I like Casey's reporting on social media and I like the #forkiverse stuff. This latest issue of Platformer has a paragraph:
In the week since, amid a general craze for Claude Code, I have let my imagination run wild. In just a few days, I built a series of tools I expect to return to over and over again to help me in my work. As with my website, I don’t understand the underlying code at all. But for the moment, anyway, I’m not certain it matters.
The thing is, the likelihood that he wrote something that didn't exist before is super low. He might have built some tool that is super Casey-specific, where he'd have to learn how to work a more general-purpose tool. It is very unlikely that a non-programmer will encounter a problem where there are no existing solutions—especially on something as basic as a chore related to running a web page.
So this building of basic tools feels like reinventing the wheel instead of learning. It's burning the climate to make a personalized solution to a long-solved problem.
While hubris is one of the three attributes Larry Wall said great programmers have, LLMs have allowd hubris to go too far. Now people who know very little in a domain feel like they can skip learning that domain and just make a personal tool with an LLM that will meet their needs.