Defining the Debate: Is AI Replacing Software Engineers?
The question of AI replacing software engineers refers to whether advanced coding tools and autonomous agents will remove the need for traditional developer roles, or instead transform those roles into broader product-building jobs where far more people can create software without deep programming expertise. At Anthropic, this debate is not theoretical; it plays out through Claude Code, an agentic coding tool described as one of the fastest-growing in its category. Its creator, Boris Cherny, argues that for much of his own work, coding is already “solved,” and he has not written a line of code in more than six months. This claim sits at the center of a wider argument: are tools like Claude Code a force for AI career displacement, or a way to spread software creation to many more “builders” than today’s engineering workforce?
Anthropic’s Coding Expert: The End of the ‘Software Engineer’ Title?
Cherny’s most provocative view concerns the software engineer future as a job title. He predicts that the label “software engineer” could start to disappear, replaced by roles closer to general “builders” who combine product thinking, design, and AI agents to ship code. In Cherny’s world, designers, product managers, and other colleagues are already becoming effective coders through Claude Code, shrinking the gap between idea and implementation. This outlook treats coding AI job impact as direct job displacement for many traditional developers, especially those focused on straightforward implementation work. At the same time, Cherny says he is consciously automating his own job, showing he expects rapid change rather than a slow transition. His perspective captures a growing concern in tech: as coding becomes more automated, what happens to the classic career ladder built on years of writing and maintaining code by hand?
A ‘Golden Age’ for New Founders, Not the End of Coding
Alongside warnings about the end of the software engineer title, Cherny offers a strikingly optimistic view of opportunity. He urges 22-year-old computer science graduates who feel entrepreneurial to start companies, calling this “the golden age” for launching a startup with AI agents. He says that thanks to tools like Claude Code, “you and your agents can build a giant company,” shrinking the need for large technical teams at the earliest stages. During a talk with Y Combinator founders, he found that about half the room let Claude Code write 100% of their code, while only one person out of a couple hundred used no model-generated code at all. These figures suggest coding is being automated for a large share of new projects, but also that more people can build working products from day one.
More People Writing Code, Fewer Traditional Engineers
Cherny’s forecast does not imagine a world with less software creation; he expects far more of it. He predicts that if we count everyone who writes code or uses agents to write code, there will be 100 times more such people than today. In this view, the coding AI job impact is not a simple story of AI career displacement, but a reshaping of who participates in building software. Technical fluency still matters, but product insight and user understanding gain equal weight. Investors such as former Y Combinator leaders echo this trend, saying they are keen to fund founders who “can’t code at all” but deeply understand their users. The likely outcome is a mixed landscape: traditional engineering roles may shrink or change title, while a broader class of builders and founders use AI to assemble and ship complex systems at high speed.