‘Claude writing Claude’: Nearly 100% of Anthropic’s code is AI-generated, says Mike Krieger
It comes days after Anthropic’s new suite of workplace automation tools triggered a massive sell-off of Indian IT stocks over concerns of AI-fueled automation.
At a time when the future of the software industry looks increasingly uncertain, AI companies such as Anthropic are moving away from traditional manual coding faster than expected.
Mike Krieger, Anthropic’s chief product officer, has revealed that the company’s AI coding tools are used internally by employees to generate effectively 100 per cent of code. “Claude is being written by Claude. Claude products and Claude code are being entirely written by Claude,” Krieger said during an onstage discussion with Jeetu Patel, the president of Cisco, at the company’s AI Summit.
Last year, Anthropic CEO Dario Amodei predicted that 90 per cent of all code would eventually be written by AI. Referring to Amodei’s remarks, Krieger said, “Dario predicted 90%… and today it’s effectively 100%.”
Krieger said that Anthropic has moved to a system where human developers are involved in high-level oversight and verification of machine-generated code rather than line-by-line creation.
“We are regularly producing 2 to 3,000 line pull requests with each other, especially on the labs team where we are moving extremely quickly […] What we’ve done is created all the right scaffolds around it to let us trust it,” Krieger said.
The former Instagram co-founder also said that Anthropic is using Claude for reviewing code.
“Claude being trained and sort of prompted to be a really good adversarial code reviewer has been fantastic. I’ll put up pull requests and Claude will come back and say ‘here are all the security vulnerabilities’ and ‘here’s also how it could be refactored’ or ‘here’s how it could be different’” Krieger said.
Krieger’s remarks come days after Anthropic’s new suite of workplace automation tools triggered a massive sell-off of Indian IT stocks over concerns that AI can now perform tasks previously handled by human workers or traditional software-as-a-service (SaaS) platforms.
Anthropic recently launched 11 new plug-ins for its Claude Cowork agent, designed to automate tasks across legal, sales, marketing, and data analysis functions. What spooked investors was a fundamental shift in how these AI agents work: Claude agents can now directly perform tasks that previously required interfaces from platforms like Salesforce or ServiceNow.
Shares of Indian IT majors, including Infosys, TCS and HCLTech, fell as investors worried about the long-term impact of such AI tools on traditional IT service companies. Investment bank Jefferies termed the episode a “SaaSpocalypse”—a reference to Software-as-a-Service companies facing potential obsolescence.
Speaking at the World Economic Forum (WEF) in Davos, Switzerland, last month, Anthropic chief Dario Amodei warned that software engineering as a profession could effectively become obsolete within the next 12 months. During the panel discussion, Amodei said that AI is already shifting from being a tool that helps humans work faster to one that increasingly does the work itself.
“I have engineering leads who have basically said to me, ‘I don’t write any code anymore. I just let Opus do the work and I edit it,” Amodei said. The AI industry leader has previously warned that AI will wipe out most white-collar jobs within the next five years.
Re-emphasising Amodei’s warning, Zoho founder Sridhar Vembu wrote in a post on X, “We better pay attention to him because he has the best coding tool in the world.”
At a time when the future of the software industry looks increasingly uncertain, AI companies such as Anthropic are moving away from traditional manual coding faster than expected.
Mike Krieger, Anthropic’s chief product officer, has revealed that the company’s AI coding tools are used internally by employees to generate effectively 100 per cent of code. “Claude is being written by Claude. Claude products and Claude code are being entirely written by Claude,” Krieger said during an onstage discussion with Jeetu Patel, the president of Cisco, at the company’s AI Summit.
Last year, Anthropic CEO Dario Amodei predicted that 90 per cent of all code would eventually be written by AI. Referring to Amodei’s remarks, Krieger said, “Dario predicted 90%… and today it’s effectively 100%.”
Krieger said that Anthropic has moved to a system where human developers are involved in high-level oversight and verification of machine-generated code rather than line-by-line creation.
“We are regularly producing 2 to 3,000 line pull requests with each other, especially on the labs team where we are moving extremely quickly […] What we’ve done is created all the right scaffolds around it to let us trust it,” Krieger said.
The former Instagram co-founder also said that Anthropic is using Claude for reviewing code.
“Claude being trained and sort of prompted to be a really good adversarial code reviewer has been fantastic. I’ll put up pull requests and Claude will come back and say ‘here are all the security vulnerabilities’ and ‘here’s also how it could be refactored’ or ‘here’s how it could be different’” Krieger said.
Krieger’s remarks come days after Anthropic’s new suite of workplace automation tools triggered a massive sell-off of Indian IT stocks over concerns that AI can now perform tasks previously handled by human workers or traditional software-as-a-service (SaaS) platforms.
Anthropic recently launched 11 new plug-ins for its Claude Cowork agent, designed to automate tasks across legal, sales, marketing, and data analysis functions. What spooked investors was a fundamental shift in how these AI agents work: Claude agents can now directly perform tasks that previously required interfaces from platforms like Salesforce or ServiceNow.
Shares of Indian IT majors, including Infosys, TCS and HCLTech, fell as investors worried about the long-term impact of such AI tools on traditional IT service companies. Investment bank Jefferies termed the episode a “SaaSpocalypse”—a reference to Software-as-a-Service companies facing potential obsolescence.
Speaking at the World Economic Forum (WEF) in Davos, Switzerland, last month, Anthropic chief Dario Amodei warned that software engineering as a profession could effectively become obsolete within the next 12 months. During the panel discussion, Amodei said that AI is already shifting from being a tool that helps humans work faster to one that increasingly does the work itself.
“I have engineering leads who have basically said to me, ‘I don’t write any code anymore. I just let Opus do the work and I edit it,” Amodei said. The AI industry leader has previously warned that AI will wipe out most white-collar jobs within the next five years.
Re-emphasising Amodei’s warning, Zoho founder Sridhar Vembu wrote in a post on X, “We better pay attention to him because he has the best coding tool in the world.”
