Gopichand Kotana

There was a time when mastery meant spending years inside a single way of doing things. To an average person, anyway.

A programmer learned a language and its idioms, a photographer learned how light fell through a particular lens, a writer learned a voice and stayed with it. Style emerged slowly, through repetition and constraint. People then applied the mastery and style to solve challenges in their field.

AI tools have shaken up this arrangement. A person can imagine almost anything and ask for it directly. The limitation is no longer access to the technique but the clarity of the prompt and the patience to refine it. The tools prescribe multiple solutions, ranging from competent and elegant to downright disastrous. The tool does not usually decide the outcomes. Agents are an exception, but I believe they won’t be used for general problem-solving. People will mainly use them to solve a problem in changing contexts. An example of that would be flight ticket booking. We largely know how to do it, but depending on the person, their location, and their preferences, the agent has to make certain decisions.

Returning to AI tools, they often lack a lot of context. Let’s take a software project as an example. It is unaware of the project’s history, the unwritten team conventions, certain parts of the system that no one in the firm knows exactly how they work, or the emotional cost of complexity and changing styles. It offers possibilities, but the act of choosing remains stubbornly human. None of the suggestions is wrong in isolation. However, only one may be right for this codebase at this moment. So the role of the developer shifts from construction to judgment.

Taste becomes key to everything. Choosing what is appropriate becomes super important. Especially if the claim that AI tools shorten the iteration cycles becomes true, all the bad considerations multiply very quickly. Expertise looks less like invention and more like discernment.

AI expands the field of possibilities, but taste determines where you land. Craft still matters, but it has moved. It lives in the ability to recognize quality rather than to produce it from scratch. If there is a risk here, it is not that taste becomes irrelevant, but that it becomes underdeveloped. We are paid, increasingly, not for our ability to produce, but for our ability to notice. Attention is all you need?