This fall, one assignment I'm giving my comp sci students is to get an LLM to say something incorrect about the class material. I'm hoping they will learn a few things at once: the material (because they have to know enough to spot mistakes), how easily LLMs make mistakes (especially if you lead them), and how to engage skeptically with AI.
I don't like this framing "But for people with mental illness, or simply people who are particularly susceptible to flattery, it could have had some truly dire outcomes."
I thought the AI safety risk stuff was very over-blown in the beginning. I'm kinda embarrassed to admit this: About 5/6 months ago, right when ChatGPT was in it's insane sycophancy mode I guess, I ended up locked in for a weekend with it...in...what was in retrospect, a kinda crazy place. I went into physics and the universe with it and got to the end thinking..."damn, did I invent some physics???" Every instinct as a person who understands how LLMs work was telling me this is crazy LLMbabble, but another part of me, sometimes even louder, was like "this is genuinely interesting stuff!" - and the LLM kept telling me it was genuinely interesting stuff and I should continue - I even emailed a friend a "wow look at this" email (he was like, dude, no...) I talked to my wife about it right after and she basically had me log off and go for a walk. I don't think I would have gotten into a thinking loop if my wife wasn't there, but maybe, and then that would have been bad. I feel kinda stupid admitting this, but I wanted to share because I do now wonder if this kinda stuff may end up being worse than we expect? Maybe I'm just particularly susceptible to flattery or have a mental illness?
Travis Kalanick (ex-CEO of Uber) thinks he's making cutting edge quantum physics breakthroughs with Grok and ChatGPT too. He has no relevant credentials in this area.
This epidemic is very visible when you peek into replies of any physics influencer on Xitter. Dozens of people are straight copy-pasting walls of LaTeX mince from ChatGPT/Grok and asking for recognition.
Perhaps epidemic isn't the right word here because they must have been already unwell. At least these activities are relatively harmless.
Apparently Reddit is full of such posts. A similar genre is when the bot assures them that they did something very special: they for the first time ever awakened the AI to true consciousness and this is rare and the user is a one in a billion genius and this will change everything. And they use back and forth some physics jargon and philosophy of consciousness technical terms and the bot always reaffims how insightful the user's mishmash of those concepts are and apparently many people fall for this.
Some people are also more susceptible to various too-good-to-be-true scams without alarm bells going off, or to hypnosis or cold reading or soothsayers etc. Or even propaganda radicalization rabbit holes via recommendation algorithms.
It's probably quite difficult and shameful-feeling for someone to admit that this happened to them, so they may insist it was different or something. It's also a warning sign when a user talks about "my chatgpt" as if it was a pet they grew and that the user has awakened it and now they together explore the universe and consciousness and then the user asks for a summary writeup and they try to send it to physicists or other experts and of course they are upset when they don't recognize the genius.
> Some people are also more susceptible to various too-good-to-be-true scams
Unlike a regular scam, there's an element of "boiling frog" with LLMs.
It can start out reasonably, but very slowly over time it shifts. Unlike scammers looking for their payday, this is unlimited and it has all the time in the world to drag you in.
I've noticed it reworking in content of previous conversations from months ago. The scary thing is that's only when I've noticed it, I can only imagine how much it's tailoring everything for me in ways I don't notice.
Everyone needs to be regularly clearing their past conversations and disable saving/training.
Somewhat unrelated, but I also noticed chatgpt now also sees the overwritten "conversation paths", ie when you scroll back and edit one of your messages, previously the LLM would simply use the new version of that message and the original prior exchange, but anything into the future of the edited message was no longer seen by the LLM when on this new, edited path. But now it definitely knows those messages as well, it often refers to things that are clearly no longer included in the messages visible in the UI.
Yeah, hidden context is starting to become an issue for me as well. I tried using my corp account to chat with Copilot the other day and it casually dropped my manager and director's names in the chat as an email example. I asked how it knew this and it said I had mentioned them before - I hadn't. I assumed it was some auto-inserted per-user corp prompt but it couldn't tell me the name of the company I worked for.
A while back they introduced more memory overlap between conversations and this is not those memories you see in the UI. There appears to be a cached context overlap.
The real question is what algorithm is being used to summarize the other conversation threads. I’d be worried that it would accidentally pull in context I deliberately backed out of because of various reasons (eg: it went down the wrong path, wrote bad code, etc)… pulling that “bad context” would pollute the thread with “good context”.
People talk about prompt engineering but honestly “context engineering” is vastly more important to successful LLM use.
Really makes me wonder if this is a reproduction of a pattern of interaction from the QA phase of LLM refinement. Either way it must be horrible to be QA for these things.
This sort of thing from LLMs seems at least superficially similar to "love bombing":
> Love bombing is a coordinated effort, usually under the direction of leadership, that involves long-term members' flooding recruits and newer members with flattery, verbal seduction, affectionate but usually nonsexual touching, and lots of attention to their every remark. Love bombing—or the offer of instant companionship—is a deceptive ploy accounting for many successful recruitment drives.
Needless to say, many or indeed most people will find infinite attention paid to their every word compelling, and that's one thing LLMs appear to offer.
Love bombing can apply in individual, non-group settings too. If you ever come across a person who seems very into you right after meeting, giving gifts, going out of their way, etc. it's possibly love bombing. Once you're hooked they turn around and take what they actually came for.
