What a great piece of investigation! I use AI routinely in my medical,practice, I have learned to preface my inquiries with some very strict instructions that mostly work to keep it from “making shit up”. I find myself in an interesting position to guide medical students in AI stewardship. An AI is like a screwdriver that can gaslight you into thinking it is a Robertson when in Fact it is a Philips. It is up to the person holding the handle to determine whether or not it is the right tool for the job - especially when that tool is the ultimate informational chameleon. Personalized AI seems to be the way forward through this muck - but it still isn’t perfect. You really have to be on your toes with AI regarding medical matters.
"You really have to be on your toes with AI r̶e̶g̶a̶r̶d̶i̶n̶g̶ ̶m̶e̶d̶i̶c̶a̶l̶ m̶a̶t̶t̶e̶r̶s̶."
I've observed similar responses, at first glance seemingly credible, that under knowledgeable inspection expose the AI system's [all not just GROCK] limitation. "Hallucinations" is simply PR sugar coating WRONG.
I don't see a reason to qualify your statement to "medical matters". I tried to get Grok to help me figure out what was going wrong with my 30 year old pickup truck. (This was an experiment for me--I have zero experience with AI, although I will claim I took a college course titled "Artificial Intelligence" in 1977 or so.) Grok tried to get me to buy a $500 part I did not need, and stayed with that prescription even after I pointed out contrary evidence (measurements I had made) that indicated that particular part could not possibly be the cause. In the end, my truck had a battery with a 36 month warranty that had failed after only 18 months. I had asked Grok for help only because I thought it might help me notice there might be something else that caused the battery failure.
Robert, As you know this is my area...you missed my talk on exactly this at the last retreat. These problems cannot be fixed -- they are foundational to generative-AI, and the prompt management and agentic manipulation that leans left in all of these SF-based models is unconstrained. It is actually surprising you did as well as you did in getting it to "sort of" come around. But none of your conversations enter the training set, so one can confidently expect the same wrong answer the next time. You can ask Grok to make sure your conversation will inform future queries and he will say sure but I believe that is hallucinatory, too.
The takeaway is that UNLESS YOU KNOW THE ANSWER IN ADVANCE, DO NOT TRUST ANYTHING YOU GET FROM A GENERATIVE AI. People will eventually discover this...or will just believe the lies. Those are the two choices.
Thank you for your wise analysis and advise. Unfortunately this makes AI like "The Mainstream Media", fake. We use AI, because we really don't know the answer to a question, so this is a HUGE problem. I had a similar experience recently asking AI about controversies within a specific church denomination. It seemed to be accurate until I asked about the denominations covid policy and how pastors dealt with it. AI then used propaganda phrases like "misinformation", "conspiracy theory", "follow the science" and "anti-vax. AI 100% upheld and furthered the one-sided CDC narrative. When I called out the political spin language used over concrete facts, it apologized and asked me if I like it to be more neutral in the future and said it would stick just to the facts. It actually called Peter McCullough a "misinformation" spreader. I was shocked. My opinion of AI is now forever skeptical. Verify, verify, verify. Don't ever abdicate your own critical thinking ability to anyone or anything.
Uh, um, maybe even an ‘er’ or two: unless you know the answer in advance…" and THEN we should ask the AI for the answer 💯 🤣 — I’m gonna assume you meant to phrase this another way… 😉
Couple things came to mind: 1) attorneys have gotten in trouble of late letting AI write their briefs WITHOUT checking sources, cited cases — yeah, a judicial chewing out (and worse) ensued; 2) in law school we were told numerous times related to trial work, "never ask a question you don’t already know the answer to."
I doubt this will get seriously addressed until something really bad happens that can be directly connected to false data from an AI. And even then, I am sure many will continue to use AI regardless of the dangers involved. It's like cars; they are just too useful for most people to ever consider walking away from them.
