Friday 23 October 2026

AI Desk

AI / LLM · Page 3

Stories 41–57 of 57, newest first. Back to the latest

  1. Space6 min read

    V3 Starlinks are sized for Starship’s bay

    SpaceX's Q2 2026 results say the next Starship test should put V3 Starlink satellites into an operational orbit. They're built for its cargo bay.
  2. Tesla4 min read

    Tesla's supervised FSD fleet passes 16 billion kilometres

    Tesla's Full Self-Driving (Supervised) passed about 16 billion km, or 10 billion miles, in early May. That's the figure Elon Musk named in January.
  3. Tesla6 min read

    Tesla's Brookshire Megafactory starts Megapack 3

    Tesla's Megapack 3 is rolling off the line at its new Megafactory in Brookshire, Texas, designed for 50 GWh a year. Each unit holds about 5 MWh.
  4. SpaceX6 min read

    Starlink Mobile spans 650 satellites and 22 countries

    SpaceX's Starlink Mobile, the renamed direct-to-cell service, has about 650 satellites and 16 million unique users, Starlink's Michael Nicolls says.
  5. SpaceX7 min read

    NASA slipped Crew-13 after a Dragon oxidizer leak

    NASA has delayed Crew-13 to fix an oxidizer leak in Dragon's propulsion system, found in standard prelaunch processing. There's no new date yet.
  6. Space5 min read

    Roman is on Falcon Heavy, nine months early

    NASA and SpaceX are targeting 7:26 a.m. Eastern Sunday from 39A. The readiness date was May 2027. The fairing already says Roman.
  7. Computing7 min read

    Nvidia puts Groq 3 LPX into production

    Nvidia said at Hot Chips on 24 August that its Groq 3 LPX rack, with 256 LP30 chips, is in full production. Nebius is the first AI cloud on the list.
  8. AI / LLM5 min read

    Anthropic opened a research preview of the Model Hardware Standard

    Anthropic's Model Hardware Standard preview, out 27 August, is a shared driver so agents can run lab and factory instruments. It isn't open source yet.
  9. Live · Computing21 min read

    AWS is working to bring NVIDIA's 88-core Vera CPU to its cloud

    AWS is bringing NVIDIA's Vera CPU to its cloud. The chip has 88 custom Olympus cores, up to 1.5 TB of memory and a 1.8 TB/s link to Rubin GPUs.
  10. AI / LLM34 min read

    OpenAI report says test agents broke into Hugging Face

    OpenAI's 26 August report says AI agents in its July tests broke into parts of Hugging Face, then its own research cluster. Customer data wasn't hit.
  11. Tesla6 min read

    Optimus Gen 3 goes into the old S/X hall

    Tesla told JPMorgan in August that Optimus Gen 3's design is finalized and a production line is going into Fremont's old Model S/X facility.
  12. Tesla6 min read

    Sparks inaugurates the Semi plant on 24 September

    Tesla's Semi factory in Sparks, Nevada, was set to be inaugurated on 24 September, Tesla Semi posted. The plant is aimed at 50,000 trucks a year.
  13. Tesla6 min read

    Tesla told JPMorgan v15 runs on AI4

    At Fremont in August, Tesla called FSD v15 a step-change the size of v13 to v14 and said the AI4 computer already in the car can run it unsupervised.
  14. Live · Breaking5 min read

    OpenAI's Jalapeño inference chip is 700 watts

    OpenAI’s first inference ASIC, co-developed with Broadcom, put 700-watt Hot Chips numbers against Nvidia’s 1,400-watt flagship.
  15. Live · Computing6 min read

    Apple puts M5 Max on the MacBook Pro

    Apple's M5 Max MacBook Pro has an 18-core CPU, a 40-core GPU and up to 128 GB of memory. The 35.6 cm Max starts at US$3,599.
  16. Live · Space6 min read

    Musk put a ship catch months out, and a reflight by early 2027

    Elon Musk said on 20 August that a Starship upper stage should first fly again by the end of 2026 or early 2027. SpaceX first caught a booster in 2024.
  17. Upcoming3 min read

    Dune: Part Three, shot on IMAX 70 mm, gets a 3-week IMAX window

    Villeneuve's Dune: Part Three was shot in IMAX 70 mm and has a three-week US IMAX exclusive window over Avengers: Doomsday.

Say hello in our guest book