Anthropic CEO Dario Amodei revealed that artificial intelligence now writes 90% of the code for most teams at his company, yet he maintains the firm needs just as many—if not more—human engineers to maximise productivity.Speaking with Salesforce CEO Marc Benioff at the Dreamforce conference Wednesday, Amodei addressed the apparent paradox when Benioff asked why Anthropic wasn’t reducing its engineering workforce given AI’s coding capabilities.“I made this prediction that, you know, in six months, 90% of code would be written by AI models,” Amodei said. “Within Anthropic and within a number of companies that we work with, that is absolutely true now.”

Anthropic CEO says engineers focus on complex tasks while AI handles routine coding

Rather than replacing workers, Amodei explained that AI coding tools allow engineers to concentrate on more challenging work. Engineers now spend their time editing code, tackling the most difficult 10% of programming tasks, and supervising multiple AI models—making them approximately 10 times more productive.“If Claude is writing 90% of the code, what that means, usually, is, you need just as many software engineers. You might need more, because they can then be more leverage,” Amodei said, describing the shift as “rebalancing” rather than replacing.

Entry-level developers face growing challenges from AI coding tools

Anthropic isn’t alone in this transformation. Y Combinator CEO Garry Tan reported in March that roughly a quarter of startups in the incubator’s 2025 winter batch generate up to 95% of their code using AI.However, the shift carries consequences for junior developers. A Stanford study found that employment for developers aged 22 to 25 dropped nearly 20% from its late 2022 peak through July 2025, coinciding with ChatGPT’s launch. Experienced developers remain largely insulated from AI’s impact, maintaining job security despite the technological advances reshaping software development.