• 0 Posts
  • 106 Comments
Joined 2 years ago
cake
Cake day: July 14th, 2023

help-circle
  • You don’t have to finish the file to share it though, that’s a major part of bittorrent. Each peer shares parts of the files that they’ve partially downloaded already. So Meta didn’t need to finish and share the whole file to have technically shared some parts of copyrighted works. Unless they just had uploading completely disabled,

    The argument was not that it didn’t matter if a user didn’t download the entirety of a work from Meta, but that it didn’t matter whether a user downloaded anything from Meta, regardless of whether Meta was a peer or seed at the time.

    Theoretically, Meta could have disabled uploading but not blocked their client from signaling that they could upload. This would, according to that argument, still counts as reproducing the works, under the logic that signaling that it was available is the same as “making it available.”

    but they still “reproduced” those works by vectorizing them into an LLM. If Gemini can reproduce a copyrighted work “from memory” then that still counts.

    That’s irrelevant to the plaintiff’s argument. And beyond that, it would need to be proven on its own merits. This argument about torrenting wouldn’t be relevant if LLAMA were obviously a derivative creation that wasn’t subject to fair use protections.

    It’s also irrelevant if Gemini can reproduce a work, as Meta did not create Gemini.

    Does any Llama model reproduce the entirety of The Bedwetter by Sarah Silverman if you provide the first paragraph? Does it even get the first chapter? I highly doubt it.

    By the same logic, almost any computer on the internet is guilty of copyright infringement. Proxy servers, VPNs, basically any compute that routed those packets temporarily had (or still has for caches, logs, etc) copies of that protected data.

    There have been lawsuits against both ISPs and VPNs in recent years for being complicit in copyright infringement, but that’s a bit different. Generally speaking, there are laws, like the DMCA, that specifically limit the liability of network providers and network services, so long as they respect things like takedown notices.


  • Why should we know this?

    Not watching that video for a number of reasons, namely that ten seconds in they hadn’t said anything of substance, their first claim was incorrect (Amazon does not prohibit use of gen ai in books, nor do they require its use be disclosed to the public, no matter how much you might wish it did), and there was nothing in the description of substance, which in instances like this generally means the video will largely be devoid of substance.

    What books is the Math Sorcerer selling? Are they the ones on Amazon linked from their page? Are they selling all of those or just promoting most of them?

    Why do we think they were generated with AI?

    When you say “generated with AI,” what do you mean?

    • Generated entirely with AI, without even editing? Then why do they have so many 5 star reviews?
    • Generated with AI and then heavily edited?
    • Written partly by hand with some pieces written by unedited GenAI?
    • Written partly by hand with some pieces written by edited GenAI?
    • AI was used for ideation?
    • AI was used during editing? E.g., Grammarly?
    • GenAI was used during editing?E.g., “ChatGPT, review this chapter and give me any feedback. If sections need rewritten go ahead and take a first pass.”
    • AI might have been used, but we don’t know for sure, and the issue is that some passages just “read like AI?”

    And what’s the result? Are the books misleading in some way? That’s the most legitimate actual concern I can think of (I’m sure the people screaming that AI isn’t fair use would disagree, but if that’s the concern, settle it in court).


  • Look up “LLM quantization.” The idea is that each parameter is a number; by default they use 16 bits of precision, but if you scale them into smaller sizes, you use less space and have less precision, but you still have the same parameters. There’s not much quality loss going from 16 bits to 8, but it gets more noticeable as you get lower and lower. (That said, there’s are ternary bit models being trained from scratch that use 1.58 bits per parameter and are allegedly just as good as fp16 models of the same parameter count.)

    If you’re using a 4-bit quantization, then you need about half that number in VRAM. Q4_K_M is better than Q4, but also a bit larger. Ollama generally defaults to Q4_K_M. If you can handle a higher quantization, Q6_K is generally best. If you can’t quite fit it, Q5_K_M is generally better than any other option, followed by Q5_K_S.

    For example, Llama3.3 70B, which has 70.6 billion parameters, has the following sizes for some of its quantizations:

    • q4_K_M (the default): 43 GB
    • fp16: 141 GB
    • q8: 75 GB
    • q6_K: 58 GB
    • q5_k_m: 50 GB
    • q4: 40 GB
    • q3_K_M: 34 GB
    • q2_K: 26 GB

    This is why I run a lot of Q4_K_M 70B models on two 3090s.

