home.social

#gtc26 — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #gtc26, aggregated by home.social.

fetched live
  1. Speculative decode is an inferencing optimization that was mentioned a few times at #GTC26. I'd heard of it but didn't know how it worked, so I spent some time figuring it out. Some notes (and toy code that illustrates its benefits!) are here: glennklockwood.com/garden/specu... #AI

    speculative decode

  2. Speculative decode is an inferencing optimization that was mentioned a few times at #GTC26. I'd heard of it but didn't know how it worked, so I spent some time figuring it out. Some notes (and toy code that illustrates its benefits!) are here: glennklockwood.com/garden/spec

    #AI

  3. My latest: Dell Technologies jumps into the ring with NetApp and VAST Data, unveiling a new #AI #dataorchestration product built on its Dataloop acquisition, as NVIDIA #STX shakes up the #enterprisedatastorage industry.

    Key quote: "The #cuDF and #cuVS integrations quietly showing up inside #Snowflake, #Starburst, #watsonx -- those matter more to an #enterpriseIT leader's next 12 months than anything Jensen [Huang] showed on the big stage."

    Find out why that is, as well as how the Dell Data Orchestration Engine stacks up to competitors: techtarget.com/searchstorage/n #GTC26 #Nvidia

  4. Nvidia outlines AI factories and agentic AI.
    Jensen Huang highlights physical AI for robotics, industry, and next-generation computing platforms in GTC keynote. #GTC26
    buff.ly/CNFaYbp

  5. These new NVIDIA STX storage node implementations from Quanta and Supermicro appear identical even down to the LED and USB-C placement. I wonder if the drive carriers are differentiated.

    #gtc26

  6. NVIDIA’s Kyber rack combines NVLink switch ASICs into what used to be the cable backplane. It’s physically huge—height of the whole rack. I wonder what the FRU is on this monstrosity. #GTC26

  7. What stood out from the avalanche of #Nvidia #GTC26 news for many observers wasn't a chip or hardware system -- it was deeper forays by NVIDIA into securing and even creating #AIapplications with its new #NemoClaw for #OpenClaw reference architecture and #AgentToolkit.

    Hear from analysts Jim Mercer, Torsten Volk, Jim Frey, and Michael Leone, alongside keynote comments by Jensen Huang and an interview with Yuval Fernbach, as they break down the ways enterprise applications -- and #AppSec -- are dramatically changing: techtarget.com/searchitoperati

  8. Stoked seeing the OpenSearch Project featured by Jensen Huang on keynote! 😍

    One of the innovations in V3 has been adding GPU acceleration based on NVIDIA's cuVS. Our benchmarks, using CAGRA algorithm integrated through Facebook's Faiss library, showed:
    ✅ 9.3x faster index builds
    ✅ 3.75x lower cost
    ✅ 2x higher throughput
    ✅ 2.5x lower CPU usage

    linkedin.com/feed/update/urn:l

  9. Literally feel like I got hit in the face with the NVIDIA #gtc26 news firehose today. Ow.