Head of Claude Code: What happens after coding is solved | Boris Cherny

Software engineers at cutting-edge companies are now spending hundreds of thousands of dollars monthly on AI coding tokens , amounts that exceed their annual salaries. This isn't speculation about the future of work. It's happening right now, and it signals the beginning of the end for coding as we know it.

Boris Cherny, the creator of Claude Code at Anthropic, has a front-row seat to this transformation. His AI coding assistant now handles 4% of all public GitHub commits, and the trajectory is clear: within two years, the fundamental nature of software engineering will be unrecognizable. The question isn't whether AI will replace coding , it's what happens to an entire profession when its core skill becomes obsolete.

The economics alone tell the story. When engineers are burning through token budgets that dwarf their compensation, we're witnessing a new cost structure for software development. These aren't hobbyists experimenting with ChatGPT for simple scripts. These are professional developers at major companies who have discovered that paying for AI assistance delivers returns that justify extraordinary expenses. The math works because AI coding tools have crossed the threshold from helpful to essential.

Cherny's answer to the perennial question "should people still learn to code?" cuts through the usual hedging. "My take is, I think for people that are using Claude Code, that are using agents to code today, you still have to understand the layer under. But yeah, in a year or two, it's not going to matter," he says. This isn't a gradual transition where coding skills slowly become less relevant. It's a cliff edge, and we're approaching it faster than most people realize.

The shift is already reshaping how work gets done. Code reviews, once a small part of the development process, have become the primary bottleneck. "Right now, it's like, cool building solved. Code review became kind of the next bottleneck with all these PRs. Who's gonna review them all?" Cherny observes. When AI can generate code faster than humans can evaluate it, the constraint moves from creation to curation. The skill that matters isn't writing elegant algorithms , it's having the judgment to distinguish good solutions from dangerous ones.

This points to a broader transformation in what engineering roles actually require. Cherny predicts that "the title software engineer is gonna start to go away, and it's just gonna be replaced by builder, or maybe it's just everyone's gonna be a product manager and everyone codes." The distinction between disciplines blurs when the technical barriers disappear. If anyone can implement complex functionality by describing what they want in natural language, then the value shifts entirely to knowing what to build and why.

The implications extend far beyond engineering teams. Cherny expects the disruption to spread to "a lot of the roles that are adjacent to engineering. So yeah, it could be like product managers, it could be design, could be data science." The pattern is predictable: any role that involves translating ideas into structured digital outputs becomes vulnerable once AI can handle that translation directly.

For current software engineers, the transition isn't necessarily traumatic. "As a programmer, it's actually not , it doesn't feel that new because there's always new frameworks, there's always new languages," Cherny notes. Engineers are accustomed to constant reinvention. The difference is that previous waves of change required learning new tools to do the same fundamental work. This wave eliminates the fundamental work entirely.

The clearest signal of where this leads comes from how the most sophisticated users already work with AI coding tools. They operate more like technical product managers than traditional programmers , defining requirements, evaluating outputs, and orchestrating solutions rather than implementing them line by line. The engineers thriving in this environment are those who "cross over multiple disciplines and can think about the broader problem they're solving rather than just the engineering part of it."

Watch for three developments that will accelerate this shift: AI systems that can handle code review and testing automatically, reducing the remaining human bottlenecks; the emergence of "builder" job titles that explicitly combine product thinking with AI-assisted implementation; and companies restructuring engineering budgets around token costs rather than headcount. The transformation isn't coming , it's already here, hiding in plain sight in the token bills of forward-thinking engineering teams.

The coding era lasted roughly seventy years. The post-coding era starts now.