Login
You're viewing the functional.cafe public feed.
  • Nov 18, 2026, 9:28 PM

    @theodoraward It will never escape me that the same people trying to convince us that Claude is alive and can feel pain are so eager to dismiss the pain that people of color, queer people, and disabled people experience under oppression. They're prioritizing their mentalist fictions over real, thinking, living, breathing *humans*.

    💬 1🔄 10⭐ 7

Replies

  • Nov 18, 2026, 10:05 PM

    @xgranade @theodoraward I've been saying for a while that the thing that believing the chat bot is alive reveals is how fucking small that person thinks of the actual people around them.

    💬 1🔄 0⭐ 0
  • Nov 18, 2026, 11:01 PM

    @pathunstrom @xgranade @theodoraward I've noticed (and have said, probably not eloquently) that the reason fascists (which these people are) don't think of themselves as hypocrites is that they simply do not consider minorities, women, etc, as people.  So when they talk about being "for people", they mean a really, really, really small subset of people that just so happen to look, act, and think like themselves.

    It's... yeah.  They don't think BIPOC/LGBTQ/Disabled people are actually people.  They think, very explicitly, that we're subhuman.  They *used* to be more open about just saying that; they tend to hide it nowadays, but their actions make it *very clear* that they think that.

    💬 1🔄 1⭐ 0
  • Nov 19, 2026, 10:59 AM

    @aud @pathunstrom @xgranade @theodoraward this line of thought also goes into to how they want to use AI. If someone decent was worried it might be conscious, they'd start making plans for AI rights and abandon any idea of using AI by just paying the electricity costs, because having a conscious being do uncompensated work at your every whim of slavery (reminder here: they're dead wrong about the consciousness part). They're saying two things: "I want to use AI as a tool" and "I think AI is conscious" that put together amount to "I want to use a maybe-conscious thing as a tool." To hold that thought in your head implies that you're probably fine with treating some humans as tools too.

    I can't find the exact source again, but one interesting thing about human consciousness we can learn from anesthesia studies is that it seems to depend on brain wave patterns. Here's one article that talks a bit about this:

    neurosciencenews.com/how-the-b

    Neutral networks kind of stimulate networks of neurons, but they only simulate activation propagation in a very simple way that doesn't allow for waves of activity like those that occur in the human brain. So there's a very simple argument that for now, neural networks as currently designed lack the proper structural correlates to the human brain to support the patterns of activation that we know are a prerequisite for consciousness in humans. It's a bit of a shallow argument that misses the deeper questions about what consciousness is, but I think it's useful to cut through some of the "LLMs rival the complexity of the human brain" hype.

    As someone who has done a bit of deep learning research (although not with transformer models), I've never had an interaction with an LLM that remotely suggested to me anything close to consciousness, because I understand how they work. I've certainly been surprised by their capabilities, but I really wish I had a simpler way to share my understanding of how LLMs work that others would actually get in a way that helps inoculate them from the hype...

    💬 0🔄 1⭐ 0