vim, harfbuzz, who's next? people ask in shock, but...

i think, fundamentally, the reason Claude and Codex are becoming part of crucial FOSS projects is the same reason xz almost became the entry point for mass-scale server hacking a few years ago. we've decided to make billion dollar industries rely on burned out, lonely individual developers who never found a way to get paid for their labor. we never managed to solve that problem.

these burned out, lonely devs see a tool that spits out the boring part of their work in a more or less functional manner if you squint, and it "only" (*) costs "$200" (**) a month. i can imagine why most people take it. heck, i won't deny that i am tempted myself, but my convictions remain too strong.

i've also seen some say that these people should step down and make way for new developers. who, exactly? i know how many months it took me to find a maintainer for one of the more popular Minecraft mods, and that's a position with both far more takers and far less responsibility on either side.

i think that a lot of what's going to happen to software in the next few years is the consequence of long term systemic issues. the introduction of LLM tools to the equation is merely an illuminant and accelerant