Thread

How OpenAI Used Its Own LLMs to Design Its Jalapeño Chip

206 points · 139 comments · maxall4

amelius · 1789774081 · focus · HN ↗
At some point people will use an LLM to design an Apple M series competitor.
1. bigyabai · 1789774233 · focus · HN ↗
  They won't, because they'd need an ARM architecture license.
  1. pixl97 · 1789774463 · focus · HN ↗
    I mean you can design anything without a license. Selling it is where the problems come up. Even then there are likely places in China that would still make it for you.
  2. amelius · 1789774700 · focus · HN ↗
    Why, the LLM can make up its own architecture.
    The value lies in the design space exploration, which is what an LLM can easily do.
    <a href="https://en.wikipedia.org/wiki/Design_space_exploration" rel="nofollow">https://en.wikipedia.org/wiki/Design_space_exploration
  3. wmf · 1789775157 · focus · HN ↗
    Arm sells architecture licenses to anybody these days.
    bpicolo · 1789835001 · focus · HN ↗
    Correct me if I'm wrong, but don't legal regulations around monopolistic behavior kind of mandate that?
    wmf · 1789838587 · focus · HN ↗
    There are some investigations starting about this but no rulings yet.
  4. cmrdporcupine · 1789777733 · focus · HN ↗
    Or they'll just build a competitor in RISC-V instead and that's fine.
    Except the problem is not restricted to the actual ISA or its HDL implementation, etc.
    It's even just getting space / time in a fab at that advanced of a process node.
  5. nr378 · 1789778729 · focus · HN ↗
    Qualcomm have an architecture license and the Snapdragon X2 Elite Extreme X2E-96-100 isn't too far off the M5 Pro.
    [1] <a href="https://browser.geekbench.com/processors/snapdragon-x2-elite-extreme-x2e-96-100" rel="nofollow">https://browser.geekbench.com/processors/snapdragon-x2-elite...
    [2] <a href="https://browser.geekbench.com/macs/macbook-pro-14-inch-2026-18c-cpu-20c-gpu" rel="nofollow">https://browser.geekbench.com/macs/macbook-pro-14-inch-2026-...
    bigyabai · 1789782206 · focus · HN ↗
    Are they using LLMs to close that gap, or is this their Nuvia acquisition doing the heavy lifting?
    pazimzadeh · 1789789813 · focus · HN ↗
    watt about in performance per watt?

Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.