DALLE 3

    • intensely_human@lemm.ee
      link
      fedilink
      English
      arrow-up
      2
      ·
      8 months ago

      Video training and production is gonna be a whole new ballgame for the AI in terms of understanding the world.

      Like imagine trying to interpret the human world from photos without ever, not even once, seeing it move. Just frozen images. Patterns in the images, but not always things you understand. Balls sometimes in the air, often at the ground, sometimes blurred in a particular direction. Then you see a movie of a ball bouncing, and other movies of things falling, and it all comes together.