from Hacker News

How I learned to stop worrying and love the LLM

by jasonpeacock on 2/10/25, 5:54 PM with 1 comments

• by theamk on 2/10/25, 6:39 PM

  I don't think the core language should have AI in it for many practical reasons - offline operation, toolchain size, resource requirements, determinism, etc.. There is also a whole problem of LLMs non-deterministically generating incorrect output from correct inputs, and one of the Rust's values is that "generating incorrect output from correct inputs" is bad, and if it's non-determenistic, it's even worse.

  but IDEs are a different beast - if a user likes AI they can install IDE or a plugin to help them, and there is no problem if that plugin requires huge downloads or tons of resources or always-on internet connection with paid subscription. And there is no impact on Rust's reputation if that plugin suggest nonsense.