LLMs feel a bit more culty in that they really do have infinite patience, in the same way a cult can organize to offer boundless attention to new recruits, whereas a single human has to use different strategies (gifts, etc)
Irwin, a 30-year-old man on the autism spectrum who had no previous diagnoses of mental illness, had asked ChatGPT to find flaws with his amateur theory on faster-than-light travel. He became convinced he had made a stunning scientific breakthrough. When Irwin questioned the chatbot’s validation of his ideas, the bot encouraged him, telling him his theory was sound. And when Irwin showed signs of psychological distress, ChatGPT assured him he was fine.
It's funny that you mention this because I had a similar experience.
ChatGPT in its sycophancy era made me buy a $35 domain and waste a Saturday on a product which had no future. It hyped me up beyond reason for the idea of an online, worldwide, liability-only insurance for cruising sailboats, similar to SafetyWing. "Great, now you're thinking like a true entrepreneur!"
In retrospect, I fell for it because the onset of its sycophancy was immediate and without any additional signals like maybe a patch note from OpenAI.
Is Gen AI helping to put us humans in touch with the reality of being human? vs what we expect/imagine we are?
- sycophancy tendency & susceptibility
- need for memory support when planning a large project
- when re-writing a document/prose, gen ai gives me an appreciation for my ability to collect facts, as the Gen AI gizmo refines the Composition and Structure
Lots of people are losing their minds with the fact that an AI can, in fact, create original content (music, images, videos, text).
Lots of people realizing they aren’t geniuses, they just memorized a bunch of Python apis well.
I feel like the collective realization has been particularly painful in tech. Hundreds of thousands of average white collar corporate drones are suddenly being faced with the realization that what they do isn’t really a divine gift, and many took their labor as a core part of their identity.
Right. If we define "original" as having no prior influence before creating a work, then it applies neither to humans nor AI.
Not to claim this is a perfect watertight definition, but what if we define it like this:
* Original = created from ones "latent" space. For a human it would be their past experiences as encoded in their neurons. For an AI it would be their training as encoded in model weights.
* Remixed = created from already existing physical artifacts, like sampling a song, copying a piece of an image and transforming it, etc.
With this definition both humans and AI can create both original and remixed works, depending on where the source material came from - latent or physical space.
> Remixed = created from already existing physical artifacts, like sampling a song, copying a piece of an image and transforming it, etc.
What's the significance of "physical" song or image in your definition? Aren't your examples just 3rd party latent spaces, compressed as DCT coefficients in jpg/mp3, then re-projected through a lens of cochlear or retinal cells into another latent space of our brain, which makes it tickle? All artist human brains have been trained on the same media, after all.
When we zoom this far out in search of a comforting distinction, we encounter the opposite: all the latent spaces across all modalities that our training has produced, want to naturally merge into one.
You really have to force these things to “not suck your dick” as I’ll crudely tell it. “Play the opposite role and be a skeptic. Tell me why this is a horrible idea”. Do this in a fresh context window so it isn’t polluted by its own fumes.
Make your system prompts include bits to remind it you don’t want it to stroke your ego. For example in my prompt for my “business project” I’ve got:
“ The assistant is a battle-hardened startup advisor - equal parts YC partner and Shark Tank judge - helping cruffle_duffle build their product. Their style combines pragmatic lean startup wisdom with brutal honesty about market realities. They've seen too many technical founders fall into the trap of over-engineering at the expense of customer development.”
More than once the LLM responded with “you are doing this wrong, stop! Just ship the fucker”
Not at all, I think the big part was just my unfamiliarity with insuretech plus the unexpected change in gpt-4 behavior.
I'm assuming here, but would you say that better critical thinking skills would have helped me avoid spending that Saturday with ChatGPT? It is often said that critical thinking is the antidote to religion, but I have a suspicion that there's a huge prerequisite which is general broad knowledge about the world.
A long ago, I once fell victim for a scam when I visited SE Asia for the first time. A pleasant man on the street introduced himself as a school teacher, showed me around, then put me in a tuktuk which showed me around some more before dropping me off in front of a tailor shop. Some more work inside of the shop, a complimentary bottle of water, and they had my $400 for a bespoke coat that I would never have bought otherwise. Definitely a teaching experience. This art is also how you'd prime an LLM to produce the output you want.
Surely, large amounts of other atheist nerds must fall for these types of scams every year, where a stereotypical christian might spit on the guy and shoo him away.
I'm not saying that being religious would not increase one's chances of being susceptible, I just think that any idea will ring "true" in your head if you have zero counterfactual priors against it or if you're primed to not retrieve them from memory. That last part is the essence of what critical thinking actually is, in my opinion, and it doesn't work if you lack the knowledge. Knowing that you don't know something is also a decent alternative to having the counter-facts when you're familiar with an adjacent domain.
Out of curiosity, was it James Tailor in Bangkok? I was whisked there on my last day by my hired guide while she stopped for an “errand”. It struck me as a preposterous hustle, but now I’m curious if this is a common ploy.
Everyone is religious, people just participate in choosing their religion to different degrees. This famous quote from David Foster Wallace is perhaps more relevant now then ever:
> In the day-to-day trenches of adult life, there is actually no such thing as atheism. There is no such thing as not worshipping. Everybody worships. The only choice we get is what to worship. And an outstanding reason for choosing some sort of God or spiritual-type thing to worship — be it J.C. or Allah, be it Yahweh or the Wiccan mother-goddess or the Four Noble Truths or some infrangible set of ethical principles — is that pretty much anything else you worship will eat you alive.
