Skip Navigation

Posts
333
Comments
283
Joined
9 mo. ago

Rekall is a company that provides memory implants of vacations, where a client can take a memory trip to a certain planet and be whoever they desire.

  • This is how we know American oligarchs have crossed the metaphorical event horizon.

    I will note that you'll never find a russian oligarch complaining about buying their parents/relatives a nice house or whatever in 1999 that could have been invested into something else. It is a very human thing to buy your parents something nice when you have a windfall of money (irrespective of who you are and what your ethical framework is like).

    On the other hand, Huang's theorycrafting about investing the 1999 cost of the S-class shows a lack of humanity. One could argue it is a sign of mental illness or severe maladjustment somewhat comparable to having a long term heroin addiction. Huang has more than enough resources for the S-class cost to never even cross his mind and yet he publicly brought it up in a very pompous and almost child-like manner.

  • Apple using Intel as a fab is an interesting development.

    FWIW, the headline used the term "manufacturing", which to me implies fabrication (more so than design).

  • Sounds extremely stupid, but what do I know? My primary computing platform is a dual monitor desktop.

  • I remember practicing for the SAT ~20 years ago. I hated it, other exams could be somewhat interesting, it felt like it was designed to be a chore.

    Back then, my family ordered very expensive practice books (kids didn't usually take SAT exams in the country I was living in). I don't know what the situation is now, but this actually seems like a good use of LLMs.

  • Heatsinked, but I still feel like some of the early PCIe 4.0 SSDs are less refined than later 4.0/5.0 models.

    I am not dropping the speeds (I do video stuff as a hobby and it helps).

  • I have a pretty well ventilated case that I clean regularly, I think the case temp is fine. The early PCIe 4.0 SSD is what I am most worried about.

  • For real, I can't see me moving off my AM4 build for another ~3 years at least. Thankfully it works pretty well, except the 5800X is starting to show its age.

  • When questioned about the rock-and-hard-place situation these users are in, McAfee stated that AMD "[is] certainly looking at everything that [it] can do to bring more supply and kind of reintroduce products back into the [AM4] ecosystem to satisfy the demands of gamers that maybe want that significant upgrade in their AM4 platform without having to rebuild their entire system", further adding that he thinks this is "definitely something [AMD is] very actively working on."

    It looks like they are considering some sort of AM4 release, doesn't mean it will be with X3D cache.

  • Is this confirmed? I am not seeing any official confirmation from AMD.

  • It's been a while since I've tuned GPUs, and that too was an ancient laptop 760M (modded BIOS to unlock core and memory clock, I used the laptop as a desktop and it had additional cooling), the general impression I get is that tuning with modern computer hardware is a case of diminishing returns (RAM OC notwithstanding).

    That being said, do you have something specific in mind? Running the GPU cool and quiet? Getting the most out of it irrespective of power consumption? Don't have any experience with Arc (only had shit-tier intel iGPUs from before 2020), but I think this would help people answer your question.

  • I would have preferred an AM4-based 5850X3D!

  • It might take a while. And one might end up in a situation where there will be other things to worry about if the collapse of the AI bubble wrecks the economy.

  • Sure, but overwhelming majority of people buy Macbooks because of the design and ecosystem, not because of unified memory.

    Of course it doesn't require desktop-level cooling and power availability, but chances are if you're not running MacOS (which has a broad set of use cases, including ones where unified memory is not the key adoption driver), you do want to make use of the additional power and cooling. That's why I cited the Nvidia Spark, the Framework desktop and the AMD-based mini-PCs with unified memory). There is a reason why Framework, known for making laptops, went with a desktop solution for a unified memory desktop.

    I am willing to bet, Nvidia is not going to offer large memory sizes on their N1 SoC, they want people to buy the Spark and that's not their target area.

  • I wouldn't trust a local LLM solution from a large American company. Not saying that they would try to "pull a quick one", but they are unreliable and corrupt.

  • Nadella maybe knows a lot more than any of us about LLMs/GenAI tech, but one doesn't need to know anything about LLMs (or even technology) to know that an oligarch like Nadella cannot be trusted (in any context).

  • I don't mean restrict, I am talking more about market demand.

    Desktops are a bettee fit in terms of cooling and power availability for long throughout tasks (not burst type use cases).

    These is a reason why both Framework and Nvidia have released desktops with large unified memory options (as well as many smaller mini-pc manufacturers) and not laptops.

  • Thanks for the suggestion, I am so not looking forward to moving the family off Nova (lots of tech support), but it must be done.

  • Of course, of course!

    You're not at all engaging in demagoguery and trying cheaps tricks (no one can read the thread OP what it says specifically)! So what kind of discussion can there be about no one buying into your narrative, right?