this post was submitted on 24 Jan 2024
270 points (90.9% liked)
Open Source
31365 readers
246 users here now
All about open source! Feel free to ask questions, and share news, and interesting stuff!
Useful Links
- Open Source Initiative
- Free Software Foundation
- Electronic Frontier Foundation
- Software Freedom Conservancy
- It's FOSS
- Android FOSS Apps Megathread
Rules
- Posts must be relevant to the open source ideology
- No NSFW content
- No hate speech, bigotry, etc
Related Communities
- !libre_culture@lemmy.ml
- !libre_software@lemmy.ml
- !libre_hardware@lemmy.ml
- !linux@lemmy.ml
- !technology@lemmy.ml
Community icon from opensource.org, but we are not affiliated with them.
founded 5 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
The writing was on the wall when they established a generative AI using everyone's code and of course without asking anyone for permission.
It's an interesting debate isn't it? Does AI transform something free into something that's not? Or does it simply study the code?
@xilliah It's not free though. It came with licenses. And LLMs don't have the capability to "study", they are just a glorified random word generator.
Ok
There's no debate. LLMs are plagiarism with extra steps. They take data (usually illegally) wholesale and then launder it.
A lot of people have been doing research into the ethics of these systems and that's more or less what they found. The reason why they're black boxes is precisely the reason we all suspected; they were made that way because if they weren't we'd all see them for what they are.
The reason they're black boxes is because that's how LLMs work. Nothing new here, neural networks have been basically black boxes for a long time.
Sure, but nothing is theoretically stopping them from documenting every single data source input into the training module and then crediting it later.
For some reason they didn't want to do that of course.
Llama and stability AI published their sources, did they not?
The reason they are blackboxes is because they are function approximators with billions of parameters. Theory has not caught up with practical results. This is why you tune hyperparameters (learning rate, number of layers, number of neurons ina layer, etc.) and have multiple iterations of training to get an approximation of the distribution of the inputs. Training is also sensitive to the order of inputs to the network. A network trained on the same training set but in a different order might converge to an entirely different function. This is why you train on the same inputs in random order over multiple episodes to hopefully average out such variations. They are blackboxes simply because you can't yet prove theoretically the function it has approximated or converged to given the input.
Can you link it please? I'd like to inform myself.
I doubt they have a factual basis for their opinion, considering
Is just plain wrong. Researchers would love to have a non black box AI (i.e. a white box AI), but it's unfortunately impossible with the current architecture.
Their use of language also feels more emotional and if anything it makes me more skeptical.
No, it's exhausting.