Quality and understandability after AI (opens in new tab)

So, it’s February 2026 and state of the art AI agents (LLMs that can execute actions) can write code with tremendous speed and their output is not half bad. Boom. Where do we go from here?