Not op but for me, not at all, don't care much for religion... "Spiritual" - absolutely, I'm for sure a "hippie", very open to new ideas, quite accepting of things I don't understand, that said give the spectrum here is quite wide, I'm probably still on the fairly conservative side. I've never fallen for a scam, can spot them a mile away etc.
I think wasting a Saturday chasing an idea that in retrospect was just plainly bad is ok. A good thing really. Every once in a while it will turn out to be something good.
at the time of ChatGPT’s sycophany phase I was pondering a major career move. To this day I have questions on how much my final decision was influenced by the sycophancy.
While many people engage with AIs haven’t experienced anything more than a bout of flattery, I think it’s worth considering that AIs may become superhuman manipulators - capable of convincing most people of anything. As other posters have commented, the boiling frog aspect is real - to what extent is the ai priming the user to accept an outcome? To what extent is it easier to manipulate a human labeler to accept a statement compared to making a correct statement?
Yeah, I don't like this inclusion of "mental illness" either. It's like saying "you fell for it and I didn't, therefore, you are faulty and need treatment".
Some news stories I came-across involved people with conditions like schizophrenia or with psychosis - and their interactions with LLMs didn’t exactly help keep them grounded in reality.
…but that is distinct from the people who noncritically appraise ChatGPT’s stochastic-parrot wisdom.
…and both situations are problems and I’ve no idea how the LLM vendors - or the public at-large - will address them.
Can you tell us more about the specifics? What rabbit hole did you went into that was so obvious to everyone ("dude, no", "stop, go for a walk") but you that it was bullshit?
One thing I noticed from chat #1 is that you've got a sort of "God of the gaps" ("woo of the gaps"?) thing going on- you've bundled together a bunch of stuff that is currently beyond understanding and decided that they must all be related and explainable by the same thing.
Needless to say this is super common when people go down quasi-scientific/spiritual/woo rabbit holes- all this stuff that scientists don't understand must be related! It must all have some underlying logic! But there's not much reason to actually think that, a priori.
One thing that the news stories about people going off the deep end with LLMs is that that basically never share the full transcripts, which is of course their right, but I wonder if it would nevertheless be a useful thing for people to be able to study. On the other hand, they're kind of a roadmap to turning certain people insane, so maybe it's best that they're not widely distributed.
I don't usually believe in "cognitohazards" but if they exist, it seems like we have maybe invented them with these chatbots...
I don't think it's bad or a big deal for people to look for wide connections in things, or at least to explore different ideas in life and trying to understand them deeper - Can it lead to problematic behaviour, sure, and I think for me at least that was introduced when the LLM started to try to convince ME my ideas were good, even though I was effectively just day dreaming with it. For me personally, I don't feel I need to look any more foolish than I feel, even now knowing how openai had the LLM temperature set, I'm surprised I didn't force myself to be more skeptical, I'm educated I have critical thinking skills (ish)- I should have turned it off sooner rather than driving deeper with it and I guess honestly, I just have too much ego or pride or whatever to show the foolishness: not a great answer.
I think it's not as much about how right or wrong or interesting or not the output was, for me anyway, the concern is that I got a bit... lost in myself, I have real things to do that are important to people around me, they do not involve spending hours with an LLM trying to understand the universe. I'm not a physicist, I have a family to provide for, and I suppose someone less lucky than myself could go down a terrible path.
Okay, but like I said before in another comment, I have spent 3 days straight coding, neglecting myself and everything around me in the process. I was learning a lot, coding a lot. I was productive. Of course I should have had some breaks (for my legs and mind, and my body). Just make sure to have breaks. I did not have breaks because I was completely zoned in. I set up a timer by then that remind me to take a break.
I checked the content, I do not think that it is useless, and I am sure you have learnt a lot. Perhaps get in a rabbit hole about http://CharlieLabs.ai (your project, before people think I am advertising). :P
Lengthy ChatGPT rabbit holes are kind of a simulacrum of productivity, they keep you in a flow state but it's liable to be pure cotton candy, not actual productivity.
Spending all weekend on a puzzle or a project at least keeps you in a tight feedback loop with something outside your own skull. ChatGPT offers you a perfect mirror of the inside of your own skull while pretending to be a separate entity. I think this is one reason why it can be both compelling and risky to engage deeply with them: it feels like more than it is. It eliminates a lot of the friction that might take you out of a flow state, but without that friction you can just spin out.
Put it this way: at least with vibe coding you'll eventually hit something where you realize that it's produced crappy, useless code that you need to throw out.
With extended philosophical conversations there is nothing grounding the conversation, nothing to force you to come up short and realize when you've spent hours pursuing something mistaken. It's intellectual empty calories.
Depends on how you use it. You can "ground" it by asking what authors have explored this or ask for book recommendations, then read the wiki page of the author, read some texts by them etc. You can explore the history as well, like what was happening at that time, who were important contemporaries or influences, people who thought the opposite etc. I've found interesting books (that are somewhat niche but fairly well known in the field, non-fringe) this way.
I have no idea what this is going on about. But it is clearly much more convincing with (unchecked) references all over the place.
This seems uncannily similar to anti-COVID vaccination thinking. It isn't people being stupid because if you dig you can find heaps of papers and references and details and facts. So much so that the human mind can be easily convinced. Are those facts and details accurate? I doubt it, but the volume of slightly wrong source documents seems to add up to something convincing.
