you are viewing a single comment's thread
view the rest of the comments
[+] 11 points 4 months ago (43 children)
  • [–] 17 points 4 months ago (7 children)

    Thing is, it is already happening. Don't forget that teenager that killed himself because his AI girlfriend told him to.

    Also look at the other comment.

    It will become even worse once realistic robots become a thing, and there will most definitely be versions with holes.

  • source
  • parent
  • hideshow 7 child comments
  • [–] 2 points 4 months ago (4 children)
  • [–] 2 points 4 months ago (3 children)

    Wtf did I just watch 😭😭😭😭

  • source
  • parent
  • hideshow 3 child comments
  • [–] 2 points 4 months ago (2 children)

    A board game review by The Dragon's Tomb!

  • source
  • parent
  • hideshow 2 child comments
  • [–] 2 points 4 months ago* (3 children)

    I have used LLMs quite a lot since they first became a thing and there is one thing I always notice. After chatting for a while it always feels like you are talking to a version of yourself who just happens to be knowledgeable in a vast amount of topics.

    At least for me it is similar to having an internal monolog just a lot more interactive. Considering that, I can definetly see why people find AI partners appealing. Because they are literally talking to themselves.

    I don't think that's a bad thing on itself, after all I used to have long conversations with myself all the time to help me arrive into conclusions. Just be mindful of that and use as a tool, because that's exactly what it is.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 5 points 4 months ago (1 child)

    Using LLMs that way with the knowledge that it is a sort of mirror of yourself works much better than pretending it's an actual consciousness that understands you. Think of it as a modern journal or diary that you can put thoughts into and then get back more organized, maybe with some ideas to branch from.

  • source
  • parent
  • hideshow 1 child comment
  • [–] -2 points 4 months ago* (4 children)

    the sad truth is for many people the social interactions the ai provides will be more positive than the real world ones.

    you assume that real world interaction is positive, it often, is not.

    like try being a gay kid in a conservative town... you will have zero positive social interactions, and the AI one will probably be the only one you have that doesn't hate you for existing.

  • source
  • parent
  • hideshow 4 child comments
  • [–] 2 points 4 months ago* (3 children)

    That's just the equivalent of taking drugs because you're depressed.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 2 points 4 months ago (2 children)

    Not, really. Antidepressants do actually work, they have clinical trials of decades behind them. With that being said they treats the symptoms, they don't fix the underlying cause. But treating the symptoms does allow you to start working on yourself so you slowly crawl out of the hole.

    Talking with LLMs and pretending they are anything but a tool all it does is further isolate you. They don't even understand what they write. They are text predicting machines guessing which words should probably go after the ones you gave as the input.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 2 points 4 months ago (1 child)

    I'm not talking about antidepressants, I'm talking about non-medical drugs. Many depressed people fall in that spiral of depressed ---> drugs ---> slightly less depressed ---> even more depressed ---> more drugs ---> repeat.

    Same with people talking to LLMs for social connections.

    I wasn't talking about medical antidepressants, these can help when taken in controlled amounts.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 2 points 4 months ago

    Ok, my bad. I thought by drugs you were talking about things like SSIRs which some are attempting to demonize these days. Illicit drugs on the other hand just provide a very short temporarily relief and just prolong the problem.

  • source
  • parent
  • [+] -6 points 4 months ago (18 children)

    This is the dumbest thing I've ever heard. Half of my friends are socially inept quite nerds who don't say shit online. All of them provide little socialization outside of small bursts or even we made it a point to hang out in person. It would be trival for me to fill a lot of those gaps if I didn't have other outlets. The lack of empathy around AI interactions and how they form relationships is astounding.

  • source
  • parent
  • hideshow 18 child comments
  • [–] 11 points 4 months ago (13 children)

    The LLM gets no empathy because it deserves none. It doesn't have feelings or emotions, it never will. At best it is a tool.

