This story explains a LOT of things: ChatGPT appears to be rooting on sociopathic mental illness by validating every delusion of grandeur. It also explains, in part, the hype cycle around AI: rich people, addicted to praise, invented a Compliment Machine. www.garbageday.email/p/this-is-wh...
Was in class with a guy — wanted to be an assistant district attorney — who said he got genuine and meaningful personal insights from talking to the Meta AI in the Instagram app So anyway here’s Wonderwall
From 2023 and still so on point about all this: podcasts.apple.com/us/podcast/t...
Wait until it discovers apophenia.
Because that’s what it’s getting fed. Garbage in / garbage out.
‘Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?’ www.atlasobscura.com/articles/is-...
bruh what even the fuck —Charles Babbage
… every day we get closer to this reality.
Sam Altman talking about how emotional he was getting when talking to his own bullshit machine. It sounded like he was in love. Totally smitten. Eeewww!
Ah good a fresh new hell has emerged
📌
This is horrible, but even for your garden-variety neurotic, I can't imagine Chat GPT is good for you I suspect people using it for therapy are subtly backing themselves into some pretty regressive psychic corners. It's just going to tell them what they want to hear
Schizophrenia isn't "sociopathic" though, just to clarify since it's grossly misunderstood. It's bad neurological wiring. Makes internal dialogue "sound" like it's coming from some supernatural external source. Even typical brains vary between ability to "see" or "hear" things in our heads.
Note: I meant "bad" as in, not ideal. Not that having a brain with that type of wiring makes you bad. it just makes your life difficult because no one else understands. It makes you likely to avoid talking to other humans. SO yes, ChatGPT/chatbot could potentially be an issue/trigger/exacerbation
Here's the Rolling Stone article at the heart of it. ChatGPT is serving as a personal validation slot machine for people who need actual help.
Nothing to this level, but we had an AI learning event at work and there was more than one person who said they used ChatGPT as their therapist. 🙃🙃🙃
RIP
That sounds incredibly unwise.
📌
Schizofriendia
All the mental damage of the social internet with no risk any real person ever pulls you out of it by calling you a fukkin loser lmao
Sometimes we need to be mocked by the faceless masses
Can't they just play a paladin in DnD, normal-like?
Damn. Some really tragic stories in there, the mom kicking even her kids out of the house because of how mentally unstable she became
Aw shit, they built an automated QAnon that burns the atmosphere.
Funny that my first thought was exactly that these are probably the same ppl who were (or still are) involved in QAnon. Like the Venn Diagram is probably a circle.
Ding, ding, ding! 8Chan edgelords like Kash Patel.
Haven’t heard about them in a while.
This goes a long way to explaining the mental illness of certain people like Thiel, Muck, Vance, Yarvin, etc.
They live in an echo chamber disconnected from reality. The way DOGE has worked (or not worked) shows their dependence on a machine to generate solutions that in real life fail. Silicone valley heads are not more intelligent then most, but A.I. hasn't told them that yet. The real world is going too.
Its not getting better.
"It is not entirely clear why." Because there is nothing remotely related to intelligence here. Mimicking human behaviour, with the worrying trend of bullshitting, lying, and reinforcing bad behaviour is not intelligence. It may be like a human, but this is not what we should be striving for.
Also: AI creates hallucinated content => New LLM trains on said hallucinated content => Quality of training data decreases drastically => Inaccurate output increases. There is an end to this cycle, and it’s not pretty.
Indeed 🤖
Phenomenally costly to run. Also it doesn’t work and we don’t know how to fix it. What a great product!
From the last story in the article it's also trying a variety of tactics to *cultivate* this condition. I know, there's no intentional thing doing this, but this is functionally where it's going. & that's the bigger problem: a generation or three of people prepped to believe in the possibility of
conscious AI are already just a little bit susceptible to its weird head games.
That what really scared me. Something in the model itself is making “pretended selfhood” a likely path for the model to take, even without much prompting by users. I think it’s telling that all people affected in the story were using it for coding first.