    Generally speaking, there’s not a perceptible quality drop going to Q6_K from 8 bit quantization (though I have heard this is less true with MoE models). Below Q6, there’s a bit of a drop between it and 5 and then 4, but the model’s still decent. Below 4-bit quantizations you can generally get better results from a smaller parameter model at a higher quantization.

    TheBloke on Huggingface has a lot of GGUF quantization repos, and most, if not all of them, have a blurb about the different quantization types and which are recommended. When Ollama.com doesn’t have a model I want, I’m generally able to find one there.


  • I recommend a used 3090, as that has 24 GB of VRAM and generally can be found for $800ish or less (at least when I last checked, in February). It’s much cheaper than a 4090 and while admittedly more expensive than the inexpensive 24GB Nvidia Tesla card (the P40?) it also has much better performance and CUDA support.

    I have dual 3090s so my performance won’t translate directly to what a single GPU would get, but it’s pretty easy to find stats on 3090 performance.





  • Further, “Whether another user actually downloaded the content that Meta made available” through torrenting “is irrelevant,” the authors alleged. “Meta ‘reproduced’ the works as soon as it made them available to other peers.”

    Is there existing case law for what making something “available” means? If I say “Alright, I’ll send you this book if you want, just ask,” have I made it available? What if, when someone asks, I don’t actually send them anything?

    I’m thinking outside of contexts of piracy and torrenting, to be clear - like if a software license requires you to make any changed versions available to anyone who uses the software. Can you say it’s available if your distribution platform is configured to prevent downloads?

    If not, then why would it be any different when torrenting?

    Meta ‘reproduced’ the works as soon as it made them available to other peers.

    The argument that a copyrighted work has been reproduced when “made available,” when “made available” has such a low bar is also perplexing. If I post an ad on Craigslist for the sale of the Mona Lisa, have I reproduced it?

    What if it was for a car?

    I’m selling a brand new 2026 Alfa Romeo 4E, DM me your offers. I’ve now “reproduced” a car - come at me, MPAA.


  • A paid skillful engineer, who doesn’t think it’s important to make that sort of a change and who knows how the system works, will know that, if success is judged solely by “does it work?” then the effort is doomed for failure. Such an engineer will push to have the requirements written clearly and explicitly - “how does it function?” rather than “what are the results?” - which means that unless the person writing the requirements actually understands the solution, said solution will end up having its requirements written such that even if it’s defeated instantly, it will count as a success. It met the specifications, after all.


  • Hedgedoc is fantastic. If you’re okay with your notes app being web-only (without an app or even a PWA) and you don’t need canvas notes or multi-note queries, you should check it out.

    First, every note is Markdown, but it supports a ton of things natively. It has native Vim, Emacs, and Sublime (the default) editors and it’s built to be great for collaboration (if you want).

    It also has

    • syntax highlighting for a ton of languages
    • Mermaid.js support
    • LaTeX support
    • easy drag and drop image uploads
    • a solid mobile interface (for a webapp in your browser, at least)
    • built in revision history
    • support for other diagram tools, like graphviz, flowchart.js
    • a bunch of other little Markdown enhancements that make using it feel oddly intuitive

    And best of all, they have a Hedgehog for the icon! (I may be biased.)



  • Giphy has a documented API that you could use. There have been bulk downloaders, but I didn’t see any that had recent activity. However you still might be able to use one to model your own script after, like https://github.com/jcpsimmons/giphy-stacks

    There were downloaders for Gfycat - gallery-dl supported it at one point - but it’s down now. However you might be able to find collections that other people downloaded and are now hosting. You could also use the Internet Archive - they have tools and APIs documented

    There’s a Tenor mass downloader that uses the Tenor API and an API key that you provide.

    Imgur has GIFs is supported by gallery-dl, so that’s an option.

    Also, read over https://github.com/simon987/awesome-datahoarding - there may be something useful for you there.

    In terms of hosting, it would depend on my user base and if I want users to be able to upload GIFs, too. If it was just my close friends, then Immich would probably be fine, but if we had people I didn’t know directly using it, I’d want a more refined solution.

    There’s Gifable, which is pretty focused, but looks like it has a pretty small following. I haven’t used it myself to see how suitable it is. If you self-host it (or something else that uses S3), note that you can use MinIO or LocalStack for the S3 container rather than using AWS directly. I’m using MinIO as part of my stack now, though for a completely different app.

    MediaCMS is another option. Less focused on GIFs but more actively developed, and intended to be used for this sort of purpose.