• 0 Posts
  • 60 Comments
Joined 2 years ago
cake
Cake day: July 1st, 2023

help-circle
  • I’m still young at 25, but I can see the hallmarks of aging. I’ve moved to a new state for 5 years now and when I visited my old home it felt half foreign, half familiar. I’m the youngest so my mom’s age is starting to show.

    Things I consider recent are now described as “years ago”. I’m seeing things evolve through life. Things that felt like they had a beginning, middle, end now are starting up again. Almost like a ride that’s resetting for the next ones in line.


  • I’m surprised Harkon didn’t try to take Blackreach for himself. They’re immune to Falmer poison, they can use their full power without sunlight, they’re smarter and have access to invisibility, illusions, and charms.

    In return he could have a country sized cavern, with at least two points of contact to Skyrim one either end, all the dwemer tech to hoard and use with their ancient knowledge. He could subjugate the falmer to be blood livestock, they’re already crafted to be good slaves by the dwemer.

    He could slowly build up a vampire/thrall/dwemer machine army right under Skyrim then wage war overnight. Could even use dwemer machines by day to maintain territory until you complete this master prophecy you have.



  • I stayed up to date on ai and machine learning, including language models. I remember hearing that one learned math from language and wondering where things will go. I watched ai safety videos before they felt relevant. Then I heard Openai, which had a good rep at the time, is releasing their new model online, called ChatGPT. Having played with DungeonAI and NovelAI before I was gonna fiddle with this as well.

    Then headlines broke, it became a phenomenon. Even then I figured this would be this week’s Thing before getting bored, as was common with these ai.

    Down the line I remembered hearing ChatGPT on a gas station ad for some travel app. That was when I realized this is permanent. People who aren’t even online are likely hearing about this. Suddenly my niche hobby and hopeful dreams of the future became an actual enshittified crisis.

    I don’t think I need to explain how everyone using language models now is just god awful for everyone. And the attention hasn’t gotten us closer to answering long standing questions of ethics, economic change, what is intelligence or consciousness. We’ve just got a bunch of the lowest common denominator shouting their answers now.









  • There’s a fascinating idea that Judas was the one who committed the ultimate sacrifice. That god chose him to be his human incarnate, to truly experience humanity and guilt by committing an ultimate betrayal and becoming the villain of biblical history. All allowing him to finally understand and forgive humanity’s sin, by committing one himself. It follows that this is supposedly maddening knowledge as it breaks the illusion of Christ’s sacrifice.

    I’m definitely butchering and ad-libbing the original idea, but I think this makes for a grander story than the traditional “birth myself to sacrifice myself to myself to forgive everyone else” interpretation.



  • So a tricky thing with LGBT words is they seem to vary around specificity.

    Like people can say gay to mean gay men, gay people, or any kind of queer.

    Similarly transgender can mean a gender identity not matching your biological sex, which includes genderfluid; or it can mean specifically a binary trans man/woman.

    These aren’t different valid opinions about the label. They are literally just different definitions for the words that depend on context. The words mean all of those things but usually only one in each context.





  • Wirlocke@lemmy.blahaj.zonetoTechnology@lemmy.world*Permanently Deleted*
    link
    fedilink
    English
    arrow-up
    8
    arrow-down
    2
    ·
    1 year ago

    What resonated with me is people calling LLMs and Stable Diffusion “copyright laundering”. If copyright ever swung in AI’s favor it would be super easy to train an AI on stuff you want to steal, add in some generic training, and now you have a “new” piece of art.

    LLMs and Stable Diffusion are just compression algorithms for abstract patterns, only one level above data.