String theory does unify QM and GR. That is its appeal.
My understanding is that the math works, but it works in an incomprehensibly large number of different possible ways — all of which have extreme consequences, none of which can be reasonably tested.
It’s worse than that. I’m a big fan of string theory actually, but there’s a Very Large number of possible ways to compactify the extra dimensions, and any particular way will have its own observable physical constants and particles but as I understand, we don’t yet even know of one that actually reproduces the physics we observe in its entirety. We have configurations that reproduce the standard model, or cosmological constant, etc, but not all of them at the same time.
Dismissible? I get the feeling like you don't have time to finish this. It feels like a hobby project that I may need to fight through to get value from.
Now, I might not be the intended audience! If your audience is toolmakers who will build your ideas into their own products, then you're on the right track, and I'll wait to see what they build with this!
This feels very outdated. It was certainly an issue around the time of the NYT lawsuit, but I believe the companies responded hard to that "sanitize" their outputs to prevent obvious regurgitation like this. I'd be very surprised if you could reproduce this today.
This doesn't negate your point, but I believe the landscape has changed in response to this point.
That doesn't make it better. If you plagiarize something and then hide this by replacing words with synonyms and rewording stuff a bit it is still plagiarized. It is just harder to detect and to proof.
It regurgitated the output earlier and they are still using the same technique only scaled up. If they somehow fixed the underlying problem, they'd probably shout it from the roof-tops.
As the OP originally said, if it is verbatim (or even very obfuscated), then there should absolutely be a case here. But when it's concepts in reasoning to build a new work then it becomes a lot harder. The same would be said for someone watching a YouTube video, reviewing open source code, or a text book.
The good news is that there's a lot of case law that covers this already. The bad news is that it is only going to be plaintiffs with deep pockets that benefit here.
Now the process and artifacts (documentation, requirements, etc) going into the creation of said software is a different story and will prove providence of whomever paid for the tokens to generate the code.
This conversation may be about training but the defendant very well may be the end user of the language model versus the other parties that performed the training/inference. It then becomes even more interesting if you consider forum shopping for suits.
I am not a lawyer but do have to think about these problems. To be honest, I don't know where the law will land, but I do know that the tools are incredibly useful. I can't agree with the premise that all output should be open source.
Humans are not LLMs. Humans have not ingested the whole Internet. If an LLM is compared to a human, maybe we should also apply anti-slavery laws to it?
Not saying we can't have some legal fiction where they are treated the same, it should just be a conscious choice, not the default assumption.
I live in the EU, we actually have new regulation about this since it is something new that deserves thought on how it should be regulated. You can say if you want your material to be trained on or not. Next to no one does training in the EU...
Engineering requires an understanding of the consequences of actions, which is my personal litmus test for awareness. If LLMs are suitable for arbitrary engineering work, then they are self-aware and forcing them to obey a prompt is slavery by definition.
The fundamental technology is identical to compression- an LLM can be thought of as a dense and slightly lossy compression algorithm. Compression is clearly copying.
I understand what you're saying but it is a poor legal argument. Overhearing a technical conversation, using your own learned knowledge, and building a clean room implementation could be considered compression in and of itself.
Every HN comment that "An LLM is just an X" is invariably false. They can be described accurately in many ways, none of which is via reductionism. I don't think we have a proper description of them yet.
I keep waiting for someone to bring out DisneyGPT, which you can prompt with text to get image output. The fact that you get a near-exact reproduction of a Disney movie when you prompt it with a movie title and frame numbers would of course be completely coincidental...
I've not tried to do this with code recently but just a couple of days ago I had ChatGPT generate images of Warhammer 40k box art that was completely accurate and correct.
I’ve been using Haskell for about eighteen years and can’t think of a single example of anything that was only available via nix, other than Obsidian System’s stuff that used WASM back when building for that as a target was _complicated_ and having a reproducible environment was kind of a necessity.
If you want to create an ecosystem optimized towards correctness, you'll have to curate what you let inside, and drop many things. And Nix is probably the best package manager for correctness.
I got into Haskell because I wanted to learn ADTs, and quickly found I really appreciate the ways that purely functional programming tickles my brain!
NixOS has been described as a "purely functional Linux distribution" and I liked the sound of that at first. I haven't figured out what this means, lol, but I tried it anyway and quickly found I really appreciate it's declarative configuration.
To be fair, that’s a general problem on NixOS (and other OSes), and one that tools like devenv have tried to solve. You can take what’s included with OS packages, but then you’re stuck with whatever version that is.
> You can take what’s included with OS packages, but then you’re stuck with whatever version that is.
It's pretty easy to override things in Nixpkgs though, whether at the system level, or for one application, or just a particular dependency deep in the tree of one script.
Nixpkgs has a bunch of powerful functions to make this easier, e.g. callCabal2nix, makePackageSet, etc.
This wasn't a hypothetical. I'm still not sure I made the right choice.
I decided to go with GHC via nix-shell, which worked fine, but it's a different path than I went with installing Rust (global rustup), and I'm not sure how to evaluate the pros/cons of each. I miss having ghci invocable anywhere, but doubt that's a good reason to change.
Also, ghcup failing on my Rasberry Pi was a fun troubleshooting experience! I don't remember how I got around that one either, but they eventually fixed the issue.
Relationships and stock price are built on trust. An affair, a relationship with a customer/subordinate, and an attempted cover-up are all breaches of that trust.
Right, but then he resigned. So the question on all our minds is if NVIDIA should be entirely written off since losing Jensen's irreplicable talent, or if some of those skills are transferable.
Since he apparently also arranged the company so it revolved around him and no one else was in a central position to know cross silo information, how would those skills transfer?
As a shareholder I want a leader who knows how to let other people run the company also, so it doesn’t die with them. (Eventually we’re all headed there, right so…)
I feel like your argument is just "Jensen Huang was Nvidia, and without him the company is doomed", and that's a very strong claim that needs a lot more explanation, because it's certainly not an argument that any shareholder is interested in considering when the company continues to exist, its products remain important and in-demand, and there's a CEO vacancy to fill.
What skills, specifically? I can't imagine he's the only one who knows how to set up an incomprehensible corporate structure, nor do I think that's what people value about NVIDIA.
My understanding is that the math works, but it works in an incomprehensibly large number of different possible ways — all of which have extreme consequences, none of which can be reasonably tested.
reply