Things are changing way too fast with the arrival of AI. I think most of us programmers no longer write code "by hand" like we used to. This has radically changed our profession. For those of us who love programming, writing code, and those of us who love our editor because we've always used it (did I tell you I've been using Emacs for more than 20 years?), those of us who love TDD and extreme programming, this feels weird: going from writing code to writing prompts, which isn't the same thing... It doesn't feel the same. I'm not saying it's good or bad, I'm just saying it doesn't feel the same, although I've managed to get into "flow" many times doing "Vibe Coding" and also "Agentic Programming/Agentic Engineering". I learned that last term recently from Simon Willison. I feel that for those of us who are more or less like this, and maybe for most people too, going from writing code or programming to interacting with an agent inevitably involves some grief. And the sooner we accept it, the sooner we'll be able to move on to what's new, which is exciting too, different of course, but exciting. Now the limits of what we can do in a short amount of time have shifted, and that's a good thing.
On the other hand, I don't think there's a settled opinion on the best way to generate code or create a program. What's the "best" way? Is it writing a prompt that's as detailed as possible? Is it interacting in baby steps? Is it doing TDD? Should we read the generated code? How much of it should we read?
All these questions carry one word: trust. How do I build something by vibe coding and then trust it? Well, I think all these questions and more are questions the entire IT community is trying to answer, and for now, from what I can see, there are no absolute answers.
Spanish original: C1 correction · AI intervention: I1/5 — spelling, accents, punctuation, and formatting. This English version was translated with AI. See the full scale (in Spanish).