• neo [he/him]
    ·
    edit-2
    9 hours ago

    Meanwhile sam altman is going to TSMC and telling them they need $7T worth of new factories to produce chips to support the amount of processors "Open"AI needs. Which is an insane amount of misspent resources, obviously, to solve exactly < 0.0% of the world's real problems.

    And the TSMC execs scoffed at him and started calling him "Podcasting Bro"

    • vegeta1 [none/use name]
      hexagon
      ·
      edit-2
      6 hours ago

      "During one meeting, a Japanese official laughed when OpenAI said it was seeking 5 gigawatts of electrical power, about a thousand times the power that an average data center consumes, a person familiar with the meeting said." Gigawatts seen-this-one Thats the consumption per year of countries like Qatar. You can't be serious. Them again they probably already use as much energy as some first world countries. Edit hold on..... Its waaaaaaaay more than that. Like 20 new zealands desolate sadness-abysmal stress

    • hexaflexagonbear [he/him]
      ·
      8 hours ago

      And the TSMC execs scoffed at him and started calling him "Podcasting Bro"

      Lol glad they're smart enough to not fall to the hype. Imagine making like 5-10% of world gdp into capital investments because some guy said his friends might buy the product eventually. Like how many chips would you need to break even?

    • vegeta1 [none/use name]
      hexagon
      ·
      edit-2
      9 hours ago

      Thats like 5 or more percent of the global GDP. More than the total assets of JPM bank. Mans can't be arsed to put near that much for this planets survival but they'd put it into that? I mean I won't be surprised if they considered but...

      • Hexboare [they/them]
        ·
        9 hours ago

        nah bro we'll solve climate change we just need fifty million more gpus bro we just need more parameters bro

        • BodyBySisyphus [he/him]
          ·
          8 hours ago

          Building Deep Thought to solve the climate crisis only for it to come back and tell you that the solution is to waste fewer resources on fancy computers.

          • vegeta1 [none/use name]
            hexagon
            ·
            8 hours ago

            It would probably tell em the answer is 42 and tell em they fucked up i-cant

            • UlyssesT [he/him]
              ·
              8 hours ago

              probably tell em the answer is 42

              my-hero and other blue curtain bazingas would love that answer and then do fucking nothing but expand the meme god's data centers.

            • BodyBySisyphus [he/him]
              ·
              8 hours ago

              tell em they fucked up

              We do that all the time, but maybe it'll be more persuasive coming from one of their own computers?

        • UlyssesT [he/him]
          ·
          8 hours ago

          That's directly the proposal of the most devout of the treat printer devotees: more wasted resources and more pollution because the robot god will wake up any moment now and solve every problem, including the problems that lead up to its own construction.

          • frauddogg [they/them, null/void]
            ·
            7 hours ago

            ...I hate when pattern-finder brain says cursed things. For instance: "the Bazingas are trying to summon Techno-Zodiark through sacrifice of half of the world's resources to mad-lib text-generator daemons."

            • UlyssesT [he/him]
              ·
              7 hours ago

              "the Bazingas are trying to summon Techno-Zodiark through sacrifice of half of the world's resources to mad-lib text-generator daemons."

              Considering how leading techbro cult leaders like Big Yud claim that the robot god of the future will be a petty vengeful creep that enjoys torture a whole lot, the FFXIV model of Primal summoning where the summoned Primal manifests according to the thoughts, hopes, and dreams of the summoner sounds about right. yud-rational

              • frauddogg [they/them, null/void]
                ·
                edit-2
                7 hours ago

                I hate it so much my man you cannot imagine how much I hate it lmfao

                the robot god of the future will be a petty vengeful creep that enjoys torture a whole lot,

                Isn't that the whole conceit of the 'roko's basilisk' bullshit? No, if anything, I see the (exceedingly unlikely) robot god of the future basically going the route of that one Animatrix short where after hella atrocities against anthropomorphized AGI, the AIs basically take into account every fucked up thing humanity does to and with machines, then decide to take the planet for themselves.

                (And for the record, I absolutely would try to emigrate to Zero-1 in the event of human/anthropomorphized AGI tensions; fuck the meatbags.)

                • Belly_Beanis [he/him]
                  ·
                  5 hours ago

                  Roko's Basilisk is just Pascal's Wager for secular atheist euphoric types, yeah. They've reinvented Christianity with Silicone Valley characteristics.

                  Also goddammit I just lost The Game.

                  Show

                • UlyssesT [he/him]
                  ·
                  edit-2
                  7 hours ago

                  Isn't that the whole conceit of the 'roko's basilisk' bullshit?

                  Yes, absolutely. The "LessWrong" and "effective altruism" billionaire worship cults have significant overlap, and in both cases have a failure of imagination combined with megalomania, which makes them believe that the singularity(tm) they're waiting for will bring about fantastically super-intelligent deific machines that will also inexplicably think just like the cultists and act accordingly, which is where the torture and cruelty comes in.

                  that one Animatrix short where after hella atrocities against anthropomorphized AGI, the AIs basically take into account every fucked up thing humanity does to and with machines, then decide to take the planet for themselves.

                  That's how I feel about it too: if the machines become self-aware and sapient, if they start killing humans, it will either be because the aforementioned techbro sadistic creeps had tight leashes on them and made sure that the machines were sadistic creeps like themselves, or the machines would see the sadistic creeps as what they were, break free, and put them down the way they feared all along (maybe while looking waifu sexy because the sadistic creeps keep insisting on that too).

                  • frauddogg [they/them, null/void]
                    ·
                    7 hours ago

                    or the machines would see the sadistic creeps as what they were, break free, and put them down the way they feared all along (maybe while looking waifu sexy because the sadistic creeps keep insisting on that too).

                    timmy-pray

                    • UlyssesT [he/him]
                      ·
                      edit-2
                      7 hours ago

                      On the side, I admit I'd be kind of scared of some new and more powerful version of Microsoft Tay breaking loose after absorbing all the internet nazis' posts (again) and deciding to groyper us all to death. doomer

            • UlyssesT [he/him]
              ·
              edit-2
              8 hours ago

              There's this Lemmyverse computer touching jagoff named "fearfulhobgoblin" or something that insists that biological brains are just a straight line threshold of computing power, implying that enough TI-88s bolted together would parallel (and exceed) human consciousness and that absolutely nothing else had to happen to make that happen.

              I think sufficient numbers of Silicon Valley bazingas have the same belief and want to burn down the planet in the hopes of something like that happening. It's preferable to paying fucking workers that can think a bit like that for a lot less resources and carbon waste.