‹ BackHN Continuity

Thread

The AI Race Just Got Awkward

412 points · 463 comments · allisdust

  1. reedf1 · · focus · HN ↗
    I've been running Qwen 3.8 27b (an opus 4.6 tier model), locally on a 5090 for just over two weeks @ 170 tokens/s. That's a frontier model from 9 months ago running on consumer hardware. Who knows where distillation and pruning gets us in another year.
    1. bix6 · · focus · HN ↗
      $9k for a 5090 now? Sheesh.
      1. off_with_their_ · · focus · HN ↗
        $9k is a small price to pay to experience the rapturous glory of AGI. I'd easily pay up to 3 times that to comfortably run the superintelligent models released in this post RSI world.
        1. literalAardvark · · focus · HN ↗
          Except you can do that cheaper by renting compute
Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.