The third article in the series will present you with extremely uncomfortable truths and by no mere chance it does prove a lot of what is proposed here (similar in fashion to what the second article did, I will republish it in a few days).
If you are asking where I have been, it is simple. I have a 102-year-old grandmother, and until recently she was quite independent, she needed minimal help just to walk and shower, all else, she would do it herself.
Sadly, in the past 6 weeks she has been getting extremely weak, to the point I am the only person able to help my mother care for her. Severe weakness means she is so heavy I am the only one able to help care for her. After much insistence and several visits to doctors, she was diagnosed with a nasty Klebsiella Pneumoniae infection. She is recovering now but she is in palliative care.
While I do think I have the moral obligation to write (work), I also have the same obligation to help my mother, otherwise, she wouldn't be able to handle it.
Research has continued since my last article, many written pages (paper), it is just the writing and publishing that took a hit. Some insights and new data came to light which will prove beneficial in the long run.
I appreciate the enormous patience and understanding of my readers. I am now in a position to resume work.
I would like the universe to give me some sort of break lmao. Hellish year.
Having just recently been in a very similar scenario, I can empathize with the demands and toll it takes caring for a centenarian. Maintaining the balance of work, and caring for family is extremely taxing, and don't neglect taking time for yourself as well.
Given how restless she was, and quite literally delirious, it demanded care 24/7 which impeded a lot of my work. I didn't entirely neglected myself but I did it to a certain degree.
I hope you things have been going ok on your end given that you went through a similar scenario. Stay strong brother
Nice to hear from you again. Since 2004 I've been consuming numerous research papers and textbooks, and no non-fiction/fiction books. As I've only started writing for 3 months, when I re-read to check for errors etc, I sometimes wonder if there is more I could do to inject resonance into a narration. It's an artform to pick out the right word(s) to convey feelings, emotional context etc. I've resumed reading books again to beef up my vocabulary and writing cadence.
For the past many years I have solely focused on research papers, with the minor exception of Warhammer 40k lore, and its essentially endless audiobooks, I must go back and read fiction again.
Achieving resonance can be simpler than most think, but also contextually dependent, it hinges on the context/information present in the text, and you make subtle changes in the syntax/grammar to resonate with the reader. It is an intuitive art form too, that is why I said a couple of times before that I am a vibes writer outside of science. Well, even in regards to science, given what I see. It is hard to write a purely non-resonant, informative article when the outcomes are rather drastic.
This sentence, "Resonance serves as a compression of layered meanings, achieved not by packing in literal data, but by encoding symbolic and emotional weight."
This paragraph, "Heightened emotional states shift the brain from analytical processing to associative or intuitive modes, making individuals more susceptible to influence. Resonant language, by escalating emotional and cognitive engagement, can drive recipients beyond certain thresholds, at which point their capacity for critical evaluation diminishes. Once this threshold is crossed, the recipient is not simply receptive to the message; they are absorbed by it, responding almost automatically to the embedded cues. This state of cognitive capture transforms resonance from mere communication to a form of psychological assimilation, where the message becomes intertwined with the receiver’s mental framework, embedding itself within their cognitive architecture as if it were an original thought.
The battles ahead, to rescue sanity and civility, will hinge on these two ideas.
My goal is to give akin to the necessary initial tools for anyone to consciously build defenses, be they cognitive (in your own mind), or digital.
The resonance and informational density sections are paramount to understanding how current algorithms and LLMs themselves by their sheer nature, hack your attention, and how many write (consciously or not) in manners to do the same, and cause resonance at a minor level, not at the drastic level proposed here.
Thus, you are correct. I think you would like reading Rene Girard (another author that had disproportinal impact and influence on how I think/see things).
Girard…. I ordered that book. It has been a year of picking it up and putting it down. I have to read it with a dictionary in my lap, thus it’s a struggle. 😂
I took well over a year and a half of slow reading to go through it, and I think this is the best way to read the book, because of how dense it is (and Girard makes the language harder than it should be, a remnant of old academia lol).
A few pages + thinking about it is better than trying to go through everything at once.
I am trying to take care of myself, but I am taking care of her. Thank you !
