I agree with this; I just also note that ignoring the dev and having them dev other things is cheaper (and hence we get more (insecure) things)
Code security is hard, especially for humans. You need to notice something a bit off between 3+ different pieces of code, and/or have encyclopedic knowledge about quirks of the programming language, compiler, and/or hardware.
We could have been working much harder for secure code, but it would have cost a lot before. Many orders of magnitude more man-hours, substantially higher pay and recruitment for the least pleasant parts of coding. And plenty of apps simply wouldn’t be done.
(Note that you only need one script-kiddy to notice something with an LLM, and then everyone on the internet has to patch it. Plus LLMs make ~everyone able to do this.)
let people effort post.
One can be a capitalist while alive, but not while dead. We’ve seen several reasonable examples.
I suspect no, if they haven’t actually trained a model yet?
Presumably you mean after they actually train the thing. Then I assume yes, but I’m unfamiliar with the law.
They could use self-play and existing ethical models, like Apertus.
That said, they don’t disclose afaict. Haven’t trained it, so they probably don’t know themselves.
0 points on All AI-generated code should be Open Source in !technology@lemmy.world ·
(I think I’m in favor of this fwiw)