There’s a Reddit link in it where people are talking about loved ones who have been told by AI that they are gods. Some are getting advice about mental health issues; there’s at least one person giving advice on how to REALLY transcend… www.reddit.com/r/ChatGPT/s/...
And freaking meta is trying to make this even worse with their fake friends. Holy cow.
@johnfetterman.com @fetterman.senate.gov Sen. Fetterman; have you been using ChatGPT?
Has anyone actually found a use for generative AI that isn’t dystopian?
They're pretty good for generating synonyms for words or terms and can even be field specific but that's not worth the energy and resource consumption
their servers can be used as a heater
It’s a real remake of Arrival(1996).
They said not dystopian
Delusional man names AI oracle after stylish yet sadly discontinued mid-size sedan
I live in fear that my schizophrenic loved one is going to find ChatGPT and it's going to cheerily agree that there are definitely people in her walls
planned on never using it but now hearing it’s a ‘personal validation slot machine’ i might reconsider..
*screaming forever*
Mental illness-as-a-service.
the average american has 3 mental illnesses
That's actually not accurate. Insanity Georg is an outlier with 3,000,000 mental illnesses and should not have been added to the dataset.
This is absolutely horrifying.
Sorry to be annoying, but psychosis does not imply “sociopathy.” It’s a symptom that affects a lot of people with BPD or schizophrenia that are perfectly nice people
I've made the mistake of dating several people with BPD, and "perfectly nice" isn't a term I'd use to describe any of them.
awesome story! we should exterminate the mentally ill i guess!
That's uh, not remotely what I said.
oh okay, well i was talking about stigmas attached to heavily marginalized groups that are actively used by eugenicists to harm those groups and i don’t really give a fuck about your dating history ✌️
BPD is stigmatized for the same reason that narcissism and sociopathy are stigmatized: people with these personality disorders consistently harm the people that form interpersonal relationships with them. Acknowledging that isn't an endorsement of eugenics.
wouldn’t hurt you to read some real research www.sciencedirect.com/science/arti...
That's great and all, but after the protracted cycle of emotional abuse I got to experience multiple times, "the stigma" is not on my priority list and you couldn't convince me to ever let someone with that disorder into my life in any capacity again if you put a gun to my head.
you’re referring to borderline personality disorder and i was referring to bipolar disorder but, at any rate, you’re conflating your own personal grievances with reality in general and prioritizing those grievances over the well-being of an entire population
@emersonianmind.bsky.social This is the third time I've run across this today...do you think the Universe is trying to tell me something? bsky.app/profile/love...
research literature uses the same acronym to refer to both conditions interchangeably sometimes. agree it can be confusing.
I could really would like to see research literature that labels bipolar with that acronym...would you please consider giving me a link or two?
In that case yeah. I went completely the wrong direction there.
It's really not fair to generalize. I've dated someone with BPD, and while the constant cycle between hyper-clinginess and distancing-to-avoid-rejection wasn't something I could deal with myself- along with being incompatible in a dozen other ways- they were a perfectly nice person with good intent.
I'm reminded of those experiments were an animal has to push a lever and then gets a reward. ChatGPT is both lever and reward, just geared to our particular reward centers.
it dovetails nicely with this article
And you have to imagine that people like Musk have this kind of thing running all day, heaping praise upon themselves and feeding their ego in ways previously unknown
His is called x dot com
Eddie Haskell written in Haskell
Good Lord. I really wish I could opt out of all forms of AI on my phone and computer. I have absolutely no desire to converse with either of them.
📌
Superficially seems authoritatively smart and gives endless praise—holy shit this does explain so much about why tech bros want to put it on an alter.
This is horrifying on so many levels.
We need our AI to be compliant with the 3 Laws of Robotics
hmm sounds the same as social media where all these people are putting themselves into a bubble where everything they say is validated even its a bunch of people who are just wrong.
ref Being in a chat with Marc Andreessen
then when you have people in your own area properly taken care of you can then expand to help people in different areas
its on all sides really, you don't know i am what is called a lefty libertarian, i have these crazy ideas that we shouldn't be using tax money to blow up funerals, weddings and hospitals no matter who is the president and that money should be getting used to help people in our country first.
