• 0 Posts
  • 154 Comments
Joined 3 years ago
cake
Cake day: August 27th, 2023

help-circle

  • mindbleach@sh.itjust.workstoMemes@lemmy.mlHappy Killdozer Day
    link
    fedilink
    arrow-up
    9
    arrow-down
    2
    ·
    3 days ago

    Marvin Heemeyer was a boomer terrorist asshole.

    Dipshit got a great land deal, illegally lived out of a shop with no plumbing, agreed to sell his shit-filled ditch at immense profit, reneged on that sale, and then lost his goddamn mind and embarked on a suicidal rampage through a mixed residential area with earth-moving equipment covered in mounted weaponry. Stop even pretending to idolize this bastard.







  • mindbleach@sh.itjust.workstoHacker News@lemmy.bestiver.seHating AI Is Good
    link
    fedilink
    English
    arrow-up
    1
    arrow-down
    2
    ·
    18 days ago

    Kneejerk overcorrection to performative ingroup identity. The tech is fine. When the bubble bursts, we’ll still have the chatbot that codes, and the denoiser that renders live action. Muttering about copyright, like anyone should respect copyright, will either quietly fade, or produce vegan models from bespoke and licensed data. Those vegan models will not change anything for the bitter core who shouted loudest about it, because they’re only bolstering a reactionary conclusion. Same reason they don’t care how local models address power use. Same reason Chrome pushing a four-gig update is like burning a rainforest, but Youtube’s daily bandwidth will not be examined.

    Absolutely excoriate the schmuck whose ‘AI clouds reality’ book is a brainless example of its own premise. But a Nobel laureate bouncing ideas off a hallucinating robot is already unremarkable. Cory fucking Doctorow uses one to check his posts. It’s fine. That’s what it’s for. Don’t treat it as an oracle, and don’t submit its output like you wrote it. Not even when people cannot tell, despite the highest quality standards and intense scrutiny. Granta’s uncertainty acknowledges how spicy autocomplete can already leave no trace beyond a vague suspicion. You want that capability kept out of artistic expression? Go off. But do not pretend it will simply go away.

    Imagine describing what LLMs do, a decade ago. Not the Sam Altman cocaine fantasy. Their actual current functionality, warts and all. Who would be against it, on principle? Even if you go full alarmist - ‘I came back to 2016 because this is what we have to stop!’ - you’re unlikely to persuade anyone to join the Butlerian jihad, over what the chatbot is and does. They could rattle off some obvious uses, including a few we still haven’t done. It seems fine.

    People nowadays are in antithesis because we’ve had this shoved at us from every angle. It’s been used to justify mass layoffs, constant surveillance, and computing restrictions. Every platform overflows with unfiltered garbage because the zaibatsus that own them are invested in the cocaine fantasy. You hate these companies and confuse this for the product.

    When that all collapses - we can finally talk about the tech, in itself. About having computers at least try to do what you ask for, in plain English. About the storytelling potential of describing images into existence. About people creating the tools they need, without years of comp sci. Are there problems within all that? Of fucking course there are. You live in science fiction now. It’s gonna get less stupid, any day now, but it’s gonna stay weird.





  • entitled irrational hyper privileged

    bad faith

    Yeah fuck me for trying to engage with what you think, despite continued essentialist insults. Unpleasant time vampire.

    Again: even if you had to run this a bunch to get these results - you can in fact run it a bunch, and it can in fact get these results. Even if that’s only sometimes, that’s not never. Even mere translation is useful, like moving from a slow language to a fast language, as for the Minecraft example.

    You are the only one here with a dogmatic absolute opinion on this, and it’ not even one you formed yourself.