• humanspiral@lemmy.ca
    link
    fedilink
    arrow-up
    2
    ·
    5 hours ago

    There are bigger forces against AI buildout than Democrats. Plenty of demand for datacenters in SEA region (90% of tokens consumed by China) anyway. Instead, OpenAI, and ANT valuations show them having huge difficulty being able to follow through on nearly $2T in spending commitments for those datacenters, and MSFT announcing they are cancelling projects and not renewing Oracle rental deal shows that the Nvidia dependent ecosystem does not have desperate “end user facing” demand.

    wrote 3 articles this week on c/technology@lemmy.ml

  • Phoenixz@lemmy.ca
    link
    fedilink
    arrow-up
    18
    ·
    12 hours ago

    Of course all tech CEOs, investors, and banks will go blame the Democrats for this

    Its easier than taking responsibility for their own fucked up actions, and the results that they themselves caused

  • RedstoneValley@sh.itjust.works
    link
    fedilink
    arrow-up
    45
    ·
    18 hours ago

    It’s funny how quickly the narrative shifts to “the Democrats did it”, when there are some very good other reasons for the stocks to fall, and nobody seems to notice. The circular investment AI business model was not sustainable before the election and the stocks started to fall before, and this became apparent to all with the anthropic IPO filing

    • schema@lemmy.world
      link
      fedilink
      arrow-up
      9
      ·
      12 hours ago

      “We can’t break the laws and do what we want to build datacenters we couldn’t afford to build anyways, because our debt was about to be below invetment grade, since AI doesn’t even come close to the returns we promised. It’s the Democrats fault!!!”

    • ManixT@lemmy.world
      link
      fedilink
      arrow-up
      19
      ·
      16 hours ago

      Also watch now how magically republicans start caring about the deficit and that consumes the news cycles.

    • empireOfLove2@lemmy.dbzer0.com
      link
      fedilink
      English
      arrow-up
      36
      arrow-down
      1
      ·
      21 hours ago

      Nope. The RAM they are making now is not built for consumers and cannot be used in any consumer hardware. IT will take 2-3 years for them to correct course at least.

      • wheezy@lemmy.ml
        link
        fedilink
        English
        arrow-up
        30
        arrow-down
        2
        ·
        21 hours ago

        It might be cheaper to buy the server CPUs and motherboards. But, capitalism doesn’t really do that. They’ll throw it all in a landfill before they would resell it. There is no profit in repacking it when it would just flood the market with pennies on the dollar RAM. It’s so fucking infuriating that we live in a world where throwing shit in a landfill is the solution to overproduction. Instead of people just having useful shit.

        • h0tbeef@lemmy.zip
          link
          fedilink
          arrow-up
          5
          ·
          13 hours ago

          The HBM (slop memory) doesn’t work in a real computer.

          They’ll have to just get rid of it because it’s literally useless for anything other than slop generation.

          So regardless of what our dumbass market dictates, that HBM is useless, full stop.

          • wheezy@lemmy.ml
            link
            fedilink
            English
            arrow-up
            3
            ·
            edit-2
            11 hours ago

            True. I think that’s about a quarter of the dram produced. Though it could be used in public sector research and not just thrown away.

            Astronomy and Biology research benefit from AI image analysis models. Because it’s not just generating slop text. It’s used to analyze image data to find potential points of interest. Reducing an impossible number of raw data points (stars, cells, etc) down to a manageable number that can be analyzed and confirmed by an actual human.

            This isn’t the “AI” people talk about when they say “AI”. It was just called “machine learning with computer vision models” before all the AI marketing crap.

            • h0tbeef@lemmy.zip
              link
              fedilink
              arrow-up
              1
              ·
              6 hours ago

              I don’t know enough about the mechanical workings of things like radiological image recognition to know whether or not HBM would be useful for them, but hopefully you’re right.

              I believe HBM makes up more than a quarter of current production, with all 3 major memory manufacturers prioritizing HBM.

          • Blaster M@lemmy.world
            link
            fedilink
            English
            arrow-up
            3
            arrow-down
            1
            ·
            11 hours ago

            HBM is far from useless - it’s a really good, compact memory architecture with tons of bandwidth. Find a way to adapt it to DRAM modules and we’re gold.

