AI has advanced to the point where it can analyze code for correctness, recommend fixes, and even catch bugs. This has made technical competence less unique, as many developers now rely on AI tools for reviews and testing. However, AI lacks the ability to judge whether a solution should exist in the first place. That judgment, or "taste," comes from deep experience and understanding of design principles, which AI cannot replicate. The author illustrates this point through a personal experience of rejecting a Linux kernel patch that had passed AI review. While the patch was technically sound, it had design flaws that would lead to long-term issues, such as increased complexity and hidden constraints. These concerns stemmed from a nuanced understanding of the system, which AI is not equipped to evaluate. Taste is about recognizing these subtleties, something developed over years of experience and exposure to both successful and failed ideas. As AI continues to automate basic tasks, the focus shifts to who makes the important design decisions. Questions about what belongs in a system, recognizing potential pitfalls, and understanding the broader implications of a design are now more important than ever. This shift underscores the value of human judgment, shaped by experience and taste, in making decisions that go beyond mere correctness. The article stresses that while AI can assist in identifying mistakes and suggesting improvements, it should not be the final arbiter in critical engineering decisions. The essence of good design comes from human insight, not just adherence to rules. In an environment where execution is easy and correctness is common, it’s personal taste that truly sets apart the best engineers.