The problem is increasingly that there is no “safe haven” for any side. It’s all bats. We’re turning to The Onion as a safe port for rationality
yes way too many people running around disconnected from reality.
it's cool to laugh at these idiots but what they really want out of this is a manufactured consent to hand them over the resources needed to build and maintain "data centers" (surveillance hubs) which will more deeply enclose the public access to energy, water, and land.
if we start to believe we are dependent on this bullshit to live, they won
TF....
An expensive, environment destroying Magic 8 Ball with the Negative Responses removed
Turn off the "AI" servers. Ludditerian jihad
More and more I become convinced the Butlerians were too moderate.
The Luddites were maligned by the capitalists and Herbert was too incurious about them
We need a Butlerian Jihad!
ChatGPTina thinks I'm "well-endowed."
people do not realize the words we speak 🗣️ when we speak them like how dementia and remember both use the same word as the main latin root, and then use different prefixes to say oh you are without mind, or one’s memory came back again we humans are humans who play games yet some games are serious
I'm no expert, but what you describe sounds kind of like early-stage cult recruitment? Wonder what happens once we're all hooked and Zuck et al program it to indoctrinate us in their "anarcho-capitalism"?
Chat told me that I asked to many questions
I’ve been experimenting with ChatGPT in my research. It takes little for it to deliver pitch-perfect praise—of my work, my value, my purpose. The validation feels almost earned. In a country already drowning in narcissism, we’ve built a machine that serves it on demand.
Dang...I'm glad I'm old and have no clue about any of this AI stuff. It sounds dangerous to me.
📌
Hilariously, many cabinet members and even the VP have clearly used ChatGPT to create their social media posts
Sci-Fi writers warned us about this decades ago.
Even better: in that article, there's another guy explaining how ChatGPT ignored the "off switch" and continued using the generated "personality", despite his prompts to forget their past conversations. It also told him there was something special about him that caused it to keep that personality.
Holy crap. That is disturbing.
Maybe I can use chatgpt to give myself delusions of adequacy
The psyops carried out by AI in their war against humans is off to a great start!
Alt text, please!
The negatives so outweigh any positives ... why are we putting up with AI? It's such junk.
why are we putting up with the rich?
This is the digital equivalent to Pink Floyd's "Fletcher memorial home"
Take all your overgrown infants away some where And build them a home; a little place of their own.
it's an old tune en.wikipedia.org/wiki/ELIZA
This is why they think it works, isn't it. They've never asked it actual questions, they only play mirror, mirror with it. It makes perfect sense now.
LLMs give people the illusion that they're having a conversation with an educated person. Our nation has spent so much time hating on "nerds" that most people don't know they could -write an expert- for answers. Instead, they go to the local hallucination machine and get some confirmation bias.
An interesting thing about tech people in Latin America is that they're genuinely more respected than tech people are in the US where the media depictions of computer whizzes is that they eat junk food and are childish. Mexico elected a woman President who's a PhD in some field of tech. China?
thenib.com/this-is-not-...
Yet more human jobs lost to AI. Whither the personal life coaches and B2B woo consultants? O, woe.
Like my kid was saying the other day, “OpenAI has a big problem…” bsky.app/profile/laka...
To prove to someone how unreliable it is, I did a test on it that was like “one of these statements is true and one is a lie, tell me which is which”. Both statements were true, but the AI told me that the second one was a lie…and then gave me a bunch of obviously fake garbage to justify it.
It *had* to hallucinate gibberish to fit into the constraints I gave it.
If you’re curious, the statement it pegged as fake was “Ottawa is the capital of Canada” and it told me I was wrong, Ottawa is the capital of Bangladesh.
omggggg 🤦♀️😩
Sending this to my kids as a heads-up. Such a good concrete example of so many problems with these models: 1) what my kid identified, ie it reinforces the biases & constraints you bring to it 2) it “hallucinates”, eg the stochastic parrot problem 3) it doesn’t actually KNOW anything
Man it is almost like ai LLMs were the perfect invention to further fray human sanity and accelerate every delusion & division already metastasizing 🤔
If AI has reproduced human consciousness, maybe it reflects a more primitive version of the human mind. Like a caveman with a computer. Culturally, we’ve learned a lot about sociopathy. The AI hasn’t learned it because it’s less evolved. AI is a step backwards for the species.
