Comments by "WloCkuz" (@wlockuz4467) on "Sabine Hossenfelder" channel.

  1. AIs becoming sentient is always a marketing gimmick. They way LLMs are built makes it fundamentally impossible for them to become sentient, that is assuming we can even quantify what it means to be sentient. However, you can obviously bake behaviours into the LLMs that make them output certain type of responses, Claude sounding smart for example, but that's not really becoming sentient.
    1
  2. Funny how little this gets talked about. I guess the AI hype wagon doesn't want this to be common knowledge because it means less business for them.
    1