I’m curious if any other people out there feel similar We’re over three years in, and I’m finding that the “productivity” boost these tools give me is a faux-boost.
They have, however, increased the quality of my work, since I have new tools that help me iterate, refactor, and refine better than I can when I am just coding in total isolation.
I’ve always learned the most from other developers (and still do), but LLMs are a great alternative when you just need some different takes on what you’re trying to accomplish.
I don’t necessarily complete projects that much faster, nor do I really have any interest in doing so, IF it means sacrificing insight, understanding, or maintainability.
I find that LLMs, especially agentic workflows, impact all of those negatively when I actually care about the quality of the work I’m outputting.

Lars Faye | Confident Coding

A developer shares their experience that LLMs haven't actually increased shipping speed or feature output, but have improved code quality by enabling better iteration, refactoring, and refinement. The author values LLMs as a sounding board similar to peer collaboration, but warns that agentic workflows can negatively impact insight, understanding, and maintainability when quality matters.

LLMs don't actually make me more "productive", or ship more features in a shorter amount of time.

If you’re working on only one project at a time or in a technology you’re already familiar with, then yes — it’s quite possible that the boost won’t be significant, or there might even be a drop in productivity. If, on the other hand, you have multiple projects and/or you’re working in a technology where you know you’d normally have to constantly check the documentation or video tutorials, then the gain is much greater.