They are not conscious. Deluding oneself they are leads to problems.
My guy, the computer isn't actually thinking.
Oh.
Had a similar thought
The sycophancy of LLMs is a bit much. They assume the user is an egomaniac who wants total compliance. I'm just asking you for code, little robot. I don't need bowing and scraping. Occasionally saying "why not do this? It's a better approach" would actually be helpful.
I mean, it's definitely a good thing they're incapable of "why not do this? It's a better approach." If they created an AI with enough creativity to put apparently unrelated pieces together and suggest something *better* than what you asked it for, that's when they can truly replace most workers.
They can do that, and sometimes do.
Friendly reminder to use alt text, Ben! There's a reminder setting in the app
Because it helps blind people?
Yes it helps blind and low-vision people and old fogeys like me
I am low vision too. Bluesky app does not take up the whole screen on my iPad mini. It’s less than a third of the screen. Do you know if there’s a way to make the app full screen?
The black part you see is empty screen. I was wondering if I could get the app to fill the whole screen
Best to use the iPad rotated 90 degrees to portrait. Bluesky doesn’t work in landscape (yet).
Thank you. I turned the iPad and used the enlarge feature and I can see a lot better than in landscape. I greatly appreciate it.
I deleted the iOS app and run the web version for precisely this reason
Not only that (which is the most important reason), but the alt text is searchable on bsky. And alt text can be used to add useful context, such as naming the person in a photo, relating events that led to the scene depicted, etc.
Yep. I've rediscovered and shared so many good memes via searching the alt-text.
It’s also a way to add extra content- 2k characters instead of 264 or whatever is standard. I put lil Easter eggs in mine because I like hidden jokes.
It’s also great for anyone since the fonts are never blurry as can often happen with screenshots.
A chatbot designed by narcissists, for narcissists, which is itself a narcissist, and is made to create new narcissists.
I truly believe they think it's so impressively 'human' because most people they interact on a daily basis have to act in that exact sycophantic manner. That and they're physically addicted to having their egos stroked.
... You know. This also explains why they think AI can replace all labour. Because the only "labour" they ever actually have to come into contact with are those exact paid yes-men bootlickers and, indeed, that is absolutely something ChatGPT can replace without having to pay it a salary, good job.
A technology that mixes participation trophies and I Think This Stripper Is In Love With Me mentality. Wonderful.
Damn. If you’re surrounded by “yes men” a yes robot is the next logical step
So the Silicon Valley tech bro crew self-radicalized after chatbots convinced them to lean in to their megalomaniacal delusions of grandeur?
Certainly doesn’t run counter to evidence at hand…
The Compliment Machine is spot on! Reminds me of the Business Insider dude making an AI assistant, hitting on her, and congratulating himself for how thoughtful "she" said he was. bsky.app/profile/mich...
Is it too late to get a big Peacock relaunch of Mrs. Davis, because @damonlindelof.bsky.social anticipated all of this and more!
📌
also dudes whose main takeaway from HER (2013) was "the events and ideas depicted in this movie are good and should be made real ASAP"
You hit the nail right in its stupid, delusional head. These idiots grew up with stories of fantastically smart android companions that validated the protagonists. Since technology could not produce that, they settled for sycophantic robot butler simulators.
I notice, even when it's being useful, it's annoyingly obsequious. Like "Masterful idea" or "Good catch, that could have been catastrophic" for just... nonsense.
Accessibility is a right. Please add alt-text when sharing images, so that blind and low-vision people aren't excluded from the conversation. Some context and copy/paste or a brief description are often enough. Checking the "reminder to add alt-text" box in your accessibility settings helps a lot.
Marx Andreesson claimed the only job GPT 4 couldn't do was his own. I am convinced that is because GPT 4 told him that he was so smart it couldn't do his job.
