Fast and Hard Code | Armin Ronacher's Thoughts and Writings

Armin Ronacher's Thoughts and Writings

blog<br>archive<br>projects<br>travel<br>talks<br>about

Fast and Hard Code

written on August 22, 2026

One of the memes on Twitter is that "programming is solved now." I’m not sure<br>to what degree it is, but one thing is pretty clear: the act of familiarizing<br>yourself with a language no longer matters and some of the friction that<br>mattered for humans does not matter for agents.

As a result, LLMs make language choice much less consequential than it used to<br>be. If you don’t like the choice, you can seemingly rewrite it in another<br>language and you can make it pick a language that you, as a programmer, are<br>entirely unfamiliar with.

Which in turn means that people can, and do, choose based on the marketing of<br>languages much more. As a long-term Rust programmer I found it quite<br>fascinating to see people now ship Rust code who previously might not have<br>chosen it. I attribute at least one part of this to two recent vibe shifts:<br>there is a lot more talk about wanting fast software, and about LLMs being<br>exceptional at optimizing code without regressing behavior.

Folks like Mitchell Hashimoto, Charlie Marsh, Jarred Sumner, Daniel Lemire and<br>quite a few others always carried a certain level of obsession with fast and<br>performant software and they also all happen to be receptive to agents writing<br>code. Maybe as a result, or unrelated others are now joining in. That’s<br>because with things like<br>autoresearch you don’t even<br>necessarily need to know all the tricks: you just need to put an agent on it —<br>though knowledge greatly helps!

If you look around, there are plenty of projects that want to be fast and small,<br>and they increasingly pick "hard languages". And it’s not just Rust that is<br>benefiting. Even Zig — despite the fact that the creators and parts of the core<br>community are pretty negative on the whole AI thing — is too. For instance<br>Cloudflare’s new<br>Artifacts service<br>uses a pure-Zig Git-protocol engine, compiled to a roughly 100 KB WebAssembly<br>module and Vercel released fx, a Zig coding<br>agent advertised to be small and fast. From what I can tell, all these projects<br>are largely LLM-assisted.

But it’s not just people picking less common languages but also that they are<br>increasingly working with "much harder" technologies. All of a sudden I have<br>seen people do some really impressive stuff with DWARF files, eBPF, custom<br>network drivers, custom crypto and really old computing hardware. Many of these<br>things were previously off-limits for lots of developers. In some cases (eg:<br>crypto) you were even pushed away because those things were intentionally<br>gatekept by the people in the know.

So maybe the world will have more slop, but it might also have more developers<br>in it, that want things to be fast and small.

This entry was tagged

ai,<br>programming and<br>thoughts

copy as / view markdown

&copy; Copyright 2026 by Armin Ronacher.

Content licensed under the Creative Commons<br>Attribution-NonCommercial 4.0 International License.

Contact me via mail,<br>bluesky,<br>x, or<br>github.

You can sponsor me on github.

More info: imprint &<br>AI transparency.<br>Subscribe via atom / RSS.

Color scheme:<br>auto,<br>light,<br>dark.