely self-sustaining. This also validates the capabilities of Claude Code, Anthropic’s command-line tool for agentic coding, which appears to have become indispensable to the company’s own engineering workflows. The shift from Amodei’s 90% prediction to 100% reality in such a short timeframe suggests the transition to AI-assisted development may accelerate even faster than industry leaders anticipated, potentially reshaping how software companies staff and structure their engineering organizations in the years ahead.