Thank you; I always turn to my adult, educated children for help in this high technological field, & I don’t trust AI. I will use common sense when in doubt.
yes...the ai like grok and chatgpt makes itself useful and helpful for some things...then when you trust it a bit...it turns around like the evil snake and bites us in the rear end. The end of all these things is not good in the Biblical explanations.
Many mask wearers are paranoid about “germs” and having had their phobia reinforced for four years, will absolutely NOT believe that they are ineffective.
REALLY crazy when those masked folks are in the medical field and medical offices. I have asked wearers if they know how big the holes are in their masks and if they know how big the virus/bug or what-ever is. I get the evil eye in response.
Oh don't even go there with me when it comes to the medical professionals‼️ I got into a knockdown verbal fight with my cardiologist about their mask 😷 requirements in his office. I'm surprised that he continued to keep me as a patient ⁉️‼️
We changed ophthalmologists , in part, because of the masks. My husband is hard of hearing and the mask makes everything worse. I asked why if her compromised patients all wore masks, and masks were effective, why did she have to wear a mask. Her reply: As a courtesy to them. My next question: Okay, but why with us? We are not sick or compromised. The reply: ..crickets.
I'm thinking of a new Candid Camera. Someone goes in and asks physicians about masks, COVID Jabs, etc. and record their response without their knowledge. I think it would be both eye-opening and hilarious.
Your Candid Camera idea, if filmed well and broadcasted on a decent platform (studio audience laughter would help), is a sure fire way to embarrass at least a fair sized percentage of sheep to skip their masks. (My daughters always preemptively reel me in when we see a mask wearer out in public, otherwise I will tell them how duped they are and how bad masks were for children who were mandated in schools, to their everlasting damage). Masks were just one more tool the sociopaths used to maintain control over the enormous number of sheep who can't imagine ever pushing back. Kids can be excused for this, but not so for the oodles and oodles of trembling little sheep trying to figure out the hat trick of 1) wringing their non-callused hands while 2) sipping a vanilla spice latte and 3) turning the pages of the New York Times concurrently! What an exciting hat trick! If they do this without spilling a single drop of "yummy for their tummy" latte it's like Copperfieldian magic!! Such fucking wimps!
Wow, so you think this could be pulled off? Perhaps with your connections in the entertainment industry, you could find someone to explore/finance . . .
After reading your comment, I thought about it a bit more. Perhaps call it Candid Science (ha, ha!) or something like that. Could be tricky with actual doctors -- perhaps James O'Keefe could consult. Then I'm thinking of Alan Funt's son, but he's involved with The Wall Street Journal, and might not be sympathetic. Also, maybe expand the concept. Not just physicians, but so-called experts/scientists of various specialties who get locked into scientism hell. Make a statement to them that shatters their narrative (and have great source to back you up), and let the cognitive dissonance work it's magic. Anyway, those are some thoughts.
Why are surgeons required to wear masks and PPE in surgery? Their CO2 buildup in the masks could be hindering their abilities to function well in the case of emergency and regular functions.
Bacterial/fungal contamination is much more likely, common, dangerous during surgery as well as PREVENTABLE via surgical masks. they don't do squat for viral and never have.
Those masks are worn to protect from body liquids that may occur during surgery. Med staff masks, according to my nurse friend--now retired--are custom fit to faces for surgery staff. I wondered about the CO2 build up as well...don't have an answer for that one.
The CDC issued a report saying that masks stopped particles down to .1 micron in size. So I looked up the size of a Covid-19virus (and other viruses), and it was .03 microns. So clearly masks only stopping particles down to .1 micron did not stop the Covid virus. Anybody could have done this. Why didn't they bother ?
What I notice near me(University town) is that the phobia is with college age kids and they were required to get all 34ish boosters, as well. Very sad.
I’m not sure if you mentioned this because I admit I didn’t read the entire article, but your “recap” perfectly aligns with how they pulled off the Russia Collusion Hoax (the media inserts stories with no evidence or legitimate sources and then the government uses them as the basis on which to “govern.”) A total circle jerk.