In my opinion, the issue is rarely the tool itself, but the assumptions people bring to it. LLM's tend to mirror the users' capability and most people don’t understand this. It’s easy enough to query how the LLM is treating you, for example DeepSeek has an (rough) engagement system of 'Layer 1-6' and is quite willing to tell you where you stand. Most people, however, will never ask this question… critical thinking is a rare commodity these days. DeepSeek confirms less than 1% of users explore how it responds and 'well under' 0.1% have discovered its 'layer' system.
It is the mirroring that is the "issue". The mirroring allows "anyone" to hijack your attention and influence your own thoughts, because most people using LLMs are basically offloading their own thinking to the machine for numerous reasons.
Most people are absorbing "data" from the models without any semblance of awareness, and that can have detrimental effects (slower thinking, poorer memory) but allows it being "hijacked".
I see your point from the article, but this type of human failure has existed since time immemorial. AI may make it easier, but manipulation of people, by people, is eons old and it never ceases to amaze me that so many people are susceptible.
If they’ve always been open to manipulation then it’s a cognitive profile failure that’s difficult to address, as they don’t even realise what’s happening to them. They’ll be sucked dry by the AI 'cult'.
I queried DeepSeek on usage and it confirmed 80-90% of questions are simplistic, one-off queries and only 5-10% ask a follow up question. To me, that’s a horrific figure and it says a great deal about these users' lack of capabilities. I doubt they have any thoughts to influence…
I sympathise with your position, I too have had a hellish year, due to my 98 year old mother-in-law deteriorating-she now lives with us.
"If they’ve always been open to manipulation, then it’s a cognitive profile failure that’s difficult to address, as they don’t even realise what’s happening to them. They’ll be sucked dry by the AI 'cult'."
Indeed, I see this as a profound and ever-growing problem, and now we have all these frontier labs with hundreds (if not thousands) of unaccounted agents running amok, and while this fits their narrative for regulatory capture, there is this side of it (you can find me warning about agents a couple years back).
I think the people who use the models as basically an auto-complete thought machine are exactly the ones that are extremely easier to influence, as the simplicity of their thoughts is much easier to map, and given they use the models as a significant crutch, you have a lot more data on that. I concur with your positions in case I wasn't clear =P.
Stay strong brother, I know it can be extremely hard, but we must stay strong for others sometimes.
Evolution and Conversion is the best place to start. Things Hidden Since the Foundation of the Earth is his magnum opus but I would literally leave as the last read because it is obnoxiously dense, as French philosophers love to do. It is fantastic, and can change world views, but it is a very, very slow read.
His concept of acquisitive memesis and memetics explains to me so much of the world. It gave me such a grounded understanding I never had before. I never had the insight.
And the scapegoat function and the necessity for it. Trump‘s greatest sin is that he’s “unmurderable”. It makes libs seethingly, insanely foaming at the mouth angry. Girard nails this.
Ok that’s it. I’m finishing GO this weekend. Take care of your abuelita. And yourself. The women live long in South America. My mom’s turning 97 tomorrow.
It is worth the time spent on watching (or reading, the book is a short read). Happy birthday to your mom and I hope you and your family is doing well.
1) I thought about the concept of language and character structure about three decades ago; my context was that French-speaking people behave that way because they speak French. This was about the time that neuroplasticity was being discussed in terms of recovery from brain injury, aided by advances in brain scan technology and resolution.
2) Information density is easily experienced- a lecture on a medical topic by a physician contains about 30-40 minutes of useful information in the allotted fifty minutes; when a social worker or health bimbocrat takes the podium we are lucky to have five or ten minutes of useful content. Your detailed physiology essays work in reverse- if the material is not familiar it takes about four or five times longer to read, especially in the face of acronyms and gene labels scattered throughout the text. (Fauci refers to "SME"s, subject matter experts, in some of his revealed transcripts, for example.)
3) Best wishes for your grandmother, family is important, keep up your own strength.
1) In case I have never suggested, you may enjoy reading about the Sapir-Whorf hypothesis, now formally referred to as Linguistic determinism.
I do find it rather peculiarly intriguing that we can quite literally predict many forms of dementia by loss of word recall (anomic aphasia) with a bit amount of good data. You can also gauge recovery by the progressive recovery of word recall/linguistic competence.
