• 0 Posts
  • 141 Comments
Joined 3 years ago
cake
Cake day: July 11th, 2023

help-circle
  • Think more extreme than that, as in “if you’ve ever seen any GPL code whatsoever, it has in some small way impacted your understanding of code and now all code you write forevermore must be GPL” rather than just the codebase of the thing you are re-implementing because that’s the kind of breadth you’d have to apply to the “derivative works” clause in GPL to argue all gen AI code must be GPL because the training data contains some GPL code.


  • Schadrach@lemmy.sdf.orgtoComic Strips@lemmy.world•Prompt [SMBC]
    link
    fedilink
    English
    arrow-up
    1
    arrow-down
    1
    ·
    5 days ago

    This represents such an entirely wrong idea of what’s going on that it’s almost hilarious. The model doesn’t just have an internal database of all the training data that it, like, collages together. Holy fuck, if a model internally contained all of the training data in a form it could simply be retrieved from it would be far more valuable as data compression than anything else it might be or do.

    What’s much closer is that an image generation model has seen millions of pictures, each labeled with metadata about what it is a picture of, each of which causes tiny adjustments to the neural networks weights, because the model is being trained to be able to look at a novel picture and recognize “this picture is of a dog pissing on an apple tree”. Any model that can do that can also be run in reverse - essentially “here is a block of white noise I am saying is supposed to be a really noisy picture of a dog pissing on an apple tree, try to clean up the noise and show me how it looks like a dog pissing on an apple tree”, then run that in a loop until the picture doesn’t change much between iterations. It doesn’t need to have seen something precisely identical to what it’s generating - with enough training data it has an internal model for things like dogs, clowns or what makes something look like anime and the generation just tickles the bits until the picture rates as probably being what’s being asked for.



  • Schadrach@lemmy.sdf.orgtoComic Strips@lemmy.world•Prompt [SMBC]
    link
    fedilink
    English
    arrow-up
    5
    arrow-down
    2
    ·
    8 days ago

    Hell, my interpretation of the “derivative works” clause of GPL stipulates that all llm and genai products should also be gpl.

    …an interpretation broad enough to cover that would also mean that anyone who has ever seen GPL code before would only ever be able to write GPL code ever again. The idea of which I imagine would give Richard Stallman an erection you could see from space,


  • They left out the video game version, who looks like a butch version of the anime one, except she also got super into bonesmashing and it didn’t go quite as intended. Weirdly, the person her face model will have been based on will have had a face that looks shockingly like canon despite the game model only barely resembling her at all.

    Also the Netflix adaptation would have been black. Netflix is noticeably weird about casting black actors to play characters that are canonically gingers more than any other case of casting that doesn’t resemble canon.









  • Schadrach@lemmy.sdf.orgtoComic Strips@lemmy.world•The lookup table
    link
    fedilink
    English
    arrow-up
    5
    arrow-down
    2
    ·
    1 month ago

    …but does the system that is the Chinese Room (the entire system taken as a whole) understand Chinese?

    Any given sufficiently small piece of your brain does not understand English, does no individual neuron in the language center of your brain understanding English mean that you do not understand English? If that understanding is held in the organization and structure of the brain but not in the individual parts, then why can’t the Chinese Room understand Chinese despite no individual component understanding Chinese? That’s a surprisingly difficult question to answer in a coherent fashion without asserting that knowledge in an organic brain is magical in some fashion, and nothing else counts.

    Wait until you consider the concept of philosophical zombies, which is all about pointing out the edges and contours of the hard problem of consciousness. The starting point of that discussion is essentially “Is someone who responds exactly like a regular person would but has no actual subjective experience conscious?” and gets stranger from there.






  • So, I’m sitting in a creepy house playing paranormal investigator with a machine that can listen to the past, but only if I can tell it exactly when, where and who was present. I’m supposed to be investigating a series of deaths, but they’re…weird. I’ve personally listened to three of the deaths happen, none of them with an obvious proximate cause of death, and one of them at the very beginning of the time span the machine can reach. Somehow the deaths are also connected to the weather?, and also trigger agnosia of the victim’s identity, including one man forgetting his lady entirely mid-sentence when she dies while they are together. One person seems potentially at least partly immune to this effect, since she has asked about her daughter after her daughter died, while everyone else has forgotten her daughter ever existed at all aside from the corpse of an unknown person being found.

    Amusingly, there’s one guy wandering the house during the timeline who talks about how he used to live there, but isn’t sure why he showed up other than nostalgia, and is very angry that the house is in the hands it is instead of his own - this is because the first victim was the owner of the house and his brother who invited him and absent knowing he ever had a brother, things fit together…weird.