Joel Auterson makes games and small tools, and he wrote this one after a bad week — a collapse of motivation he traces directly to generative AI. His stated position is narrow and specific: setting aside the environmental and social arguments entirely, he simply does not enjoy programming with a code assistant. “It isn’t fun for me, the output doesn’t feel like mine, and I take no pride in what it produces.”

The sting is not that the work is gone but that the audience for it is:

  • He loves making fiddly little things — he names a git branch switcher he uses every day and was proud of. Colleagues said nice things. Now anyone can order that tool, or anything like it, off a prompt, so nobody cares. He is candid that he needs those “head pats” from peers to feel fulfilled.
  • What set off the collapse was someone else’s argument that game making is becoming more like music, where the recorded artefact is cheap and the live, human event is what people pay for. After years of learning an expensive skill, that read to him as “maybe that was a waste of time.”
  • A conversation with a friend — Shad, who builds an iPhone camera app without generative AI, for the same joy-related reasons — reframed it as a choice between three paths: use AI to feel like you’re keeping up and stop enjoying the work; stop making things, which creative people rarely actually do; or keep making things the hard way for no reason except wanting to.

He is already on the third path by temperament — his current project is a game engine he wrote himself, in C++, with a faux-3D renderer. If speed were the goal he would be using Unity or Godot. The essay ends where its title does: “That’s why we make games. That’s why we make anything.”

What the thread adds

The 394-comment thread on Hacker News is the argument the essay deliberately declines to have, and it is far less unanimous than the post itself. Both sides have receipts.

  • DidntUseIt — the strongest counterexample to the gloom, from the other direction: without LLMs they were “making a contact lead form with some basic stuff and calling it good. With them, I’m making the whole SaaS.” They shipped invoicing, e-signing and email marketing solo, and argue that a developer refusing agents now only gets to do beginner work. embedding-shape disputes the premise: “If what you’re working on doesn’t exists in the training data at all, the LLMs can sometimes be more in the way than helpful… the moments I’m working without LLMs, that’s the time I’m working very much on ’not beginner stuff’.”
  • user43928 — the mirror image of the essay: 15 years coding, and “It’s never been so fun,” a 200k-line app they no longer review line by line, no sign of the model losing the thread. bradly cuts at the confidence rather than the claim — hundreds of hours and a token budget on something with no users or revenue yet, “I’m not sure it gives you the authority to calling people clueless for having a different opinion.”
  • Tade0 — the craft argument that does not depend on enjoyment: “I refuse to leave understanding the codebase to machines.” A human reviewer caught redundancy their automatic reviewer missed, and they explain why they think it keeps happening — “LLMs won’t ever create elegant and minimal code, because their style is based on the mean.” jeremyjh refines it: that describes a pre-trained base model, while post-training now dominates the behaviour of coding agents, and “the best code for an LLM is different than the best code for a human” — lots of redundancy and defensive code is reinforced because it works for the model. mpetrovich names the part that worries him: writing code used to force you to understand it, “but now the former doesn’t guarantee the latter.”
  • MachineMan — turns the essay’s own question back on its author: “Did you enjoy the crafts, or did you enjoy the compliments?” 1dom supplies the material version — the value of a career built on “thousands of solitary hours from a young age” was that it was scarce, and “that asset has been somewhat synthesized now.”
  • x3haloed — the harshest dissent: “This is 80% ego talking and 20% fear of change… You can find a way to produce unique and outstanding results that still matter, or you can slowly fade into a hardened stance of obsolescence.” pastel8739 answers with the obvious objection — “Do you actually believe that people don’t know whether they are enjoying a task or not?” — and bigmadshoe locates the real injury somewhere neither side names: not peer approval, but the finished thing. “It feels fundamentally bad to know that however hard I worked on this, it would have been better, faster, or both had I used AI to produce it.”
  • djha-skin — the most concrete prediction in the thread, from someone who tried the hard-way project and ended up handing docs, tests and the readme to an agent anyway: hobby programming is entering its “sports” era, like a jogger who drives to the top of the hill and runs down the parts they enjoy. saejox lived the cost — after 20 years they abandoned a second Steam game because “this 1000 lines of code could have been just a prompt.”
  • fsloth, Oras and timcobb — the durability argument, in three versions: nobody takes an enjoyable way of working away from you; no company will let non-developers push generated code to production; and the respect was never there to lose — “No one cares about the code” was beaten into them 10 to 20 years before generative AI. fsloth’s analogy is chess: “More players than ever, perfectly obsolete if you compare yourself to a computer, but that does not hurt.”

Pseudonymous handles, no per-comment scores, and the ordering is Hacker News’s own ranking, so this is a slice of a thread and not a consensus.