Now I'm worried. I guess we won't be talking about detectors and filtering of AI generated crates etc any more.
I suggest to attach a URL to the official announce: Rust Foundation Welcomes OpenAI As Platinum Member, Announces Donation to Rust Project - The Rust Foundation
As far as I understood (from the Reddit discussion too), it's more or less of just a donation. The seat does not give a vote, and does not affect decision making in the language development process.
The fact of sustainable donation by itself is a good news, in my opinion. I was worried that the regular RustLang contributors are underpaid for their hard work. Hopefully, that would improve the situation. And I hope it would stimulate other bigtechy to contribute into Rust more regularly.
On a general note, the fact that the new generation of IT companies have interest in Rust is also a good news. And I would say it's very natural. New languages no matter how good are they, rarely supersede already filled niches. The main chance for Rust as a relatively new language (comparing e.g. to C++) is to become recognized by the new IT companies, and to fulfill new niches. The previous generation could eventually adopt it too, but the locomotive is usually the new generation of ITs.
To note, one such locomotive was block-chain businesses in recent years.
Ah good. What you say there makes me feel better about it. I'm all for industry, AI or otherwise, contributing to the development of the tools they use and depend on. As long as they can't use their money to steer the project in bad directions.
Actually I think AI and Rust are something of a good fit. The LMM tries to palm you off with slop and the Rust compiler being so fussy about types, lifetimes, etc says no, try better. That must be whole classes of problems avoided compared to using lesser languages.
Yes, self-verifying deterministic system certainly helps. I didn't have a chance to use LLMs for programming, but my friends who actively use them together with Rust, confirm it. Rust compiler at some extent plays a role of tests for the LLM output.
Speaking of the fitness function. Perhaps ML languages (especially those with dependent type systems) might also be of interest. But they are lesser recognized, unfortunately.
I'm still of the mind that I should at least take a look at the LLM generated code I try and determine that it looks like it does what I want and does not do anything I don't want. I want to see that any LLM generated tests do actually test what they claim to. Basically code review like any other. So of course Rust's fussiness removes the need to worry about whole classes of details in that review.
The chance to use LLM for code generation is only a Google search away. I'm amazed how much code I can get the "AI mode" to generate for free.