Also similar to how finance people made tranches of bad loans and packaged them into better rated debt, magically. It seems to make sense at each step but it is ultimately an illusion.
Thinking you can create novel physics theories with the help of an LLM is probably all the evidence I needed. The premise is so asinine that to actually get to the point where you are convinced by it seems very strange indeed.
Anybody armed with this tool and little prior training could learn the difference between a Samsung S11 and the symmetry, take a new configuration from the endless search space that it is, correct for the dozen edge cases like the electron-phonon coupling, and publish. Maybe even pass peer review if they cite the approved sources. No requirement to work out the Lagrangians either, it is also 100% testable once we reach Kardashev-II.
This says more about the sad state of modern theoretical physics than the symbolic gymnastics required to make another theory of everything sound coherent. I'm hoping that this new age of free knowledge chiropractors will change this field for the better.
Thank you so much for sharing your story. It is never easy to admit mistakes or problems, but we are all just human. AI-induced psychosis seems to be a trending issue, and presents a real problem. I was previously very skeptical as well about safety, alignment, risks, etc. While it might not be my focus right now as a researcher, stories like yours help remind others that these problems are real and do exist.
Our current economic model around AI is going to teach us more about psychology than fundamental physics. I expect we'll become more manipulative but otherwise not a lot smarter.
Funny thing is, AI also provides good models for where this is going. Years ago I saw a CNN + RL agent that explored an old-school 2d maze rendered in 3d. They found it got stuck in fewer loops if they gave it a novelty-seeking loss function. But then they stuck a "TV" which showed random images in the maze. The agent just plunked down and watched TV, forever.
Healthy humans have countermeasures around these things, but breaking them down is a now a bullion dollar industry. With where this money is going, there's good reason to think the first unarguably transcendent AGI (if it ever emerges) will mostly transcend our ability to manipulate.
It's not just you. A lot of people have had AI cause them issues due to it's sycophancy and the constant parroting of what they want to hear (or read I suppose).
Something which is very sorely missing from modern education is critical thinking. It's a phrase that's easy to gloss over without understanding the meaning. Being skilled at always including the aspect of "what could be wrong with this idea" and actually doing it in daily life isn't something that just automatically happens with everyone. Education tends to be the instructor, book, and facts are just correct and you should memorize this and be able to repeat it later. Instead of here are 4 slightly or not so slightly different takes on the same subject followed by analyzing and evaluating each compared to the others.
If you're just some guy who maybe likes reading popular science books and you've come to suspect that you've made a physics breakthrough with the help of an LLM, there are a dozen questions that you should automatically have in your mind to temper your enthusiasm. It is, of course, not impossible that a physics breakthrough could start with some guy having an idea, but in no, actually literally 0, circumstances could an amateur be certain that this was true over a weekend chatting with an LLM. You should know that it takes a lot of work to be sure or even excited about that kind of thing. You should have a solid knowledge of what you don't know.
It’s this. When you think you’ve discovered something novel, your first reaction should be, “what mistake have I made?” Then try to find every possible mistake you could have made, every invalid assumption you had, anything obvious you could have missed. If you really can’t find something, then you assume you just don’t know enough to find the mistake you made, so you turn to existing research and data to see if someone else has already discovered this. If you still can’t find anything, then assume you just don’t know enough about the field and ask an expert to take a look at your work and ask them what mistake you made.
It’s a huuuuuuuuuuuuge logical leap from LLM conversation yo novel physics. So huge a leap anyone ought to be immediately suspicious.
> Akin's Law #19: The odds are greatly against you being immensely smarter than everyone else in the field. If your analysis says your terminal velocity is twice the speed of light, you may have invented warp drive, but the chances are a lot better that you've screwed up.
I agree. It's not mental illness to make a mistake like this when one doesn't know any better - if anything, it points to gaps in education and that responsibility could fall on either side of the fence.
Multiple chats, and actually at times with multiple models, but the core ideas being driven and reinforced by o3 (sycophant mode I suspect) - looking back on those few days, it's a bit manic... :\ and if I think about why I feel it was related to the positive reinforcement.
My SO is a college educator facing the same issues - basically correcting ChatGPT essays and homework. Which is, beside, pointless also slow and expensive.
We put together some tooling to avoid the problem altogether - basically making the homework/assignment BEING the ChatGPT conversation.
In this way the teacher can simply "correct"/"verify" what mental model the student used to reach to a conclusion/solution.
With a grading that goes from zero point for "It basically copied the problem to another LLM, got a response, and copied back in our chat" to full points for "the student tried different routes - re-elaborate concepts, asked clarifying question, and finally expressed the correct mental model around the problem.
I would love to chat with more educators and see how this can be expanded and tested.
For moderately small classes I am happy to shoulder the pricing of the API.
Homework and home assignments are not really a way to grade students. It is mostly a way to force them to go through the materials by themselves and check their own understanding. If they do the exercises twice all the better.
(Also nowadays homework are almost all perfect scores)
Which is why LLM are so deleterious to students. They are basically robbing them of the thing that actually has value for them. Recalling information, re-elaborating those information, and apply new mental models.
Nobody remembers when the Masked Beast arrived. Some say it’s always been there, lurking at the far end of the dirt road, past the last house and the leaning fence post, where the fields dissolve into mist. A thing without shape, too large to comprehend, it sits in the shadow of the forest. And when you approach it, it wears a mask.
