• 0 Posts
  • 251 Comments
Joined 4 months ago
cake
Cake day: March 7th, 2026

help-circle



  • Jiral@lemmy.orgtoPeople Twitter@sh.itjust.worksManagers
    link
    fedilink
    arrow-up
    1
    ·
    edit-2
    23 days ago

    Yes I think Strix Halo makes sense when low power use is a requirement. I built a custom fanless Strix Halo system for the fun of it and I guess there aren’t too many out there running Gemma 4 31B Q8 without a single fan, anywhere.

    And for MoE models that need 60-80GB + context it is perfect. Those are decently fast then as well.

    PS: If VRAM is all you care about the maxed out Mac Studio is fascinating. 512GB unified memory for around 10K EUR (pre crazy bubble prices) That should be able to run pretty large MoE models but dense models of that size would probably run glacially.


  • My system runs at 100W TDP though. That is maybe 140W at the power outlet, incl. monitor and everything.

    This is also the dense 27B model at Q8. But yeah, it is not terribly fast. I think the best use case is on MoE models. GPT-OSS-120B runs on it for example and at 50T/s speed is not a n issue anymore either. (I could get it to run even on just 64GB but the new llama.cpp might need a tiny bit more memory which pushed it just across the limit. yeah I know, for seriously using it you’d need the 128GB version)


  • Jiral@lemmy.orgtoPeople Twitter@sh.itjust.worksManagers
    link
    fedilink
    arrow-up
    1
    arrow-down
    1
    ·
    23 days ago

    At Q8 it is around 35-40GB I think + memory for required context.

    I have a Framework desktop. It gets you you around 6t/s. Not suitable for professional use but for personal use I think it is fine. I do prefer Gemma 4 though, but that comes with similar reqirements.


  • I am pretty negative on AI but there is a point there. I tried the open weight local model Gemma 4 31B and while it likely cannot compete with the best Claude has to offer today, it might be on par with Claude from a year ago, at least for certain applications. With a local model the data stays on your system and you are in control of the costs (no sudden price hikes). But local models aren’t for free either they still guzzle compute, merely on your own hardware (or rented hardware)







  • Nuclear protection for Ukraine, do you have any credible source for that claim? Materials and arms from other countries, hardly makes a case against your point that you have no army unless you have nuclear arms.

    Even if Biden said anything. Biden isn’t in power anymore and even with him it was far fetched to believe the US would risk nuclear war with Russia over Ukraine. This whole premise has become entirely absurd under Trump. Yet, the counter attacks against Russia have vastly intensified since Biden is out of office.

    Your racist classifications are not necessarily as central to foreign policy as you might believe.


  • I see, I did indeed misunderstand you.

    So you are arguing really for proper nuclear proliferation with Norway getting its own independent nuclear capabilities like France. The only small country having nuclear capabilities is Israel, and they surely pay a hefty price for it and are extremely dependend on the support of another major power nonetheless, probably more than Norway is dependent on France, even after this deal.

    If France renegades on that deal, the worst thing that happens to Norway is being in the situation form before that deal.

    Ukraine has no nuclear warheads and is bombing a nuclear power heavily (defending against its aggression), damaging a lot of its strategically important oil and gas industry btw. A nuclear deterrent is a strong power but it is not the only way to exert hard power, nor does the lack of having nuclear weapons make a country necessarily easy prey.


  • You were not talking about nuclear proliferation in the comment I was responding to, but made the ridiculous claim that Norway turned itself into a vassal of France. Following up towards another poster with a smear against “French people”. That said, I take back my comment towards you in respect of the rules. Maybe you should take back yours too.

    I am not sure how arguing for nuclear proliferation means being a Russian bot. Is there a “not” missing? In that case how is nuclear proliferation when it does not change the number of nuclear war heads, does not spread nuclear war heads to another nation nor does it give any additional government control over nuclear warheads.






  • I was maybe not clear about it, by non-US source I was talking about historical studies and works not done by US Americans. So, no need to tell me that the US did not exist in the 17th century. Studying “whiteness” is a very US American thing, and the easiest way to prevent such bias is to look at work from elsewhere. But let’s take that Michigan/North Carolinian anthropologisat for a second. What you describe there is purely external, concerning colonies, not the homeland.

    From what you describe it does not appear, that even in Italy, people were classifying people as white and non-white amont the domestic, non-immigrant population. Not in the 17th century. Now, in the 19th century of course you have the rise of modern racism but even in those works you quote it sounds like they were not separating local non-immigrant population of Italian speaking regions into populated by white and non-white.