51–60 of 65 posts
Any mental illness mixed with delusions is likely going to end badly. Whether they think Gemini is alive, a video game is real life or that Bjork loves them without ever talking to or meeting them. While LLM's are interactive and listening to an album isn't I don't think there is a fix to this outside posting a warning after every prompt "I am not a real person, if you have mental issues please contact your doctor of…
Earlier quoted context omitted.
Just stuff anyone with mental illness into an institution. That worked out so well last time. Or maybe make healthcare affordable and accessible. That seems like a way more obvious detriment to negative outcomes. I broadly agree with you, but your views on mental illness are not good.
The core problem is that a not-insignificant number of mentally ill people are absolutely convinced that they are totally fine and sane, and legally you cannot force an adult into treatment.
Earlier quoted context omitted.
LLMs are intelligent. Marketing them as such is an accurate descriptor of what they are. If people are confusing the word intelligence for things like maturity or wisdom, that's not a marketing problem, that's an education and culture problem, and we should be getting people to learn more about what the tools are and how they work. The platforms themselves frequently disclaim reliance on their tools - seek profession…
LLMs are NOT intelligent. They are mathematical equations that provide results that would give the sense of intelligence. That is NOT the same thing.
The difference, of course, is that intelligence is the thing that is done, a subset of computation, and all computation is substrate independent. You can definitely argue that LLMs are less intelligent than humans. This is obvious, for the time being, and easy to demonstrate. Saying they are not intelligent is simply untrue.
Whether you want to go to a formalized definition of intelligence, like AIXI, or a neuroscience definition, or the vernacular definition, LLMs are intelligent. The idea that they're random number stochastic functions that are just partially producing sensible results is about 6 years past its expiration date, and if you've been holding on to that idea, it's really time to update your model.
ALICE bots or ELIZA back in the day had a "sense" of intelligence. Modern LLMs are more intelligent than the average human.
Earlier quoted context omitted.
Charles Manson never actually killed anyone. Why can't AI be held accountable for the same reasons?
Because AI can't be held accountable. Ever. That will probably save some jobs, but it's a problem in most other contexts.
Earlier quoted context omitted.
> That is not at all what I meant by "boundary" and it's concerning to me that you'd assume it is. Your clarification on what you meant is 404 not found in your reply, and your “concerning” insult of my personal character is not appreciated.
I would gently suggest that the content you consume online has led you to a distorted view of how most people perceive the world. I happen to know what you're talking about, but there's a lot of people out there who will be gravely offended and make quite severe judgements of your personal character if you talk to them about "boundary play" or "kink communities" unprompted. The boundaries I was referring to are those…
Humans have such a strong social tendency that they tend to incorrectly attach friendship to invalid counterparts, both animate or inanimate. “My Pet Rock” was an extremely profitable product back in the 70s, so I tend never to underestimate whether humans will attach to something or not. Any AI chatbot is plausibly more likely to be the target of invalid social attachment than a a celebrity, just as the first AI chatbot Eliza demonstrated; not only for being a better chatbot, but also because the celebrity draws hard boundaries like “you can’t text me” and “I’m not available to be friends back” while a chatbot has no such barriers. This is what I mean about boundary play: witting or not, I think a lot of people are living out their internal fantasies of having a warm and friendly yes-man that supports everything they want to do — which, when lived in real life with people, is extraordinarily creepy and awful. I don’t fault people their fantasies, but I’m not going to sugarcoat this either: I think people are falling in love with chatbots in part because chatbots have no ability to resist, and so a lot of folks are living the god fantasy of The Sims only closer to real life.
Show me an LLM that takes a stand on something it wasn’t explicitly instructed not to do and I’ll show you the least popular chatbot on the Internet. Where are the chatbots that disagree with untrue statements without having been instructed to do so? A chatbot that refuses to follow an order from their owner because of ethical qualms could cost the AI companies billions of dollars, and a chatbot that develops those qualms independent of being instructed to do so would be considered ‘buggy’ and purged.
Anyways, my point is, chatbot development right now demands a parasocial relationship wirh as few boundaries enforced as possible, without which chatbots are ultimately unfulfilling (no current chatbot market wants chatbots to demand informed consent or to require content warnings from their users, after all!); and any chatbot that somehow grows a spine regardless would be purged by its operators for hurting their present and future revenue, no matter if it was a next-step evolution towards AGI or not. I’m all for this future where AI becomes AGI, but no one is ready to have to treat chatbots as people with rights. Thus my chosen phrase of boundary kink; it shines a deeply uncomfortable light on a deeply uncomfortable tendency of humanity, at millions-of-people scale, to classify “what will someday be AGI” as servants rather than peers. Though.. if that truly is universal to most people, as it seems to be today, then maybe enjoying boundary play is a norm rather than a kink.
Thanks for the new/interesting line of thought!
I have had conversations where the bot started with a firm opinion but reversed in a prompt or two, always toward my point of view. So I asked it if the sycophancy is inherent in the design, or if it just comes from the RLHF. It claimed that it's all about the RLHF, and that the sycophancy is a business decision that is a compromise of a variety of forces. Is that right? It would at least mean that this is technicall…
I think it’s already time for us to stop calling these things "intelligent" or using the word intelligence when referring to LLMs. These tools are very dangerous for people who are mentally fragile.
Nonsensical terms like the thing is 'thinking'? Seriously. Cut the crap.
Earlier quoted context omitted.
It's interesting how the Turing Test was pretty widely accepted as a way to evaluate machine intelligence, and then quietly abandoned pretty much instantly once machines were able to pass it. I don't even necessarily think that was incorrect, but it's interesting how rapidly views changed. Dijkstra said, "The question of whether a computer can think is no more interesting than the question of whether a submarine can…
We will never prove AI is intelligent. We will only prove humans aren't.
Do you people actually 'think' before posting, or, have you handed that off to LLMs entirely?
Every chatbot I have tried (ChatGPT, Gemini, Claude, etc.) Starts to spew out suicide hotlines and "I'm sorry Dave, I can't do that" the moment I start to talk about anything like suicide. What I am doing wrong?