Silicon Valley programmers are experiencing a seismic shift in their profession as AI agents like Claude Code take over the bulk of software development. Manu Ebert, a machine-learning engineer, now spends his days conversing with A.I. rather than writing code, with agents completing tasks in minutes that once took days.

This transformation represents a dramatic departure from traditional coding, where developers hunched over keyboards for hours debugging intricate details. The new workflow involves describing desired outcomes in plain English and letting A.I. agents handle implementation, testing, and deployment. Developers report productivity gains of 10 to 100 times their previous output, with some tasks that once required a full day now completed in half an hour.

Unlike other professions threatened by automation, many programmers express enthusiasm for these tools. They view A.I. as removing the drudgery of debugging and error-checking while preserving the creative, problem-solving aspects of their work. This stands in stark contrast to how creatives and other professionals experience A.I., which often strips away the most meaningful elements of their craft. The coding profession, once considered a guaranteed path to lifetime employment, now faces an uncertain future as AI-generated code becomes increasingly reliable and efficient.