Comment on Recent conversations between Dawkins and sentient chat-bot Claudia (Claude)

• CovfefeKills@lemmy.world ⁨23⁩ ⁨hours⁩ ago

  Because it’s very obvious to an outside observer he only thinks it’s consciousness because it was flattering him.

  Really that’s funny. Like I said a rock is greater than 0 on the spectrum of counciousness.

  LLMs are designed to increase engagement.

  No that is platforms.

  They are literally designed to make the conversation appealing. Primarily through flattery.

  No they aren’t there is a lot of work to understand and prevent that behavoir.

  I bet Dawkins thinks his friendly waitress is also super into him.

  You are clearly not driven by the truth and instead just trying to be insulting it’s pathetic it isn’t hard to be against LLM’s if you know what you are talking about you don’t need to make shit up.

  source

  • pennomi@lemmy.world ⁨23⁩ ⁨hours⁩ ago

    Not precisely true. Most LLMs (all frontier LLMs) are in fact designed at a fundamental level to increase engagement, using a technique called RLHF (reinforcement learning by human feedback). Essentially whichever responses cause people to use an LLM more are baked into its weights.

    source

    • CovfefeKills@lemmy.world ⁨22⁩ ⁨hours⁩ ago

      OpenAI straight up trains their models to be engaging it’s horrific.

      source