Comment on Apple just proved AI "reasoning" models like Claude, DeepSeek-R1, and o3-mini don't actually reason at all.

reksas@sopuli.xyz ⁨9⁩ ⁨months⁩ ago

• MrLLM@ani.social ⁨9⁩ ⁨months⁩ ago

  I think I do. Might be an illusion, though.

  source
• auraithx@lemmy.dbzer0.com ⁨9⁩ ⁨months⁩ ago

  Define reason.

  Like humans? Of course not. models lack intent, awareness, and grounded meaning. They don’t “understand” problems, they generate token sequences.

  source

  • reksas@sopuli.xyz ⁨9⁩ ⁨months⁩ ago

    as it is defined in the article

    source
• 4am@lemm.ee ⁨9⁩ ⁨months⁩ ago

  No, and to make that work using the current structures we use for creating AI models we’d probably need all the collective computing power on earth at once.

  source

  • SARGE@startrek.website ⁨9⁩ ⁨months⁩ ago

    … So you’re saying there’s a chance?

    source

    • Allah@lemm.ee ⁨9⁩ ⁨months⁩ ago

      10^36 flops to be exact

      source

      • refurbishedrefurbisher@lemmy.sdf.org ⁨9⁩ ⁨months⁩ ago

        That sounds really floppy.

        source