I’ll preface this by saying I’m not an AI buff. I use it (unpaid) mostly for a search aid or assistance with setting up services in my homeserver (new to homelabbing) when I get stuck, as I don’t have an unlimited amount of time to devote to it due to life obligations.
My question is: Do you feel it is ethical and/or moral to use AI to develop FOSS?
I’m genuinely torn on this issue. Part of me feels it’s counterintuitive, yet the another part of me knows the probability of it going away are slim to none and using it to develop FOSS is a way to “use it against itself”, for lack of a better term.
This is my first post on Lemmy. I’m asking to please keep it cordial. I’m not looking for an elitist lecture. I genuinely want to know the community’s thoughts.
P.S. I’ve toyed with local models, but I don’t have the hardware to make them remotely useful. I’m still learning and tweaking to see what can be achieved on limited hardware.
The problem is that FOSS projects that are popular already have a a high quality standard because they are developed by very smart people who’s been developing for a long time.
There is inherently a problem with current AI - It doesn’t learn.
When a new developer joins a FOSS project they first spend weeks or even months learning this project's specific coding practices, standards, and reasoning. I’ve been programming for 20 years and have a PhD. I can’t “just start contributing” to something like KDE because I’ve never looked at its codebase. I don’t know assumptions, methodology, or challenges that are baked into the codebase. I need to spend time to learn those to be a useful contributor.
AI can’t learn unless you do gradient descend. The problem that you mention about sounding elitist is that those FOSS developers spend a good amount of their time training new contributors and everyone is proud of their work. Now you have people who are not ready to be contributors tell AI to contribute and it’s doing its best but it just can’t learn and even having a 200K token context is not enough to fit all those project specific assumptions and rhetorics in.
Frontier models today have a context window of 1 million. KDE is an interesting case, because not only does it have a serious spaghetti code problem, it's also fairly modular, so Ai could probably contribute very effectively to parts of it & perhaps even explain & comprehend some past developer choices that are not that human comprehensible anymore. (and being familiar with one of the current active KDE developers, I'd be surprised if he were not vibecoding like a boss)
The problem isn’t what can be done, the problem is what is being done. An AI in hands of a good developer can make that developer better - sure.
On the other hand what actually happens is equivalent to submitting code that doesn’t compile, except it’s not quite as easy to discover. The code does compile but it often is still bogus. Maybe it doesn’t solve the issue, maybe it does but in a round-about way, etc.
Legal and economical arguments aside I’m not saying “don’t use AI” - I’m saying “if you submit AI code - verify its work before making someone else verify its work”. And that means you need to know how to verify its work - just compiling and running isn’t enough.
How about the old contributors using AI to assist coding? Is that ethical?
I have no ethical problem with that but I can’t speak for the masses. But also it’s a bit in a grey area legally since it’s unclear who owns LLM outputs.
I also think there's no ethical problem. But the latter part seems to be completely legal (so it's specific to a country's laws) rather than ethical, which I also agree with.