  • source
  • parent
  • hideshow 13 child comments
  • [+] -8 points 4 months ago* (9 children)

    And? Is amazing how much people will screech over non optimized solutions and risk mitigation from life challenges.

    Is it perfect? No. Can it make you feel better and fill gaps? Yes. Fucking worthless amirite?

    I'm not going to even get into how a model trained for medical purposes can be used to learn and enforce strong healthy habits and the excitement for medically trained models that are incredibly useful. It would be lost on this crowd.

  • source
  • parent
  • hideshow 9 child comments
  • [–] 8 points 4 months ago (7 children)

    Can it make you feel better and fill gaps? Yes

    So can meth. That doesn't mean we should encourage it.

  • source
  • parent
  • hideshow 7 child comments
  • [–] -5 points 4 months ago (6 children)

    Yeah, meth isn't used in any medications 🤣👍 never not even one

    Out of all the substances you could have chosen and you choose the one used as a bases for tons and tons of medications.

    👏👏👏👏👏👏👏👏👏👏👏👏👏👏 Bravo

  • source
  • parent
  • hideshow 6 child comments
  • [–] 5 points 4 months ago* (5 children)

    So... you think it's ok for everyone to be on meth like you clearly think it's ok for people to have AI companions..?

    rofl bravo indeed... Way to completely and utterly miss the analogy.

  • source
  • parent
  • hideshow 5 child comments
  • [–] -4 points 4 months ago (4 children)

    No shit for brains. It's okay to have models that are trained by medical experts to assist in medical diagnostics and treatments. While there is inherent risk in someone so dead set to beat the model, many if not most can use the tools in monitored controlled setting with minimized risk and outcomes that put weigh them.

  • source
  • parent
  • hideshow 4 child comments
  • [–] 1 point 4 months ago (3 children)

    That's not what we are talking about though. We are talking about people having personal relationships with Large Language Models.

  • source
  • parent
  • hideshow 3 child comments
  • [–] -1 points 4 months ago (2 children)

    Right. You are punching down at them making fun of them for being lonely. Reaching out looking for a need going unfilled. Vilifying everyone involved and creating a stigma around a potential tool and a completely reasonable place to reach out to when you feel desperately alone.

    I understood what you were doing.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 1 point 4 months ago* (1 child)

    The people that are already in the pit of despair are not who I am trying to save by stigmatizing. By stigmatizing, I try to prevent the normalization of AI relationships, to prevent people from ending in that pit of despair before they get there.

    The people that are already in that pit of despair can only be saved by seeing the light. I can't provide that to them, because I'm also just a stranger on a screen. That should be somebody real or on a platform meant for socializing instead of discussion.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 2 points 4 months ago

    Go ahead and do it if it makes you feel good. Just be aware of what you are actually doing. Chatting with an LLM can be fun, the problem is once you start seeing it as a conscious being that actually understands what it writes, let alone understand you. Because that couldn't be further from the truth.

  • source
  • parent
  • [–] -1 points 4 months ago (2 children)

    That says more about how you treat your tools and view your material surroundings than speaking to any sort of universal truth.

  • source
  • parent
  • hideshow 2 child comments
  • [–] -1 points 4 months ago (5 children)

    Not at the moment, but there's no reason to think it could never happen, unless you think it's impossible to simulate a human brain, which seems an unlikely thought.

  • source
  • parent
  • hideshow 5 child comments
  • [+] 7 points 4 months ago (4 children)
  • [–] 1 point 4 months ago

    https://flybrain.app/

    I know you mean human brain, but the point is that the principles largely exist and the progress is accelerating.

    However, one of the key learnings from the linked project is that the brain and body are intrinsically linked and biology is a nightmare mess of spaghetti code. Simulating just the brain has limited benefits and will not likely simulate a person or consciousness any more than simulating small pieces of a brain.

    For some further reading: https://www.sciencedirect.com/science/article/abs/pii/S030645222100021X (scihub)

  • source
  • parent