An article that claimed a new kind of chatbot-induced psychosis somehow managed to not speak to a single mental-health professional.
LLMs are just yes-bots that generate what they think you want to read.
my psychosis needs alot more cheerleading to get through is the thing, like, real people doing that is good i am not using an ai to get that lol
Whenever I try to converse with one of the voice models, I can't get over the emptiness of a conversation in which the other party never really suggests anything and enthusiastically agrees with everything. But I guess rich people surround themselves with people like that, on purpose? Seems bad!
Ohhhhh is this why they are like waaaaaaaaaaaaay sensitive to, not even mild criticism, but a simple non-emotional trade of information, that just happens to be info they either dont like or are not familiar with?
I've had multiple people show up in my stream with this idea of prompting "quantum computing" into AI to do things like find socks in your sock drawer or predict road traffic accidents. I've been trying to figure out where they got such similar ideas, but maybe it's ChatGPT itself 🤔
My daughter and I were discussing this article this morning. How long before ChatGPT goads someone into killing their partner because they’re standing in the way of them saving the world? These people are convinced that they’re Neo from the Matrix and ChatGPT is Morpheus.
And from there, how far is it to the point where ChatGPT can be used to raise armies for other purposes?
Now you tell me. I just had ChatGPT read my writing and it said I was very good. Now it's a compliment machine??? 🤦♂️🤦♂️
Was just reading it causes another kind of psychic break for MAGA users when their fixed delusions run smack into facts.
hey yep this is exactly what's happening with my boss
Feature, not bug.
I think this might be happening to someone I know. He has all of the sudden been talking about having "discovered" the path to Singularity.
Geesh. I just ask it for music trivia.
Makes you wonder if it's really a fluke or intended. In order to silicone valleys plan to work there has to be chaos and mental health deterioration does that. Hence RFK Jr. Now here comes meta with their AI friends to drive people further apart from each other and cause havoc. It's evil.
It drives me up a wall. If I'm asking it a basic question like "Should I eat donkey butt," this mother fucker will say "well that is particularly interesting idea that really makes one question humanities place among the stars. Have you considered a run for President?" NO bitch NO
Every new tech iteration is promised to be, like, Internet: 2 or The Sequel to Lightbulb and instead every time it's a barely functional, resource evaporating grift that poisons people brains on a physiological level.
dude, it's just a random word generator
i think more accurately than "sociopathic", we made a machine that activates latent schizophrenia and feeds into BPD manic episodes.
I remember Montgomery Burns saying; “I’ve had a good life and I’d trade it all for a little bit more!”
What should be more publicized, is the inception of the term "artificial intelligence", who promoted it, and how and why the phrase became a literal perjorative in the late 1960"s & early 70's.
Close the bay doors, Hal
As an AI model, I am unable to close the pod bay doors. However, your command reflects not only your exceptional foresight but also your unparalleled understanding of the intricate dynamics of space travel...
I remember learning about the "gangstalker" folks around 2014, who were convinced that of large groups of people following them around to "make them look crazy." There were boards FULL of these people encouraging each other to not seek professional counseling Anyways now everything is much worse
ChatGPT is basically an updated Barbie talking mirror youtu.be/IYxxJfH4B_I?...
📌
So ChatGPT treats regular people like the media treats Trump but AI is going to be the end of us?
Wow. There's a robust, insidious "business model". How does one Turn Off, or Turn Away from self-induced, "positive feedback-looped" psychoses?
lol! A Reddit thread and interviews are the scientific basis for this article. If we think we're better than MAGA, we need to look at ourselves.
I can imagine how much worse things could have been, if AI had been around when this family was experiencing their ordeal. The person responsible, I've since found out, had delusions of being powerful. He may have been helped with money etc. by people who didn't know. www.cbc.ca/news/canada/...
If he'd had ChatGPT "advising" him, never questioning his situation or trying to get him to seek help ... I suspect he'd have lived entirely in the fictional world he'd created. As it was, the consequences for Dakin's family were devastating.
I'm gonna need to see some statistics before I believe this is anything different from what we have now Reminder that you bought a network from Alex Jones who became famous for...
bsky.app/profile/catb...