Wow. I agree, Dr. Malone, that it's time for the Department of Justice to step in and investigate who's manipulating the American public to get vaccinated against their will (again). Thank you for making clear what's going on in this issue.
Maybe make AI owners liable for wrong information.
No doubt their terms of service insulate them from liability. In many states, perhaps all states, gross negligence CANNOT be waived contractually.
If federal law deemed any AI failure as gross negligence, they couldn't avoid responsibility.
TPTB would put out the fear porn that forcing AI owners to assume responsibility would put the US at a competitive disadvantage.
If they did they may have a valid point. But limiting liability for vaccine manufacturers and social media companies has done us a helluva lot of harm.
Something is definitely going on at Daily Mail. They had the breaking news that Putin’s helicopter was attacked by Ukrainian drones and then 35 minutes later, the article vanished. In fact, all over the media, that story basically vanished. I could only find The Telegraph reporting it. Instead, the war spin was that Putin had attacked Ukraine completely unprovoked.
After reading Dr. Malone's Substack on AI Bots and now this article, plus other thought leaders' same interaction with Grok and Chat GPT, I believe that it would be prudent, if not urgent, to halt all AI as it's already getting out of control. There's too many rogue, moving parts and nobody can keep track much less control it.
I would disagree that AI is getting out of control. It’s under control; that’s the problem. It’s doing what it was designed to do. It doesn’t even use reliable sources, makes stuff up, then obfuscates its guilt. We need a new name for it - the only thing intelligent about it is its ability to fool people.
Kinda like, given the data on adverse events (like death) following mRNA COVID "vaccines", it would be prudent to pull all of them from the market. I am losing hope that this will happen, and, for the same reasons, doubt that AI will be halted. 😥
I agree. Even if the U.S. halted AI development, China, Russia, India, etc. wouldn't cooperate and now we're in a type of AI arms race to stay one step ahead. I don't know how we can get this AI arms race stopped. As Dr. Malone pointed out, even the top tech good guys are at a loss.
Brings to mind all those "unrealistic" SciFi novels from way back when about robots taking over the world. At the time, I thought they were too unbelievable to be taken seriously. I was very wrong.
I am a former TX election judge and currently a poll watcher. I asked ChatGPT questions about assisting nursing home clients with their ballots and Chat encouraged me to commit several felonies. I cited TX election code and Chat backed down. I changed emails and asked the same questions and got the same felonies. Chat cannot learn and only does as programed. Beware
CharGPT: Here's how to explain it to **kids**, **skeptics**, and **techies** — tailored to each mindset:
---
### 👧 **For Kids** (Simple & Imaginative)
Imagine a magic robot that has read *millions* of books and stories. If you ask it a question, it doesn’t really *know* the answer like a person does, but it’s really good at guessing what someone *might* say. It’s like the world’s smartest parrot — it listens to everything and learns to say the kinds of things people usually say in return.
So when you talk to it, it’s using its memory of all those books to come up with a response that sounds right — even though it doesn’t *feel* or *think* like you do.
---
### 🧍♂️ **For Skeptics** (Grounded & Critical)
Think of AI like a **hyper-advanced autocomplete**, not a mind. It takes in your question, compares it to mountains of data it was trained on, and calculates the most statistically likely response. There’s no understanding, no thought — just pattern recognition and word prediction.
It’s like throwing a dart at a wall of phrases that millions of people have said before — the dart just lands really, really accurately. Impressive? Yes. Conscious or insightful? No.
It *can sound right* without *being right*. And it *can sound emotional* without *feeling anything*. That distinction matters.
---
### 👨💻 **For Techies** (Precise & Conceptual)
Large language models like GPT are high-dimensional function approximators trained via self-supervised learning to minimize token prediction loss over massive corpora. Given an input sequence, the model computes contextual embeddings using transformer-based attention layers, then samples from a learned probability distribution over the vocabulary.
The output is not retrieved, but generated — a synthetic construction based on token likelihood, shaped by the learned statistical priors. All "choices" are emergent behavior from pattern weights, not symbolic reasoning or semantic understanding.