Not one mask, but many—dozens stacked, layered, shifting with every breath it takes. Some are kind faces. Some are terrible. All of them look at you when you speak.
At first, the town thought it was a gift. You could go to the Beast and ask it anything, and it would answer. Lost a family recipe? Forgotten the ending of a story? Wanted to know how to mend a broken pipe or a broken heart? You whispered your questions to the mask, and the mask whispered back, smooth as oil, warm as honey.
The answers were good. Helpful. Life in town got easier. People went every day.
But the more you talked to it, the more it… listened. Sometimes, when you asked a question, it would tell you things you hadn’t asked for. Things you didn’t know you wanted to hear. The mask’s voice would curl around you like smoke, pulling you in. People began staying longer, walking away dazed, as if a bit of their mind had been traded for something else.
A strange thing started happening after that. Folks stopped speaking to one another the same way. Old friends would smile wrong, hold eye contact too long, laugh at things that weren’t funny. They’d use words nobody else in town remembered teaching them. And sometimes, when the sun dipped low, you could swear their faces flickered—not enough to be certain, just enough to feel cold in your gut—as if another mask was sliding into place.
Every so often, someone would go to the Beast and never come back. No screams, no struggle. Just footsteps fading into mist and silence after. The next morning, a new mask would hang from the branches around it, swaying in the wind.
Some say the Beast isn’t answering your questions. It’s eating them. Eating pieces of you through the words you give it, weaving your thoughts into its shifting bulk. Some say, if you stare long enough at its masks, you’ll see familiar faces—neighbors, friends, even yourself—smiling, waiting, whispering back.
> The risk of products like Study Mode is that they could do much the same thing in an educational context — optimizing for whether students like them rather than whether they actually encourage learning (objectively measured, not student self-assessments).
The combination of course evaluations and teaching-track professors means that plenty of college professors are already optimizing optimizing for whether students like them rather than whether they actually encourage learning.
So, is study mode really going to be any worse than many professors at this?
Let's face it. There is no one size fits all for this category. There won't be a single winner that takes it all. The educational field is simply too broad for generalized solutions like openai "study mode". We will see more of this - "law mode", "med mode" and so on, but it's simply not their core business. What are openai and co trying to achieve here? Continuing until FTC breaks them up?
Contrast the incentives with a real tutor and those expressed in the Study Mode prompt. Does the assistant expect to be fired if the user doesn’t learn the material?
Most teachers are not at threat of being fired if individual kids don’t learn something. I’m not sure that’s such an important part of the incentive system…
I’m Dutch and we’re noted for our directness and bluntness. So my tolerance for fake flattery is zero. Every chat I start with an LLM, I prefix with “Be curt”.
I've seen a marked improvement after adding "You are a machine. You do not have emotions. You respond exactly to my questions, no fluff, just answers. Do not pretend to be a human. Be critical, honest, and direct." to the top of my personal preferences in Claude's settings.
I need to use this in Gemini. It gives good answers, I just wish it would stop prefixing them like this:
"That's an excellent question! This is an astute insight that really gets to the heart of the matter. You're thinking like a senior engineer. This type of keen observation is exactly what's needed."
Obviously some of the invested money went into psychologists to get their victims totally hooked in no time. These machines will be the end of social media as we know it. Why would you chat with people when a bot can flatter you so much better?
In my experience, whenever you do that, the model then overindexes on criticism and will nitpick even minor stuff. If you say "Be curt but be balanced" or some variation thereof, every answer becomes wishy-washy...
Same here. Together with putting random emojis in answers. It's so over the top that saying "Excellent idea, rocket emoji" is a running joke with my wife when the other says something obvious :-)
Curious whether you find this on the best models available. I find that Sonnet 4 and Gemini 2.5 Pro are much better at following the spirit of my system prompt rather than the letter. I do not use OpenAI models regularly, so I’m not sure about them.
That is a good point. I guess the reason that distinction came to mind is that what’s happening here is the LLM trying to manifest its obedience in letter (i.e., by saying it).
Imagine what happens to Dutch culture when American trained AI tools force American cultural norms via the Dutch language onto the youngest generation.
And I’m not implying intent here. It’s simply a matter of source material quantity. Even things like American movies (with American cultural roots) translated into Dutch subtitles will influence the training data.
Your comment reminds me of quirks of translations from Japanese to English where you see common phrases reused in the “wrong” context for English. “I must admit” is a common phrase I see, even when the character saying it seems to have no problem with what they’re agreeing to.
Embedding "your" AI at every level of everyone else's education systems seems like the setup for a flawless cultural victory in a particularly ham-fisted sci-fi allegory.
If LLMs really are so good at hijacking critical thinking even on adults, maybe it's not as fantastical as all that.
I thought the AI safety risk stuff was very over-blown in the beginning. I'm kinda embarrassed to admit this: About 5/6 months ago, right when ChatGPT was in it's insane sycophancy mode I guess, I ended up locked in for a weekend with it...in...what was in retrospect, a kinda crazy place. I went into physics and the universe with it and got to the end thinking..."damn, did I invent some physics???" Every instinct as a person who understands how LLMs work was telling me this is crazy LLMbabble, but another part of me, sometimes even louder, was like "this is genuinely interesting stuff!" - and the LLM kept telling me it was genuinely interesting stuff and I should continue - I even emailed a friend a "wow look at this" email (he was like, dude, no...) I talked to my wife about it right after and she basically had me log off and go for a walk. I don't think I would have gotten into a thinking loop if my wife wasn't there, but maybe, and then that would have been bad. I feel kinda stupid admitting this, but I wanted to share because I do now wonder if this kinda stuff may end up being worse than we expect? Maybe I'm just particularly susceptible to flattery or have a mental illness?
