For a while now, software engineering has undergone a profound, fundamental shift: Where we once wrote most lines of code by hand, we now ask AI coding assistants and agents to generate large swaths of code for us. But AI needs supervision, so it is we, the humans, who must read the generated code, clean it up, and verify that it does what we want it to do. And because AI has a limited view of the greater context in which the code it generates must operate, it is we who define the system architecture, design the boundaries between services, and ensure the overall safety and reliability of our production environments. In this paradigm, the things that matter most in our developer tools are shifting, too. From Writing to Reviewing Historically, developers measured the productivity of a programming language largely by how easy it is to write . But when a coding agent can generate hundreds of lines of syntactically valid code in seconds, the rate at which a human can write code is no longer very important.…