This may be annoying to some users; I'm trying to understand how AI agents may impact Rust developers / programmers. Also, how people here think of AI, whether they use it and how much. (Besides simple autocompletion or SO replacement.)

Some may simply recommend to test whether it is useful to you and that is it.

But I also wonder what the trend will be like in a few years time. Are normal coders replaced by then? How much of your work can it do right now (assuming you have tested extensively and not just prompted ChatGPT) ? Do we relegate ourselves to prompt "engineers" in a few years time ?

I'd prefer not to state my position or current understanding besides saying that I have not explored as much as other people seem to. Some claim it's great, even experienced devs (For example, Syntax.fm uses AWS's Kiro in a recent tutorial but this is in some JS framework.)

Recently, I read the post linked in "This week in Rust" newsletter about using LLMs and agents for coding, and even the future or role of programmers.

I have also noticed some seasoned rust devs use it but didn't quite check the details.

It is also clear to me that researchers are using AI and LLMs for automation including discovery (although they aren't good or can not produce much novelty). I take this from other fields and would prefer not to give names, but I mean well-cited researches.

A less pretentious stance is given in Why bad AI is here to stay.

A new study claims programmers are 20% slower with AI tools (n 16). But other study claims the exact opposite. (n 96)
I won't link them unless someone requests them, I don't think either tells a full picture.

Do you have thoughts on the matter?

15 posts - 8 participants

Read full topic