• 0 Posts
  • 91 Comments
Joined 1 year ago
cake
Cake day: July 9th, 2023

help-circle







  • I didn’t see a notification for your reply!

    I think of it this way — at some point it surprised me that Microsoft doesn’t claim ownership in some way to the output of Microsoft Word. I think if “word processing” didn’t exist until this point in history there’s no way you’d be able to just write down whatever you want, what if you copied the works of recently-deceased beloved poet Maya Angelou? Think of the estate? I heard people were writing down the lyrics of Taylor Swift’s latest album and printing off hundreds of copies and sharing it with people at her concerts. Someone even tried to sell an entire word-for-word copy of Harry and Megan’s last best seller on Amazon that they claimed they “created” since they retyped it themselves until the publisher shut it down.

    Obviously all of those things (except my speculation about them claiming any ownership of the output, but look at OpenAI and their tool) don’t happen, but also I think people can write down their favorite poems if they want or print out lyrics because they want to or sit around typing up fan fiction with copyrighted characters all day long, and then there are rules about what they can sell with that obviously derivative content.

    If someone spends forever generating AI Vegetas because Vegeta is super cool or they want to see Vegeta in a bowl of soup or whatever, that’s great. They probably can’t sell that stuff because, y’know, it’s pretty clearly something already existing. But if they spend a lot of time creating new novel stuff, I think there’s a view that (for the end user) the underlying technology has never been their concern. That’s kind of how I see it, but I can understand how others might see it differently.




  • If you make a byte-for-byte copy of something why would you think copyright would not apply? If you listened to the dialogue of a Marvel movie, wrote it down line for line and so happened that the stage directions you wrote were identical to those in the movie, congrats, you’ve worked your way into a direct copy of something that’s under copyright. If you draw three circles by hand in exactly the right way, you might get a Mouse coming after you. If you digitally render those circles in Photoshop, same idea[/concept, yes I know one is a trademark issue].










  • LLMs are conversation engines (hopefully that’s not controversial).

    Imagine if Google was a conversation engine instead of a search engine. You could enter your query and it would essentially describe, in conversation to you, the first search result. It would basically be like searching Google and using the “I’m feeling lucky” button all the time.

    Google, even in its best days, would be a horrible search engine by the “I’m feeling lucky” standard, assuming you wanted an accurate result and accurate means “the system understood me and provided real information useful to me”. Google instead return(ed)s(?) millions or billions of results in response to your query, and we’ve become accustomed to finding what we want within the first 10 results back or, we tweak the search.

    I don’t know if LLMs are really less accurate than a search engine from that standpoint. They “know” many things, but a lot of it needs to be verified. It might not be right on the first or 2nd pass. It might require tweaking your parameters to get better output. It has billions of parameters but regresses to some common mean.

    If an LLM returned results back like a search engine instead of a conversation engine, I guess I mean it might return billions of results and probably most of them would be nonsense (but generally easily human-detectable) and you’d probably still get what you want within the first 10 results, or you’d tweak your parameters.

    (Accordingly I don’t really see LLMs saving all that much practical time either since they can process data differently and parse requests differently but the need to verify their output means that this method still results in a lot of back and forth that we would have had before. It’s just different.)

    (BTW this is exactly how Stable Diffusion and Midjourney work if you think of them as searching the latent space of the model and the prompt as the search query.)

    edit: oh look, a troll appeared and immediately disappeared. nothing of value was lost.