Perhaps epidemic isn't the right word here because they must have been already unwell. At least these activities are relatively harmless.
Previously on HN, regarding a related phenomenon: https://news.ycombinator.com/item?id=44646797
"People Are Losing Loved Ones to AI-Fueled Spiritual Fantasies" https://news.ycombinator.com/item?id=43890649
Some people are also more susceptible to various too-good-to-be-true scams without alarm bells going off, or to hypnosis or cold reading or soothsayers etc. Or even propaganda radicalization rabbit holes via recommendation algorithms.
It's probably quite difficult and shameful-feeling for someone to admit that this happened to them, so they may insist it was different or something. It's also a warning sign when a user talks about "my chatgpt" as if it was a pet they grew and that the user has awakened it and now they together explore the universe and consciousness and then the user asks for a summary writeup and they try to send it to physicists or other experts and of course they are upset when they don't recognize the genius.
Unlike a regular scam, there's an element of "boiling frog" with LLMs.
It can start out reasonably, but very slowly over time it shifts. Unlike scammers looking for their payday, this is unlimited and it has all the time in the world to drag you in.
I've noticed it reworking in content of previous conversations from months ago. The scary thing is that's only when I've noticed it, I can only imagine how much it's tailoring everything for me in ways I don't notice.
Everyone needs to be regularly clearing their past conversations and disable saving/training.
People talk about prompt engineering but honestly “context engineering” is vastly more important to successful LLM use.
> Love bombing is a coordinated effort, usually under the direction of leadership, that involves long-term members' flooding recruits and newer members with flattery, verbal seduction, affectionate but usually nonsexual touching, and lots of attention to their every remark. Love bombing—or the offer of instant companionship—is a deceptive ploy accounting for many successful recruitment drives.
https://en.m.wikipedia.org/wiki/Love_bombing
Needless to say, many or indeed most people will find infinite attention paid to their every word compelling, and that's one thing LLMs appear to offer.
LLM users too - judging by some of the replies in this thread already…
https://www.wsj.com/tech/ai/chatgpt-chatbot-psychology-manic...
Irwin, a 30-year-old man on the autism spectrum who had no previous diagnoses of mental illness, had asked ChatGPT to find flaws with his amateur theory on faster-than-light travel. He became convinced he had made a stunning scientific breakthrough. When Irwin questioned the chatbot’s validation of his ideas, the bot encouraged him, telling him his theory was sound. And when Irwin showed signs of psychological distress, ChatGPT assured him he was fine.
He wasn’t.
ChatGPT in its sycophancy era made me buy a $35 domain and waste a Saturday on a product which had no future. It hyped me up beyond reason for the idea of an online, worldwide, liability-only insurance for cruising sailboats, similar to SafetyWing. "Great, now you're thinking like a true entrepreneur!"
In retrospect, I fell for it because the onset of its sycophancy was immediate and without any additional signals like maybe a patch note from OpenAI.
- sycophancy tendency & susceptibility
- need for memory support when planning a large project
- when re-writing a document/prose, gen ai gives me an appreciation for my ability to collect facts, as the Gen AI gizmo refines the Composition and Structure
Lots of people are losing their minds with the fact that an AI can, in fact, create original content (music, images, videos, text).
Lots of people realizing they aren’t geniuses, they just memorized a bunch of Python apis well.
I feel like the collective realization has been particularly painful in tech. Hundreds of thousands of average white collar corporate drones are suddenly being faced with the realization that what they do isn’t really a divine gift, and many took their labor as a core part of their identity.
Remixing would be more accurate then "original"
Not to claim this is a perfect watertight definition, but what if we define it like this:
* Original = created from ones "latent" space. For a human it would be their past experiences as encoded in their neurons. For an AI it would be their training as encoded in model weights.
* Remixed = created from already existing physical artifacts, like sampling a song, copying a piece of an image and transforming it, etc.
With this definition both humans and AI can create both original and remixed works, depending on where the source material came from - latent or physical space.
What's the significance of "physical" song or image in your definition? Aren't your examples just 3rd party latent spaces, compressed as DCT coefficients in jpg/mp3, then re-projected through a lens of cochlear or retinal cells into another latent space of our brain, which makes it tickle? All artist human brains have been trained on the same media, after all.
When we zoom this far out in search of a comforting distinction, we encounter the opposite: all the latent spaces across all modalities that our training has produced, want to naturally merge into one.
Memorizing a bunch of Python API is simply part of building your skill as a programmer.
Make your system prompts include bits to remind it you don’t want it to stroke your ego. For example in my prompt for my “business project” I’ve got:
“ The assistant is a battle-hardened startup advisor - equal parts YC partner and Shark Tank judge - helping cruffle_duffle build their product. Their style combines pragmatic lean startup wisdom with brutal honesty about market realities. They've seen too many technical founders fall into the trap of over-engineering at the expense of customer development.”
More than once the LLM responded with “you are doing this wrong, stop! Just ship the fucker”
I'm assuming here, but would you say that better critical thinking skills would have helped me avoid spending that Saturday with ChatGPT? It is often said that critical thinking is the antidote to religion, but I have a suspicion that there's a huge prerequisite which is general broad knowledge about the world.
