Summary

Armin argues that LLMs have decoupled language familiarity from language choice, making the decision increasingly driven by marketing and performance aspirations rather than developer expertise. This has led to a surge in adoption of 'hard' languages like Rust and Zig by developers who previously wouldn't have chosen them. The vibe shift toward valuing fast, small software — combined with LLMs' ability to optimize code without regressing behavior — is the key driver. Beyond language choice, LLMs are also unlocking previously gatekept technical domains like DWARF, eBPF, custom network stacks, and cryptography for a broader set of developers.

Key Insight

LLMs haven't just lowered the barrier to using hard languages — they've shifted language selection from skill-based to aspiration-based, with 'fast and small' becoming an achievable goal rather than an expert's domain.

Spicy Quotes (click to share)

• 6

  The act of familiarizing yourself with a language no longer matters and some of the friction that mattered for humans does not matter for agents.

• 5

  People can, and do, choose based on the marketing of languages much more.

• 3

  With things like autoresearch you don't even necessarily need to know all the tricks: you just need to put an agent on it — though knowledge greatly helps!

• 4

  There are plenty of projects that want to be fast and small, and they increasingly pick 'hard languages.'

• 7

  In some cases (eg: crypto) you were even pushed away because those things were intentionally gatekept by the people in the know.

• 6

  Maybe the world will have more slop, but it might also have more developers in it, that want things to be fast and small.

Tone

observational, cautiously optimistic, analytical