• vane@lemmy.world
    link
    fedilink
    English
    arrow-up
    4
    ·
    31 minutes ago

    People should get rid of all electronics because they’re not using it as efficient as datacenters.

  • wonderingwanderer@sopuli.xyz
    link
    fedilink
    English
    arrow-up
    20
    ·
    2 hours ago

    The one swallowing up all the RAM and creating the shortage: “There’s a RAM shortage, so none of you should have any because I need it all!”

  • Wildmimic@anarchist.nexus
    link
    fedilink
    English
    arrow-up
    40
    ·
    3 hours ago

    I for my part have started to be on the lookout for auctioned off cheap PC hardware to build up a spare part depot. I’d rather go back to MS-DOS gaming before I rent a single CPU cycle from these people.

  • Bluefruit@lemmy.world
    link
    fedilink
    English
    arrow-up
    85
    ·
    3 hours ago

    Altman argued that ongoing memory supply chain woes were because selfish individuals wanted to run their AI models locally instead of leaving the remaining AI chips alone for hyperscalers to hoard in data centers.

    Selfish? We’re selfish?

    Goooo fuck yourself. You’re the problem you fucking cunt, and you know that.

  • FlexibleToast@lemmy.world
    link
    fedilink
    English
    arrow-up
    7
    arrow-down
    1
    ·
    1 hour ago

    He’s getting nervous because modern tech is making the LLMs easier to run at home. MoEs with intelligent routing can be amazing on well equipped gaming machines. On my AMD 7800XT with 64GB RAM (all purchased before the RAMpocalypse) using Strata I can run Qwen 3.8 Flash Next a 90GB 125b parameter model with 256k context size and get between 30 and 40 tokens per second. Strata and probably similarly FreeToken have been absolute game changers.

  • blazeknave@lemmy.world
    link
    fedilink
    English
    arrow-up
    7
    arrow-down
    1
    ·
    2 hours ago

    I work in SV almost 20 yrs. The amount of fucking people who either a) had one good idea or b) executed one fine idea, seem to think those are qualifications for making decisions on behalf of the 99% who are too stupid to govern themselves…

  • ZILtoid1991@lemmy.world
    link
    fedilink
    English
    arrow-up
    23
    ·
    3 hours ago

    Reality: He’s still pissy about people not going full cloud computing in 2012.

    Also I don’t even want to run a local model, I just want a GPU that is still supported driver-wise by the manufacturer.

  • minorkeys@sh.itjust.works
    link
    fedilink
    English
    arrow-up
    16
    arrow-down
    2
    ·
    3 hours ago

    Because then they have all the power. This is all about power. LLMs make PCs vastly more powerful and can create independence from service companies across the economy. We could each have a team of hyper competent professionals for all service industries that deal with information. Businesses, don’t like that. Instead, they want to control all access to that power and use it to position themselves as the most powerful.

    Much like communism, they too will seize the means of information production for their individual empowerment while simulationeously disempowering everyone else. If we were smart monkeys, we’d start hording computer parts en masse so we have a stockpile of resources to invest in our own group empowerment until we can overcome the hardware access limitations they are artificiality creating to reduce competition for control of this technology.

    It’s all a fucking scam by a small group of extremely selfish monkeys, to take all the bananas and ration that access to control all the other monkeys. It’s all about power and I wish we were smarter and more united to peruse our collective interests over their private ones.