Skip Navigation

InitialsDiceBearhttps://github.com/dicebear/dicebearhttps://creativecommons.org/publicdomain/zero/1.0/„Initials” (https://github.com/dicebear/dicebear) by „DiceBear”, licensed under „CC0 1.0” (https://creativecommons.org/publicdomain/zero/1.0/)WA
Posts
2
Comments
512
Joined
2 yr. ago

  • Open sourcing the training method without open sourcing the training data is essentially like making only part of your full source open to the public.

    Even going as far as making your training method source available, and a pre-trained kernel available (like what Mistral does) is essentially the same as what a lot of open source-adjacent companies provide.

    A pre-trained neural kernel isn't any different effectively than a pre-compiled binary library (like a dll). So what these companies are providing is closed-source binaries alongside the compilation instructions for them. But without the data that trained the kernel it can hardly be called "open source" as the actual "source" of the logic behind the kernel (the training data) is still closed to the public.

    You can fine-tune and re-train and re-quantize the models all you want but you're not really manipulating the "source" if all you have is the gptq or safetensors or some other pre-trained set of weights.

  • Doesn't seem like it outside this:

    Developers of general purpose AI models — from European startups to OpenAI and Google — will have to provide a detailed summary of the text, pictures, video and other data on the internet that is used to train the systems as well as follow EU copyright law.

    Which makes me think that it'll be used to require models to truly open their "source"

    The FOSS community really needs to come up with a better definition and licensing model for LLMs and other neural networks, though. I've seen multiple times where people refer to freely provided pre-trained models as "open source"

    AIs aren't truly open source unless their training code and the training data is fully provided. Anything else is at most semi-obfuscated and definitely not "open"

  • Yea it's the same for us, the complaints from people when they see a kid in public on a tablet are weird to me cause I know as kids we always had stuff like toys we brought into restaurants (or we went to restaurants with like coloring maps and stuff).

    Parents have been desperately trying to find things to occupy kids while they're in public so they don't disturb the people around them for years and now that smart phones/ipads are universal it seems like there's finally something that will just keep the kids quiet for awhile without a lot of effort.

    I think it's important to pay attention how much you/your kids are spending on "screen time" but it feels really disingenuous to say stuff like the current generation is cooked because of ipads.

  • I'm usually more supportive of labor protests but it's hard to see the strike aimed at blocking this legislation as anything but selfish.

    Are there details about them demanding labor concessions in exchange for this bill or is it simply "no more new doctors"?

  • crawls out of gnu logo-shaped hole

    I use icecat and it's pretty nice, main issue is you have to manually approve all JavaScript scripts and cross-site requests manually, so it makes visiting almost any website for the first time a pain in the ass.

    Luckily most stuff is concentrated to a few sites and I use private frontends liberally so it's not a huge pain.

    Slinks back into free software hole

  • I switched to guix and haven't looked back.

    Mostly because:

    1. I like the idea of functional package managers
    2. I like guix's dedication to making every package buildable from source (thus the no non-libre code rule)
    3. I like the expressiveness of scheme vs Nix's package description language

    Guix is the smoothest time I've ever built packages for a distro before (well outside arch). Which is good because there's a lot of out of date and unadded packages for potential.

  • This is something Japanese train companies figured out awhile ago for train engineers. Because driving locomotives can be really repetitive, they train engineers to do hand signals and call out actions out loud even when they're alone in the car in order to help keep the brain active and focused.