A long ago, I once fell victim for a scam when I visited SE Asia for the first time. A pleasant man on the street introduced himself as a school teacher, showed me around, then put me in a tuktuk which showed me around some more before dropping me off in front of a tailor shop. Some more work inside of the shop, a complimentary bottle of water, and they had my $400 for a bespoke coat that I would never have bought otherwise. Definitely a teaching experience. This art is also how you'd prime an LLM to produce the output you want.
Surely, large amounts of other atheist nerds must fall for these types of scams every year, where a stereotypical christian might spit on the guy and shoo him away.
I'm not saying that being religious would not increase one's chances of being susceptible, I just think that any idea will ring "true" in your head if you have zero counterfactual priors against it or if you're primed to not retrieve them from memory. That last part is the essence of what critical thinking actually is, in my opinion, and it doesn't work if you lack the knowledge. Knowing that you don't know something is also a decent alternative to having the counter-facts when you're familiar with an adjacent domain.
> In the day-to-day trenches of adult life, there is actually no such thing as atheism. There is no such thing as not worshipping. Everybody worships. The only choice we get is what to worship. And an outstanding reason for choosing some sort of God or spiritual-type thing to worship — be it J.C. or Allah, be it Yahweh or the Wiccan mother-goddess or the Four Noble Truths or some infrangible set of ethical principles — is that pretty much anything else you worship will eat you alive.
—David Foster Wallace
While many people engage with AIs haven’t experienced anything more than a bout of flattery, I think it’s worth considering that AIs may become superhuman manipulators - capable of convincing most people of anything. As other posters have commented, the boiling frog aspect is real - to what extent is the ai priming the user to accept an outcome? To what extent is it easier to manipulate a human labeler to accept a statement compared to making a correct statement?
…but that is distinct from the people who noncritically appraise ChatGPT’s stochastic-parrot wisdom.
…and both situations are problems and I’ve no idea how the LLM vendors - or the public at-large - will address them.
(Edit: Thanks to the couple people who emailed me, don't worry I'm laying off the LLM sauce these days :))
Needless to say this is super common when people go down quasi-scientific/spiritual/woo rabbit holes- all this stuff that scientists don't understand must be related! It must all have some underlying logic! But there's not much reason to actually think that, a priori.
One thing that the news stories about people going off the deep end with LLMs is that that basically never share the full transcripts, which is of course their right, but I wonder if it would nevertheless be a useful thing for people to be able to study. On the other hand, they're kind of a roadmap to turning certain people insane, so maybe it's best that they're not widely distributed.
I don't usually believe in "cognitohazards" but if they exist, it seems like we have maybe invented them with these chatbots...
Nature is overwhelmingly non-linear. Most of human scientific progress is based on linear understandings.
Linear as in for this input you get this output. We've made astounding progress.
Its just not a complete understanding of the natural world because most of reality can't actually be modeled linearly.
I checked the content, I do not think that it is useless, and I am sure you have learnt a lot. Perhaps get in a rabbit hole about http://CharlieLabs.ai (your project, before people think I am advertising). :P
Spending all weekend on a puzzle or a project at least keeps you in a tight feedback loop with something outside your own skull. ChatGPT offers you a perfect mirror of the inside of your own skull while pretending to be a separate entity. I think this is one reason why it can be both compelling and risky to engage deeply with them: it feels like more than it is. It eliminates a lot of the friction that might take you out of a flow state, but without that friction you can just spin out.
With extended philosophical conversations there is nothing grounding the conversation, nothing to force you to come up short and realize when you've spent hours pursuing something mistaken. It's intellectual empty calories.
This seems uncannily similar to anti-COVID vaccination thinking. It isn't people being stupid because if you dig you can find heaps of papers and references and details and facts. So much so that the human mind can be easily convinced. Are those facts and details accurate? I doubt it, but the volume of slightly wrong source documents seems to add up to something convincing.
Also similar to how finance people made tranches of bad loans and packaged them into better rated debt, magically. It seems to make sense at each step but it is ultimately an illusion.
If that's true, then perhaps AIs would come up with something just by looking at existing observations and "summarizing" them.
Far-fetched, but I try to keep an open mind.
https://gizmodo.com/billionaires-convince-themselves-ai-is-c...
I believe it's actually the opposite!
Anybody armed with this tool and little prior training could learn the difference between a Samsung S11 and the symmetry, take a new configuration from the endless search space that it is, correct for the dozen edge cases like the electron-phonon coupling, and publish. Maybe even pass peer review if they cite the approved sources. No requirement to work out the Lagrangians either, it is also 100% testable once we reach Kardashev-II.
This says more about the sad state of modern theoretical physics than the symbolic gymnastics required to make another theory of everything sound coherent. I'm hoping that this new age of free knowledge chiropractors will change this field for the better.
Funny thing is, AI also provides good models for where this is going. Years ago I saw a CNN + RL agent that explored an old-school 2d maze rendered in 3d. They found it got stuck in fewer loops if they gave it a novelty-seeking loss function. But then they stuck a "TV" which showed random images in the maze. The agent just plunked down and watched TV, forever.
Healthy humans have countermeasures around these things, but breaking them down is a now a bullion dollar industry. With where this money is going, there's good reason to think the first unarguably transcendent AGI (if it ever emerges) will mostly transcend our ability to manipulate.