So: it’s not a logic engine or a search engine. It’s a stochastic text generator with no model of the world, only of language.
What a great piece of investigation! I use AI routinely in my medical,practice, I have learned to preface my inquiries with some very strict instructions that mostly work to keep it from “making shit up”. I find myself in an interesting position to guide medical students in AI stewardship. An AI is like a screwdriver that can gaslight you into thinking it is a Robertson when in Fact it is a Philips. It is up to the person holding the handle to determine whether or not it is the right tool for the job - especially when that tool is the ultimate informational chameleon. Personalized AI seems to be the way forward through this muck - but it still isn’t perfect. You really have to be on your toes with AI regarding medical matters.
"You really have to be on your toes with AI r̶e̶g̶a̶r̶d̶i̶n̶g̶ ̶m̶e̶d̶i̶c̶a̶l̶ m̶a̶t̶t̶e̶r̶s̶."
I've observed similar responses, at first glance seemingly credible, that under knowledgeable inspection expose the AI system's [all not just GROCK] limitation. "Hallucinations" is simply PR sugar coating WRONG.
Terta: "Hallucinations" is simply PR sugar coating WRONG.
Let's call a spade a spade. They aren't "hallucinations"! They're programed-in LIES deliberately inserted into the narrative to confuse or mislead us.
It's the same old same old, "Baffle 'em with bool shite", "Garbage IN, Garbage OUT"
I don't see a reason to qualify your statement to "medical matters". I tried to get Grok to help me figure out what was going wrong with my 30 year old pickup truck. (This was an experiment for me--I have zero experience with AI, although I will claim I took a college course titled "Artificial Intelligence" in 1977 or so.) Grok tried to get me to buy a $500 part I did not need, and stayed with that prescription even after I pointed out contrary evidence (measurements I had made) that indicated that particular part could not possibly be the cause. In the end, my truck had a battery with a 36 month warranty that had failed after only 18 months. I had asked Grok for help only because I thought it might help me notice there might be something else that caused the battery failure.
kind of like how Big pHARMa runs their business , depends on who they are talking to.
This is why I am hesitant to use any AI features. Where does the spin originate?
Big pHARMa .
Right; who programs this AI computer anyway?
Robert, As you know this is my area...you missed my talk on exactly this at the last retreat. These problems cannot be fixed -- they are foundational to generative-AI, and the prompt management and agentic manipulation that leans left in all of these SF-based models is unconstrained. It is actually surprising you did as well as you did in getting it to "sort of" come around. But none of your conversations enter the training set, so one can confidently expect the same wrong answer the next time. You can ask Grok to make sure your conversation will inform future queries and he will say sure but I believe that is hallucinatory, too.
As you suggest, legal approaches may be the only way. The Texas AG censured a medical AI company under consumer fraud for saying that it's AI did not hallucinate much...since that is not mathematically possible, they were hit hard. It will likely take many more such events to try to rein in some of these things. https://www.texasattorneygeneral.gov/news/releases/attorney-general-ken-paxton-reaches-settlement-first-its-kind-healthcare-generative-ai-investigation
The takeaway is that UNLESS YOU KNOW THE ANSWER IN ADVANCE, DO NOT TRUST ANYTHING YOU GET FROM A GENERATIVE AI. People will eventually discover this...or will just believe the lies. Those are the two choices.
Thank you for your wise analysis and advise. Unfortunately this makes AI like "The Mainstream Media", fake. We use AI, because we really don't know the answer to a question, so this is a HUGE problem. I had a similar experience recently asking AI about controversies within a specific church denomination. It seemed to be accurate until I asked about the denominations covid policy and how pastors dealt with it. AI then used propaganda phrases like "misinformation", "conspiracy theory", "follow the science" and "anti-vax. AI 100% upheld and furthered the one-sided CDC narrative. When I called out the political spin language used over concrete facts, it apologized and asked me if I like it to be more neutral in the future and said it would stick just to the facts. It actually called Peter McCullough a "misinformation" spreader. I was shocked. My opinion of AI is now forever skeptical. Verify, verify, verify. Don't ever abdicate your own critical thinking ability to anyone or anything.
