The definition of top tier coding is changing month-on-month in the AI era.
Boris Cherny, the creator of Claude Code at Anthropic, has been one of the most candid voices documenting this shift — partly because he’s living it in real time. In a recent interview, he laid out exactly what his relationship with coding looks like today, and it bears almost no resemblance to what it looked like a year ago.
“In November, I uninstalled my IDE,” he said, “’cause, well, I wasn’t using it. In the last month, I just haven’t opened it, so I just uninstalled it.” The baseline he’s comparing against is already striking: “The way that I coded a year ago was I wrote code with some autocomplete in the IDE. At that point, I was running maybe five, ten Claudes in parallel, and my coding was prompting Claude to write code.”
That itself was already a significant departure from conventional software development. But Cherny says things have since moved to another level entirely.
“Now it’s actually leveled up, I think, again, to the next wave of abstraction where I don’t prompt Claude anymore. I have loops that are running. They’re the ones that are prompting Claude and figuring out what to do. My job is to write loops.”
The implication is worth sitting with. Cherny has effectively abstracted himself one layer above the AI. He’s writing the systems that talk to Claude, evaluate outputs, and decide what to do next. The human role, in his workflow, has become one of orchestration rather than execution — defining the logic of how AI should operate, rather than operating it directly.
“This is this next transition that I think we’re gonna see in the next few months and maybe through the rest of the year,” he added.
This isn’t the first time Cherny has signalled how far ahead his workflow is relative to the mainstream. In December 2025, he disclosed that he hadn’t opened an IDE even once during the prior month, with Claude Code writing every single line across 259 pull requests. And in March 2026, he confirmed that Claude Code is now 100% written by Claude Code itself — a recursive milestone that would have seemed like science fiction just two years ago. More broadly, 4% of all public GitHub commits are now being made by Claude Code, with projections suggesting that figure could cross 20% by the end of 2026.
The loop-writing paradigm Cherny describes is the logical endpoint of a trajectory that’s been building for some time. First came AI autocomplete. Then came prompting AI to write whole functions or files. Then came running multiple AI agents in parallel. Now comes writing automated loops that do the prompting themselves, assess what comes back, and iterate — with the human setting the rules of the game rather than playing it. Each layer has compressed the timeline to the next. If Cherny’s read is right, the developers who thrive in the next phase won’t be the best at writing code or even the best at prompting AI — they’ll be the best at designing systems that prompt AI well.