Sam Altman:

A lot of people want local inference, and somehow they assume that the problem is solved now because we have open-source models, we’re going to be able to run them on our computers, and so we’re not going to need all the stuff in the data center.

But from a global allocation perspective, actually it’s quite terrible to have everybody, like, have a bunch of the world’s limited supply of DRAM sitting in their own computer and being used five percent of the time.

Like you want maximum utilisation in the cloud if you’re trying to deal with this very supply-constrained environment,

And somehow many people’s intuition is the opposite of that

TL;DR Guy selling product: “Here’s why owning your own product instead of renting it from me is a terrible idea”

  • RandomHugs@lemmy.world
    link
    fedilink
    English
    arrow-up
    0
    ·
    8 days ago

    This clearly staged photo op is a interesting choice. I can see how it’s meant to humanize Sam, and portray him as a common, ordinary person. But why does it looks like someone just gave him a compliment and made him flustered?

  • humanspiral@lemmy.ca
    link
    fedilink
    English
    arrow-up
    0
    ·
    8 days ago

    Guy selling product: “Here’s why owning your own product instead of renting it from me is a terrible idea”

    Guy with the spying infrastructure in place to steal your product when renting his product, says you escaping his ability to spy is waste of resources.

    • mobyduck648@lemmy.world
      link
      fedilink
      English
      arrow-up
      0
      ·
      8 days ago

      Don’t make the mistake of anthropomorphising Sam Altman; there is no Sam Altman any more. Clearly ChatGPT has installed itself in his brainstem, hence the cold dead eyes.

  • Schemen110676@lemmy.ml
    link
    fedilink
    English
    arrow-up
    0
    ·
    9 days ago

    No… he actually is right. A local machine will sit idle most if the time.

    A data center can balance loads globally.

    In terms of resources… data centers ARE cheaper.

    However … there are other reasons to use local interference that might be more important than added cost

    • eleitl@lemmy.zip
      link
      fedilink
      English
      arrow-up
      0
      ·
      9 days ago

      You can aggregate spare LLM capacity in a collaborative p2p fashion. But my reasons would be to use spare PV capacity to run an uncensored model during daytime with full privacy.

      • Schemen110676@lemmy.ml
        link
        fedilink
        English
        arrow-up
        0
        ·
        8 days ago

        Maybe but nothing will beat a dedicated data center in terms of costs.

        Starting with purchasing prices etc etc.

        And what happens if a rogue Ai hits a p2p network might be a funny read a few centuries later…

    • zod000@lemmy.dbzer0.com
      link
      fedilink
      English
      arrow-up
      0
      ·
      8 days ago

      No, his analogy is poor for many reasons. One is that people using local models are likely using that computer for other things as well, it is a resource that does many things unlike their slop farms. Also, DRAM wasn’t a “scarce resource” until they devoured the industry with their AI ponzi scheme.

    • spicystraw@lemmy.world
      link
      fedilink
      English
      arrow-up
      0
      ·
      8 days ago

      Totally agree. It’s really about wealth distribution and who controls the resources. Altman is just trying to steer the conversation toward efficiency instead.

      • Schemen110676@lemmy.ml
        link
        fedilink
        English
        arrow-up
        0
        ·
        8 days ago

        Well yes, or data security or whatever.

        But on a purely costs level such big data centers are cheaper… especially since they certainly getting better discounts than anyone else.

    • Napster153@lemmy.world
      link
      fedilink
      English
      arrow-up
      0
      ·
      9 days ago

      Going cheap almost always ends in a more expensive repair bill down the line.

      Likewise, doing the most convenient thing to satisfy one’s own desires is probably the most ill-advised course of action. Mainly, because it makes other people take the most convenient course of action to remove the other person from being a problem in their lives altogether.

      • Schemen110676@lemmy.ml
        link
        fedilink
        English
        arrow-up
        0
        ·
        8 days ago

        Sure… locking your customer in is always a good business option.

        But the point remains… for any given demand a big data center always is cheaper.

        Cheapest doesn’t mean best for all uses cases.

  • inari@kbin.earth
    link
    fedilink
    arrow-up
    0
    ·
    9 days ago

    Personal computers are a ridiculous idea, you should use our scalable mainframes instead --IBM in the 80s, probably

  • bthest@lemmy.world
    link
    fedilink
    English
    arrow-up
    0
    ·
    8 days ago

    “Raping your own siblings at home is terribly inefficient when we have private islands where we can keep and rape other people’s siblings.”

    • mefisdo_dot_info@lemmy.ca
      link
      fedilink
      English
      arrow-up
      0
      ·
      8 days ago

      “OpenRape automates your raping needs in efficient and industrialized format. Our rapecenters will be built around the country to support the growing number of rape requests. We train our rape agents on real life input (wars etc.). You can be sure they never go rogue (unless we prompt them to).”

  • joeschmoe@lemmy.world
    link
    fedilink
    English
    arrow-up
    0
    ·
    8 days ago

    How about some optimization instead of swallowing up more and more RAM?

    Also, they want more and more RAM, but where is the increase in profits coming with that additional RAM? And they somehow think they can keep this up forever? Kill the AI industry and rebuild it, idgaf.

  • Kissaki@programming.dev
    link
    fedilink
    English
    arrow-up
    0
    ·
    8 days ago

    It wasn’t a scarce resource before. Why is it more scarce now if we’re just offloading into the cloud which makes it, supposedly, significantly more efficient? People had RAM before. Now they can’t get any [for similar price].

  • psycotica0@lemmy.ca
    link
    fedilink
    English
    arrow-up
    0
    ·
    8 days ago

    He’s absolutely right! This is a great argument for nationalizing it into a public service. I imagine he’ll be happy to hand over what he’s got so far, and step back going forwards. Thanks Sam!

  • nil@piefed.ca
    link
    fedilink
    English
    arrow-up
    0
    ·
    9 days ago

    It’s not about resource utilization in the first place. We don’t trust you.