Avatar from Dicebear.

  • 1 Post
  • 30 Comments
Joined 1 month ago
cake
Cake day: September 14th, 2025

help-circle







  • In fact, according to The Register, the GPU computing performance of the GB10 chip is roughly equivalent to an RTX 5070. However, the 5070 is limited to 12GB of video memory, which limits the size of AI models that can be run on such a system. With 128GB of unified memory, the DGX Spark can run far larger models, albeit at a slower speed than, say, an RTX 5090 (which typically ships with 24 GB of RAM). For example, to run the 120 billion-parameter larger version of OpenAI’s recent gpt-oss language model, you’d need about 80GB of memory, which is far more than you can get in a consumer GPU.

    Or you could’ve just made GPUs, and then we’d all be gaming and calling each other shitheads in Valorant instead of - checks notes - literally stealing the water from poor communities.







  • “It’s almost certain” that AI will reach that level eventually, one researcher told Nature.

    Semafor doing so much work trying the launder this into a story. “One scientist” in the original article, to multiple scientists in their headline.

    This is the first of three waves of AI in science, says Sam Rodriques, chief executive of FutureHouse — a research lab in San Francisco, California, that debuted an LLM designed to do chemistry tasks earlier this year.

    And the one “scientist” seems to have switched tracks from doing actual research to doing capitalism.


  • The Yale researchers’ nothingburger result has precedent. In 2023, a study by the United Nations International Labour Organization (ILO) concluded that generative AI would probably not replace most workers.

    A study of Danish workers published in April determined that generative AI had no material impact on wages or jobs. Another such study published in February found “overall employment effects are modest, as reduced demand in exposed occupations is offset by productivity-driven increases in labor demand at AI-adopting firms.”

    There is some contradictory data.

    No shit.






  • Launching October 1st, Gemini For Home is a suite of new AI-powered features for Google’s smart home hardware and software.

    The biggest change: Gemini is replacing Google Assistant on all of Google’s smart speakers, all the way back to the original Google Home speaker. This LLM-powered upgrade, announced at Google I/O, will be available through an Early Access program at first, with a wider rollout planned for next year.

    On smart speakers, Gemini brings an entirely new voice assistant that uses and understands natural language, can interpret context, and can pull in more real-time information. You still activate it with the wake words “hey Google,” but Google Assistant has been evicted.

    “Gemini for Home is the intelligence for your entire home,” Anish Kattukaran, head of product at Google Home and Nest, tells The Verge. “It’s not going to just replace Assistant on speakers and displays, but it’s going to upgrade your other devices as well, your cameras and doorbells, where you interact with those devices, and bring those smarts collectively to your entire home.”

    I’m not excited for Apple to invent smart homes after this, completing the duopoly of LLMs being in everyone’s homes even harder than before.

    Long live Home Assistant