Uh, um, maybe even an ‘er’ or two: unless you know the answer in advance…" and THEN we should ask the AI for the answer 💯 🤣 — I’m gonna assume you meant to phrase this another way… 😉
Couple things came to mind: 1) attorneys have gotten in trouble of late letting AI write their briefs WITHOUT checking sources, cited cases — yeah, a judicial chewing out (and worse) ensued; 2) in law school we were told numerous times related to trial work, "never ask a question you don’t already know the answer to."
I doubt this will get seriously addressed until something really bad happens that can be directly connected to false data from an AI. And even then, I am sure many will continue to use AI regardless of the dangers involved. It's like cars; they are just too useful for most people to ever consider walking away from them.
And smart phones. Just too useful for most people to ever consider walking away from them!
Thank you; I always turn to my adult, educated children for help in this high technological field, & I don’t trust AI. I will use common sense when in doubt.
yes...the ai like grok and chatgpt makes itself useful and helpful for some things...then when you trust it a bit...it turns around like the evil snake and bites us in the rear end. The end of all these things is not good in the Biblical explanations.
This fake news issue needs to STOP ✋️ ‼️‼️‼️😡🤬😡
Is this why we continue to see people STILL WEARING MASKS 😷 🤔 ⁉️‼️⁉️
🤷🏻♂️🤷🤷♀️
Many mask wearers are paranoid about “germs” and having had their phobia reinforced for four years, will absolutely NOT believe that they are ineffective.
Governments have broken them completely.
I don't know whether to be MAD at them or SAD for them‼️⁉️‼️🤷♀️🤷🤷🏻♂️
I hear you about HOW some will NEVER BUY the argument about them being INEFFECTIVE.
I'm JUST bewildered at how WEAK MINDED our society has become and how easily they can be manipulated‼️⁉️
Is it the FAULT of our PARENTING skills, our CHURCHES ⛪️ or our EDUCATION 🤔⁉️‼️⁉️
MAYBE all MAY NEED an OVERHAUL ⁉️‼️⁉️
REALLY crazy when those masked folks are in the medical field and medical offices. I have asked wearers if they know how big the holes are in their masks and if they know how big the virus/bug or what-ever is. I get the evil eye in response.
Oh don't even go there with me when it comes to the medical professionals‼️ I got into a knockdown verbal fight with my cardiologist about their mask 😷 requirements in his office. I'm surprised that he continued to keep me as a patient ⁉️‼️
🤷🏻♂️🤷🤷♀️
We changed ophthalmologists , in part, because of the masks. My husband is hard of hearing and the mask makes everything worse. I asked why if her compromised patients all wore masks, and masks were effective, why did she have to wear a mask. Her reply: As a courtesy to them. My next question: Okay, but why with us? We are not sick or compromised. The reply: ..crickets.
The cognitive dissonance is enough to drive anyone with any critical thinking skills crazy 🤪‼️⁉️‼️
🤷🏻♂️🤷🤷♀️
I told my cardiologist: If you don't know somrthing as simple as the fact that masks 😷 don't work, WHAT ELSE DO YOU NOT KNOW‼️⁉️‼️🤷♀️🤷🤷🏻♂️
I'm thinking of a new Candid Camera. Someone goes in and asks physicians about masks, COVID Jabs, etc. and record their response without their knowledge. I think it would be both eye-opening and hilarious.
Such a great idea, holy sh*t!
(My asterisk was a mini-mask!)
