The answer, it turned out, was that seniority still mattered — just differently. The AI could write code, but it couldn’t tell you what code to write. It didn’t understand your business requirements, your users, your technical constraints. It didn’t know which shortcuts would come back to haunt you and which were fine. It would confidently generate solutions to the wrong problem if you weren’t careful. The job shifted from writing code to directing code — knowing what to ask for, evaluating what came back, understanding the system well enough to spot when the AI was leading you astray. The discourse was polarized, predictably. Some people declared that programming was dead, that we’d all be replaced by AI within two years. Others dismissed the whole thing as hype, insisting that AI-generated code was buggy garbage that real engineers would never use. The truth was somewhere in between and more interesting than either extreme. AI didn’t replace developers, but developers who used AI became noticeably more productive. You could attempt projects that would have been too tedious before. You could learn new domains faster. You could build more. I found myself building things I wouldn’t have attempted a few years ago. Side projects that would have taken months became weekend experiments. Areas where I had no expertise — machine learning, game development, unfamiliar frameworks — became accessible because I could have a conversation with something that knew more than I did. The barrier to trying new things dropped dramatically.
I know I’m including a snippet here about the impact of AI on development, but don’t sleep on the rest of this post. It reads like my own history from a folder on a unix server all the way to present day. Amazing read and great screenshots.
Quote Citation: artmann.co, “30 Years of
Tags”, December 13, 2025, https://www.artmann.co/articles/30-years-of-br-tags