A Developer's Guide to Generative AI in FreeBSD

[MG]

What about a noob guide? There are no specific local requirements beyond computing power and storage, so it can be 1 independent universal application like a port.

 

[Alain De Vos]

.

 

[pollito_cpc]

cracauer@ just wanted to say that I think this is an excellent guide, and it should be official. Thank you for writing this - I will be using it from now on for the other projects I'm involved in.

 

[cracauer@]

cracauer@ just wanted to say that I think this is an excellent guide, and it should be official. Thank you for writing this - I will be using it from now on for the other projects I'm involved in.

I did not write this, I just linked to it.

 

[hunter0one]

As long as developers practice what they preach in this guide, it sounds good to me. I can't help but think of poorly coded software (because it was lazily vetted) when I think of LLM's.

 

[DanielBowman]

Rule #1: You are ultimately and solely responsible for every line you commit.

That about sums it up.
If you dream it yourself, overheard in the pub or borrowed it from StackOverflow, read it in original pre-print of C language or some probability machine came up with it, does not matter in the end.

 

[hedwards]

That about sums it up.
If you dream it yourself, overheard in the pub or borrowed it from StackOverflow, read it in original pre-print of C language or some probability machine came up with it, does not matter in the end.

And that's an issue because if you're using an LLM to write the code you don't have a license to do so and the LLM likely also doesn't have a license to it either.

I think Linus is correct in that it's going to be increasingly hard to identify LLM generated code over time, unless they're sloppy, which makes the whole thing a massive mess either way and I don't think anybody really wants a repeat on the A&T lawsuit or the SCO suit that just wastes a tremendous amount of time and money without any real merit.

 

[DanielBowman]

And that's an issue because if you're using an LLM to write the code you don't have a license to do so and the LLM likely also doesn't have a license to it either.

I think Linus is correct in that it's going to be increasingly hard to identify LLM generated code over time, unless they're sloppy, which makes the whole thing a massive mess either way and I don't think anybody really wants a repeat on the A&T lawsuit or the SCO suit that just wastes a tremendous amount of time and money without any real merit.

Intellectual Property is just scam for lawyers anyway.
LLM or someone copied part of my code. Yey, nice to be of service. I can even brag that my code was used in some successful (nobody sues the other ones) project. They might even hire me.

If you don't want it copied, lock it in the safe with the recipe for Coca Cola.

But you are right, lawyers and other parasites exist and someone will eventually sue. Maybe Scam Altman veryClosedAII will ask for fees on every project. Won't surprise me one bit.