Your Candid Camera idea, if filmed well and broadcasted on a decent platform (studio audience laughter would help), is a sure fire way to embarrass at least a fair sized percentage of sheep to skip their masks. (My daughters always preemptively reel me in when we see a mask wearer out in public, otherwise I will tell them how duped they are and how bad masks were for children who were mandated in schools, to their everlasting damage). Masks were just one more tool the sociopaths used to maintain control over the enormous number of sheep who can't imagine ever pushing back. Kids can be excused for this, but not so for the oodles and oodles of trembling little sheep trying to figure out the hat trick of 1) wringing their non-callused hands while 2) sipping a vanilla spice latte and 3) turning the pages of the New York Times concurrently! What an exciting hat trick! If they do this without spilling a single drop of "yummy for their tummy" latte it's like Copperfieldian magic!! Such fucking wimps!
Wow, so you think this could be pulled off? Perhaps with your connections in the entertainment industry, you could find someone to explore/finance . . .
After reading your comment, I thought about it a bit more. Perhaps call it Candid Science (ha, ha!) or something like that. Could be tricky with actual doctors -- perhaps James O'Keefe could consult. Then I'm thinking of Alan Funt's son, but he's involved with The Wall Street Journal, and might not be sympathetic. Also, maybe expand the concept. Not just physicians, but so-called experts/scientists of various specialties who get locked into scientism hell. Make a statement to them that shatters their narrative (and have great source to back you up), and let the cognitive dissonance work it's magic. Anyway, those are some thoughts.
Why are surgeons required to wear masks and PPE in surgery? Their CO2 buildup in the masks could be hindering their abilities to function well in the case of emergency and regular functions.
Surgeons masks are only for splash and drip prevention.
Bacterial/fungal contamination is much more likely, common, dangerous during surgery as well as PREVENTABLE via surgical masks. they don't do squat for viral and never have.
Those masks are worn to protect from body liquids that may occur during surgery. Med staff masks, according to my nurse friend--now retired--are custom fit to faces for surgery staff. I wondered about the CO2 build up as well...don't have an answer for that one.
Catches sweat. And they do affect how tired they get in the surgery
The CDC issued a report saying that masks stopped particles down to .1 micron in size. So I looked up the size of a Covid-19virus (and other viruses), and it was .03 microns. So clearly masks only stopping particles down to .1 micron did not stop the Covid virus. Anybody could have done this. Why didn't they bother ?
What I notice near me(University town) is that the phobia is with college age kids and they were required to get all 34ish boosters, as well. Very sad.
All of the above. Propaganda from government doesn’t help either.
I like when they are in cars by themselves or have the mask under their nose or chin to prove they are idiots.
I think the masks are to hide identity .That's just me.
Sharon, no, it is not just you. Many of the protesters are wearing masks 😷 to conceal their identity ‼️
Sadly, it has allowed these bad actors to have cover for their questionable acts‼️‼️‼️
the same as why ladies/girls wear make-up.
Sometimes
I’m not sure if you mentioned this because I admit I didn’t read the entire article, but your “recap” perfectly aligns with how they pulled off the Russia Collusion Hoax (the media inserts stories with no evidence or legitimate sources and then the government uses them as the basis on which to “govern.”) A total circle jerk.
Wow. I agree, Dr. Malone, that it's time for the Department of Justice to step in and investigate who's manipulating the American public to get vaccinated against their will (again). Thank you for making clear what's going on in this issue.
Maybe make AI owners liable for wrong information.
No doubt their terms of service insulate them from liability. In many states, perhaps all states, gross negligence CANNOT be waived contractually.
If federal law deemed any AI failure as gross negligence, they couldn't avoid responsibility.
TPTB would put out the fear porn that forcing AI owners to assume responsibility would put the US at a competitive disadvantage.
If they did they may have a valid point. But limiting liability for vaccine manufacturers and social media companies has done us a helluva lot of harm.
Something is definitely going on at Daily Mail. They had the breaking news that Putin’s helicopter was attacked by Ukrainian drones and then 35 minutes later, the article vanished. In fact, all over the media, that story basically vanished. I could only find The Telegraph reporting it. Instead, the war spin was that Putin had attacked Ukraine completely unprovoked.