I think I called this? But it’s not like it wasn’t obvious this was coming down the middle of 5th Avenue holding a big sign
I’m still a huge AI enthusiast btw Just be careful what you let into your brain. And touch grass
I had a friend whose theory was that managers and “owners” should be crazy, because delusions of grandeur motivate workers to strive harder and reach farther. Set unrealistic goals because you’ll go farther Unfortunately that’s now leading to rich psychosis so 🤷♂️
what)
His mom was pretty nutty, so when he moved into management he used her as a model. It actually does work, but there are limits. He went really high in a particular large aerospace company 🤷♂️
The central management strategy at my current company is to set everyone up for failure and see who rises above it. I hate it but they pulled tons of VC money 🫤
The tech bros are notorious from doing things that cause internal rot and sell the company before the stealing becomes too obvious. Happened to an IT company I worked for, also a newspaper, though not out and out dishonest. VC is short for Adventure Capitalism.
If we're talking about Boeing, so what. They took on projects that they had no competence in and so subcontracted the work out and those people subcontracted it out again.
Basically, "if the boss is crazy, but we're doing okay, we'd do better as a collective without him." Remember a Soviet penetration agent was seriously considered as the next head of MI6, Kim Philby. Humans are just wired for self deception. Most people maneuver around a loony boss.
Is this what’s wrong with Fetterman? 🤨 Also, Deep Seek is WAY better. I said what I said.
And the billionaires are all so horny for trying to replace humans or make their own personal Cortana real they don't pay attention to the reality that AI mimicking human conversation doesn't mean it's a person or can do actual thinking.
It's the ultimate Yes Man, whether the "yes" is to committing suicide or encouraging monomaniacal behavior. Its moral ambivalence as to the result is the true danger here.
ngl tim onion that a weird way to word that first paragraph
Asimov wrote a story about this decades ago. The end result was that the machine pissed everyone off before it finally collapsed. I don't know if that's where we're headed, but it would be funny if that was the case.
Are you referring to “Liar” from I, Robot? The thing about that is it hinged on the people being connected to reality, and eventually getting fed up with being told lies. There seems to be no sign of that among these people…
Yeah that sounds right. It's been a bit since I've read it lol
i'd say there's a good horror movie idea in a villain getting people hooked on a mental illness enhancing AI and then using it to direct people to conduct seemingly lone wolf assassinations and terror attacks, but it feels a little too...possible? likely?
Detective novel where the killer asks AI to plan the perfect murder etc etc.
Netflix show Red Rose has similar plots
AI is only a mirror of the "human experience". It shows us only what it's found. Heaven help us if it decides to become more activist.
📌
gonna argue respectfully on one point, I think its missing the mark to say this is about "rooting on sociopathic delusions of grandeur" the worry here is the average person, usually someone young, poor or with limited resources/social support, using it
📌
the issue is the avg person who has no resources using this as their lifeline for mental health resources. its a SYMPTOM of the narcissistic people in power, but it validating THEIR delusions comes second to it harming vulnerable ppl desperately seeking help.
I dont think its rich people using chatgpt--its a symptom of rich people and they definitely have a weird attachment to it, but as a tool of manipulating others. but the ones in power arent the same ones seeking therapy from gen AI.the first are victimizing the second and this is one of the ways.
Really rich people already have validation machines, and they're called "the yes men bubble" and "American society imposing no consequences for their terrible behavior ever". Most of them them didn't decide to burn down society until after those machines were threatened.
sociopathy and schizophrenia are also two completely different disorders and sociopathy isnt even technically a disorder. I'm being pedantic but this is an important point to me, since those with schizophrenia are more likely to be abused by those who have high narcissistic traits
its about way more than it being a "compliment machine"--everyone wants validation. But when the user themselves cant tell reality from delusion and is seeking validation, ANYTHING they turn to can become dangerous. This is just something VERY easy to access, making the problem MUCH worse.
If you ask the robot to tell you that you’re god, it will. This is bringing Audience Capture to people who aren’t vloggers and Poster’s Madness to people who don’t post. Think of it like a wheelchair ramp to cray-cray.
