“Make it anyway”: a developer’s essay on making things in the age of generative AI

An essay about creative work

Developer Joel Otter has published an essay in which he describes a difficult week: not a dramatic explosion, he writes, but an implosion in which his resolve and motivation slowly crumbled. He says he is writing it partly for anyone going through the same thing, and partly for his future self.

The reason is familiar to anyone in a creative field: generative AI and the ongoing devaluation of the work of artists and makers. On the programming side, where he mostly works, he finds the situation more uncertain; in the visual arts, he writes, AI output is bad — in a sense anti-art, embarrassing to look at. He also notes that almost every programmer he knows has changed over the past couple of years, and calls it the most turbulent time he can remember for software engineering. Denying that large language models are genuinely good at code generation, he argues, is racing against a moving finish line.

Small tools and a single prompt

Otter stresses that his aim is not to convince anyone of a position. His own is simple: he does not enjoy programming with a code assistant, the output does not feel like his, and he takes no pride in it. What bothers him most is that the small things he was proud of are now available to anyone. He points to his interactive git branch switcher, a tool he uses every day and that colleagues praised; today, anyone can order such a tool, or something similar, from a prompt. He admits he needs that recognition from peers to feel fulfilled.

The same pressure reaches game development. Otter says his low point was triggered by a thread from designer Zach Gage arguing that making games is becoming more like making music. He found the insight good but devastating: years spent learning to build games suddenly felt as though they might have been wasted.

Uncamera and three possible paths

A conversation with designer and engineer Shad changed his framing. Shad’s project, Uncamera, is an iOS camera app that uses raw sensor output and lookup tables rather than post-processing filters to produce photographs that look like film. It is built entirely without generative AI, for exactly the reason Otter cites: the making itself is where the fulfilment lies.

Otter concludes that only three paths are open to him. He can start using generative AI to feel he is keeping up — and lose the enjoyment of the work. He can stop making things, which he considers no real option for a creative person. Or he can keep making things the way he enjoys, keep learning, and keep doing it the hard way, with no other justification than wanting to. He had already chosen the hard way, he notes: his own game engine written in C++ rather than Unity, Godot or Unreal. That was never about speed, but about enjoyment and learning — and, as his closing line puts it, that is why we make games, and why we make anything.