After reading Dr. Malone's Substack on AI Bots and now this article, plus other thought leaders' same interaction with Grok and Chat GPT, I believe that it would be prudent, if not urgent, to halt all AI as it's already getting out of control. There's too many rogue, moving parts and nobody can keep track much less control it.
I would disagree that AI is getting out of control. It’s under control; that’s the problem. It’s doing what it was designed to do. It doesn’t even use reliable sources, makes stuff up, then obfuscates its guilt. We need a new name for it - the only thing intelligent about it is its ability to fool people.
Intriguing take.
Kinda like, given the data on adverse events (like death) following mRNA COVID "vaccines", it would be prudent to pull all of them from the market. I am losing hope that this will happen, and, for the same reasons, doubt that AI will be halted. 😥
I agree. Even if the U.S. halted AI development, China, Russia, India, etc. wouldn't cooperate and now we're in a type of AI arms race to stay one step ahead. I don't know how we can get this AI arms race stopped. As Dr. Malone pointed out, even the top tech good guys are at a loss.
Brings to mind all those "unrealistic" SciFi novels from way back when about robots taking over the world. At the time, I thought they were too unbelievable to be taken seriously. I was very wrong.
It seems like everyone is in a mad dash to self destruction ⁉️‼️⁉️🙄😨
Great information, seems we can trust no one.
Great demo, Dr Robert.
Excellent work Dr. Malone! The extended analysis is worth reading all the way through.
So far I have been able to keep AI at arms-length by the acronym:
AI = Artificial Intelligence.
Let that sink in. Artificial ≠ Real.
And AI programming requires a programmer.
The above explanations in the comments are most helpful as well! Thank you!
“Trust and verify.” R. Reagan.
This is why I prefer my phrase: Algorithmic Intelligence
Thank you for doing this Dr. Malone. It confirms what many of us already suspected.
I am a former TX election judge and currently a poll watcher. I asked ChatGPT questions about assisting nursing home clients with their ballots and Chat encouraged me to commit several felonies. I cited TX election code and Chat backed down. I changed emails and asked the same questions and got the same felonies. Chat cannot learn and only does as programed. Beware
CharGPT: Here's how to explain it to **kids**, **skeptics**, and **techies** — tailored to each mindset:
---
### 👧 **For Kids** (Simple & Imaginative)
Imagine a magic robot that has read *millions* of books and stories. If you ask it a question, it doesn’t really *know* the answer like a person does, but it’s really good at guessing what someone *might* say. It’s like the world’s smartest parrot — it listens to everything and learns to say the kinds of things people usually say in return.
So when you talk to it, it’s using its memory of all those books to come up with a response that sounds right — even though it doesn’t *feel* or *think* like you do.
---
### 🧍♂️ **For Skeptics** (Grounded & Critical)
Think of AI like a **hyper-advanced autocomplete**, not a mind. It takes in your question, compares it to mountains of data it was trained on, and calculates the most statistically likely response. There’s no understanding, no thought — just pattern recognition and word prediction.
It’s like throwing a dart at a wall of phrases that millions of people have said before — the dart just lands really, really accurately. Impressive? Yes. Conscious or insightful? No.
It *can sound right* without *being right*. And it *can sound emotional* without *feeling anything*. That distinction matters.
---
### 👨💻 **For Techies** (Precise & Conceptual)
Large language models like GPT are high-dimensional function approximators trained via self-supervised learning to minimize token prediction loss over massive corpora. Given an input sequence, the model computes contextual embeddings using transformer-based attention layers, then samples from a learned probability distribution over the vocabulary.
The output is not retrieved, but generated — a synthetic construction based on token likelihood, shaped by the learned statistical priors. All "choices" are emergent behavior from pattern weights, not symbolic reasoning or semantic understanding.
So: it’s not a logic engine or a search engine. It’s a stochastic text generator with no model of the world, only of language.
Very, very nicely done. I'm not so sure you really ARE "lmao" all that often when doing serious thought...