            Worst case we get GPUs that can play AAAA games at 8k 120fps

            • h0tbeef@lemmy.zip
              link
              fedilink
              arrow-up
              2
              ·
              6 hours ago

              The GPUs designed for inference (slop) are not the same kind of GPUs that run video games either.

              Here’s an article explaining why HBM is really difficult to integrate into a real computer, and why doing so would eliminate nearly all of the advantages of using HBM (while failing to eliminate the things that make HBM much more expensive to both produce and power than regular DRAM).

              Your proposed “worst case” scenario would be far better than reality’s “best case” scenario.

              (The real best case scenario would have been to not build all this worthless bullshit in the first place, but its too late for that).

        • CarrotsHaveEars@lemmy.ml
          link
          fedilink
          arrow-up
          8
          ·
          20 hours ago

          I remember about ten years ago, Intel Xeon family released a product line which included Xeon E3-1230 v2 or something similar to that. It is a server grade CPU but with a consumer grade price tag. Some suspected Intel did that to fight off the same tier of AMD CPU.

          So capitalists do do that, as long as there is competition. It was a good time.

          Now they fucking banded together, a fucking trade union for corporations which fights workers and consumers.

          • h0tbeef@lemmy.zip
            link
            fedilink
            arrow-up
            2
            ·
            13 hours ago

            They’re also really gonna want to recoup some of the money they’re about to lose once they realize that their HBM customers don’t actually have the money to pay for all the memory they ordered, and never will.

        • adhdsergio@lemmy.world
          link
          fedilink
          arrow-up
          1
          ·
          18 hours ago

          Well yeah if we all get their extras we won’t want to shell out thousands for whatever they’re gonna be making next

    • frazw@lemmy.world
      link
      fedilink
      English
      arrow-up
      3
      ·
      20 hours ago

      Stock market demands the line goes up. There are two ways to do that. Reduce headcount to reduce your outgoings or increase profitability which means either selling more or raising prices.

      I am not expecting lower prices.

      • h0tbeef@lemmy.zip
        link
        fedilink
        arrow-up
        3
        ·
        13 hours ago

        Supply and Demand have a role to play here

        The supply is currently thin because they’ve converted most of their fabrication facilities to make HBM (slop memory) that doesn’t work in real computers. This has significantly diminished the supply of useful memory in circulation.

        Once they can’t sell the HBM anymore they’ll be forced to convert back to making normal RAM, which will increase the supply and relieve demand (and prices).

        Their other available choice would be to die and let one of the new memory upstarts take their place.

        I don’t think they’ll choose to die, but we’ll see soon.

    • stoy@lemmy.zip
      link
      fedilink
      English
      arrow-up
      3
      ·
      21 hours ago

      And I’ll happily oblige them, I need more ram and would also buy some sticks for use in future machines.

      I bought my current computer back in 2021, and it is time to start looking for components for a new one, I usually run my machines for 7-8 years with incremental upgrades before they need to be replaced, so I have some more time…

    • wheezy@lemmy.ml
      link
      fedilink
      English
      arrow-up
      7
      arrow-down
      4
      ·
      20 hours ago

      The more the establishment fails and the more they fight DSA the better. We are in the “blue hopium” phase right now. Trying to let the libs get there on their own. I am confident the establishment Democrats will continue to betray the American people and give no accountability for the crimes of the Trump administration. Which, is not necessarily bad as long as the socialist we elect stand strong against them and don’t make compromises to obtain minor improvements. It will taint their reputation and the movement if they do.

      The Democrats are already talking about “moving forward” and saying the American people don’t want retribution. Not even the most “vote blue no matter who” liberals I know agree with that. They want to see fucking heads roll. So, the Democrats are already doing it. They are already showing how weak they are in their strongest hour.

      If DSA politicians are blocked it just shows the failures of the existing Democrats more and builds momentum for 2028.

      If DSA politicians betray the movement; we’re gonna have to do the work to keep the movement going outside of electoral politics. Because if we don’t maintain this momentum we could lose to a Tucker Carlson style fascist that is actually competent in the white house. And the electoral portion of building class consciousness will be impossible by the time they take power. The next two years will be about preparing for both scenarios and directing our messaging based on the changing conditions.