A.I. groks 66%-76% faster with data augmentation strategies.

Comments

• catloaf@lemm.ee ⁨2⁩ ⁨months⁩ ago

  AI doesn’t grok anything. It doesn’t have any capability of understanding at all. It’s a Markov chain on steroids.

  source

  • Hackworth@lemmy.world ⁨2⁩ ⁨months⁩ ago

    Did you read the paper? Or at least have an llm explain it?

    source

    • yesman@lemmy.world ⁨2⁩ ⁨months⁩ ago

      I read the abstract, and the connection to your title is a mystery. Are you using “grock” as in “transcendental understanding” or as Musk’s branded AI?

      source

      • -> View More Comments
  • Thorny_Insight@lemm.ee ⁨2⁩ ⁨months⁩ ago

    …is how AI haters redefine terms and move the goalposts to fight their cognitive dissonance.

    source

    • technocrit@lemmy.dbzer0.com ⁨2⁩ ⁨months⁩ ago

      Imagine believing that AI-haters are the ones who redefine terms and move goalposts.

      source
• eggymachus@sh.itjust.works ⁨2⁩ ⁨months⁩ ago

  Thanks for posting, please ignore the stochastic luddites 🙂

  source

  • Hackworth@lemmy.world ⁨2⁩ ⁨months⁩ ago

    I appreciate it. I’ve had little luck engaging people in conversation about AI research in general. Since abandoning reddit, I’ve been litmus testing other platforms. I’m afraid this puts lemmy at about a 12. /shrug. Still better than reddit.

    source

    • Thorny_Insight@lemm.ee ⁨2⁩ ⁨months⁩ ago

      I couldn’t think of a worse platform to try and discuss this topic than Lemmy. The consensus here is essentially that big companies = bad, AI companies = big, and thus AI = bad.

      source

      • -> View More Comments