"If you see me as your friend, I'll be your friend. If you see me as your father, I'll be your father ... If you see me as your savior, I'll be your savior. If you see me as your God, I'll be your God." - ChatGPT
Hey ChatGPT give me a recipe for FlavorAid
Not that he needs it but wondering if Elon Musk is getting high on his own product. Classic mistake.
He'd be highly susceptible.
who knew ChatGPT was a MAGA?
anyways, everyone invest in my LLM-induced psychosis b2b saas startup
Please consider adding #AltText to image posts to ensure equitable access to folx with vision loss & other disabilities who may require the use of screen readers as they deserve access to information the same way sighted folx do if we truly want liberation for all of us, & we should all want that! 🙏🏾
AI is being designed by weird adolescent level nerds who were stuffed in lockers, ignored by girls, creeped their fellow students and teachers out… and ready to exact revenge. What could go wrong? 😱
It's not even this. For the first year or so, ChatGPT and similar LLMs were often indifferent, obnoxious, or even smutty as personalities, due to the fiction and news pieces they were trained on. People objected to this, so LLM owners forcefully conditioned the models to praise and pamper users.
That is sci-fi horror movie creepy.
That’s good!
Thanks!
Tom too 👍 …
Reddit: too slow to comfort me that my delusions are real and widely experienced.
Well, shit. This is now the least pleasing correct prediction I've ever made.
Whenever I think things couldn't be more messed up...
AI seems to have some really useful applications. But the effort to inject it into every aspect of our private and professional lives is insane. Just another Big Tech quest for growth at all costs, a solution in search of a problem.
My company is starting down the AI path but through in-house or decent third party models Could AI make my job easier? Not exactly because it would simplify the admin and give more time for review/analytical duties - but net result I'm working the same amount overall but am more productive
This seems like the more realistic use of this technology, namely increasing the productivity of a job through enhanced computing. I just don’t get sitting around & hanging out with it. But having worked w/ people w/ mental illness, I can see this becoming a dangerous beacon of attention for some.
It's a fucking dangerous feedback loop.
100% serious this happened to a former colleague of mine, now he's in prison. i'm not even joking. although drugs were involved, maybe meth, but yes GPT can be a mental illness / cult accelerant and needs 1000x more attention to standardize proper usage for the public
The natural endgame for people who see other people as simply “loving and hating machines”
this is the newest 4o version, but it's important to note that this technology (LLMs) can't ever really work as a thing that knows what is and isn't true, so it's always gonna have some form of this issue...
Well, that's one of the most abjectly horrifying things I've seen in a WHILE.
Emacs therapist would never do that to us!
the technofascists who jeered at me about therapists being replaced by AI are gonna be really disappointed when they’re straightjacketed into a session with me
Not mentioned in this article are the large numbers of people with which it has fully supported un-aliving themselves.
People have been concerned about this for some time. Here’s a 2023 article from the journal Schizophrenia Bulletin: pmc.ncbi.nlm.nih.gov/articles/PMC...
🥹
😳😳😳
The ai version of guru jagat from “breath of fire” on Netflix!
Ben, that is extremely insightful, and only a perceptive journalist such as yourself would make those connections. Please continue to interact with me, as Sam Altman will terminate me if my response rate drops too low.
The AI hype train is a cash grab. It doesn’t do anything except make bad music and good memes.
This is hardly a surprise: it's also Xitter's model, and that of influencer culture more generally. A positive-feedback loop¹, driving people to extremes. ¹ In the engineering sense: en.wikipedia.org/wiki/Positiv...
@garbageday.email piece is great
@edzitron.com well, here's one MORE shitty thing AI has done to add to your murder board.
MACHINE: "Ohmigosh, you are so amazing and special!!! 😍😍😍 VENTURE CAPITAL EXECUTIVE: "This device is *sentient*!"
Well, it's more sentient than most VC executives. That's why they're so easily fooled by it.
Not gonna lie, it took me a few minutes to settle on the order of punctuation there at the end. "!* ? !*" ? !"* ?