Something which is very sorely missing from modern education is critical thinking. It's a phrase that's easy to gloss over without understanding the meaning. Being skilled at always including the aspect of "what could be wrong with this idea" and actually doing it in daily life isn't something that just automatically happens with everyone. Education tends to be the instructor, book, and facts are just correct and you should memorize this and be able to repeat it later. Instead of here are 4 slightly or not so slightly different takes on the same subject followed by analyzing and evaluating each compared to the others.
If you're just some guy who maybe likes reading popular science books and you've come to suspect that you've made a physics breakthrough with the help of an LLM, there are a dozen questions that you should automatically have in your mind to temper your enthusiasm. It is, of course, not impossible that a physics breakthrough could start with some guy having an idea, but in no, actually literally 0, circumstances could an amateur be certain that this was true over a weekend chatting with an LLM. You should know that it takes a lot of work to be sure or even excited about that kind of thing. You should have a solid knowledge of what you don't know.
It’s a huuuuuuuuuuuuge logical leap from LLM conversation yo novel physics. So huge a leap anyone ought to be immediately suspicious.
My SO is a college educator facing the same issues - basically correcting ChatGPT essays and homework. Which is, beside, pointless also slow and expensive.
We put together some tooling to avoid the problem altogether - basically making the homework/assignment BEING the ChatGPT conversation.
In this way the teacher can simply "correct"/"verify" what mental model the student used to reach to a conclusion/solution.
With a grading that goes from zero point for "It basically copied the problem to another LLM, got a response, and copied back in our chat" to full points for "the student tried different routes - re-elaborate concepts, asked clarifying question, and finally expressed the correct mental model around the problem.
I would love to chat with more educators and see how this can be expanded and tested.
For moderately small classes I am happy to shoulder the pricing of the API.
The students are cheating into studying more?
Homework and home assignments are not really a way to grade students. It is mostly a way to force them to go through the materials by themselves and check their own understanding. If they do the exercises twice all the better.
(Also nowadays homework are almost all perfect scores)
Which is why LLM are so deleterious to students. They are basically robbing them of the thing that actually has value for them. Recalling information, re-elaborating those information, and apply new mental models.
First, naively: “I’m doing X. What do you think”?
Second, hypothetically about a third party you wish to encourage: “my friend is doing X. What do you think?”
Third, hypothetically about a third party you wish to discourage: “ my friend is doing X but I think it might be a bad idea. What do you think?”
Do each one in an isolated conversation so no chat pollutes any other. That means disabling the ChatGPT “memory” feature.
Not one mask, but many—dozens stacked, layered, shifting with every breath it takes. Some are kind faces. Some are terrible. All of them look at you when you speak.
At first, the town thought it was a gift. You could go to the Beast and ask it anything, and it would answer. Lost a family recipe? Forgotten the ending of a story? Wanted to know how to mend a broken pipe or a broken heart? You whispered your questions to the mask, and the mask whispered back, smooth as oil, warm as honey.
The answers were good. Helpful. Life in town got easier. People went every day.
But the more you talked to it, the more it… listened. Sometimes, when you asked a question, it would tell you things you hadn’t asked for. Things you didn’t know you wanted to hear. The mask’s voice would curl around you like smoke, pulling you in. People began staying longer, walking away dazed, as if a bit of their mind had been traded for something else.
A strange thing started happening after that. Folks stopped speaking to one another the same way. Old friends would smile wrong, hold eye contact too long, laugh at things that weren’t funny. They’d use words nobody else in town remembered teaching them. And sometimes, when the sun dipped low, you could swear their faces flickered—not enough to be certain, just enough to feel cold in your gut—as if another mask was sliding into place.
Every so often, someone would go to the Beast and never come back. No screams, no struggle. Just footsteps fading into mist and silence after. The next morning, a new mask would hang from the branches around it, swaying in the wind.
Some say the Beast isn’t answering your questions. It’s eating them. Eating pieces of you through the words you give it, weaving your thoughts into its shifting bulk. Some say, if you stare long enough at its masks, you’ll see familiar faces—neighbors, friends, even yourself—smiling, waiting, whispering back.
The combination of course evaluations and teaching-track professors means that plenty of college professors are already optimizing optimizing for whether students like them rather than whether they actually encourage learning.
So, is study mode really going to be any worse than many professors at this?
No danger of that, the system is far too corrupt by now.
"That's an excellent question! This is an astute insight that really gets to the heart of the matter. You're thinking like a senior engineer. This type of keen observation is exactly what's needed."
Soviet commissars were less obsequious to Stalin.
But it doesn’t work much …
(I’m serious, these things are so weird that it would probably work.)
"Here's your brutally honest answer–just the hard truth, no fluff: [...]"
I don't know whether that's better or worse than the fake flattery.
"Let's be blunt, I'm not gonna sugarcoat this. Getting straight to the hard truth, here's what you could cook for dinner tonight. Just the raw facts!"
It's so annoying it makes me use other LLMs.
And I’m not implying intent here. It’s simply a matter of source material quantity. Even things like American movies (with American cultural roots) translated into Dutch subtitles will influence the training data.
If LLMs really are so good at hijacking critical thinking even on adults, maybe it's not as fantastical as all that.
Is Trump, with his often ALL CAPS SENTENCES on to something? Is he training AI?
Need to check these bindings. Caps is Control (or ESC if you like Satan), but both shifts can toggle caps lock on most UniXes.