Commit

AI is making me a faster developer, but is it making me a better one? I've been writing PHP for nearly 30 years (since PHP 3), and recently I've been thinking a lot about how AI is changing the way we all work. I've been experimenting heavily with AI-assisted development, coding agents and tools like Amp / Claude / Codex etc for a few years now, and the productivity gains are undeniable for sure. I'm shipping features faster, raising pull requests quicker and getting through work that would previously have taken considerably longer, but there's something about it that makes me uncomfortable and feel a bit "Icky". Historically, when I built something, I understood every single line. Not just what the code did, but why it worked, how the pieces fitted together and what might happen when something went wrong. That understanding came from spending hours writing code, debugging problems, making mistakes and figuring things out. Now? I can describe what I want, review the implementation, run the tests and have something working in a fraction of the time, but I'm increasingly finding myself understanding the solution at a higher level rather than knowing every detail of its implementation. Are we becoming better software engineers, or simply better at directing AI to write software? does it even matter? do we need to know all the implementation details or are we abstracted away from that now? I don't think avoiding AI is the answer. The productivity benefits are too significant, and the industry is clearly moving in this direction, but I do wonder whether we're gradually trading depth of understanding for speed of delivery. (product managers and CEOs rejoicing) I'm still figuring this out, my opinion changes depending on which way the wind is blowing btw :) Are you finding AI is improving your engineering skills, or are you starting to feel further removed from the code you ship?