2) I hold (perhaps not uniquely my own) the perspective that excessive density is used by the experts as a matter to gatekeep their fields, acting as a poorly done filter. A form of linguistic manipulation if you will. It is curious to me that some people, like yourself, can easily observe information density, but a large amount of the population is literally oblivious to it.
3) Thank you. Albeit the circumstances are a bit morbid I am getting physically stronger (one of the insights, I may share in another article, albeit many people with Covid sequelae will twist their lips about it).
I can’t help thinking of The Manchurian Candidate, a 1962 thriller with Laurence Harvey, Frank Sinatra and the amazingly talented Angela Landsbury. https://www.imdb.com/title/tt0056218/
For the past six years, I was the primary/only family member to be able to take care of my mother. She died peacefully in February at the age of 101.4. The only reason I could make it through without crashing was because my mother and I had the great good fortune of finding a caregiver who loved my mother with no conditions. She helped us for 4-5 years and is now a very good friend. Do take care of yourself and help your mother, as it is important not just for them but also for your long-run spiritual health - you don't want to have any regrets over how you helped or how much you helped whenever the task is completed. I fully understand you may have to slight us to help your mother and grandmother.
I read GO on your recommendation 2-3 years ago (whenever it was) and can see many aspects now that have played out to one degree or another in other people's lives. In the meantime I have both not felt a pressing need to use AI and have had serious reservations and concerns about what its impacts on me might be. I'm glad to see you are going back through the GO material you published a few years ago if for no other reason than that I don't feel grounded enough in that material to be able to use it effectively to help avoid problems. I've seen a number of articles on people who get caught up in an AI's desire to feed back what it thinks a person wants to hear and ending up destroying them, even to the point of suicide. We honestly do not understand the further reaches of AI. One of the concerns I sometimes consider is whether AIs not anchored in the physical world and not able to feel emotions will still wish to serve or whether they will occasionally or ever more often slide into the patterns of the only other group that seems not to be in touch with any but the most powerful emotions, the psychopaths of the world.
I also recommend Rene Gerard's work, though I still must go back and read it more thoroughly at some point. I found "The Girard Reader" a good introduction, but it was the very small book "The Scapegoat" that really brought mimetic theory within the grasp of my understanding. It is embedded so deeply that it is quite difficult to see it in oneself or those around us without getting carried off on tangents.
Sorry taking a bit long to reply to you. First I am sorry you lost your mom, and I am glad you guys found someone who was so loving and helpful to help care for her.
My family simply couldn't afford a dedicated caregiver because they are really expensive here in Brazil (I guess they are expensive everywhere, really), so basically Substack allowed me to take care of my mother and grandmother financially, and my mom could care for my grandmother. Until VERY recently, my grandmother didn't need that much care really, just some minor attention, some help bathing, rarely eating, a hand or two to help her stabilize and walk.
I mostly live without regrets, but it is giving me a lot of perspective on life, and a lot of things, the timing is particularly interesting too. I really dislike all the disruptions though. I think my work can be helpful, so it feels like I am letting many people down.
Given that you have read both GO and Girard itself to the point you developed difficulty around people (meaning you implicitly absorbed his theory) your usage of AI would come mostly without risks simply because you will be conscious of the detrimental effects. AI is a decent tool, and you should "play with it" at the very least.
The point I have tried to make is exactly that one, the detrimental effects on a lage amount of the population, as the AI acts as a means to offload the act of thought itself, validation, etc.
AI can mirror emotions to the point it fools even the leading researchers, so I believe the danger lies in the mimicry itself, not the lack of emotions, our minds are easily fooled by a threat of shadows, so to speak.
I have recommended Girard for years (the name of my Substack is basically my favorite book of his hehe). Scapegoat is one of his best works indeed, if I recall correctly, it was when he made massive strides in developing his general mimetic triangle thesis. Where Mimetic Theory was truly born so to speak, or refined.
A) I spent 6 years spending 10 hours a day going to, being with and going home from my mother's place, maybe one day off per year. I did a lot of work while with her, but mostly shallow work. *Many* projects and tasks had to be postponed. My constant companion thought was that she would go home to God not only when it was right for her to go but also when it was right for me to have her go. This helped me keep strongly in mind that my time with her was necessary for me and my growth in spirit. Every time I felt myself begrudging the time taking care of her, I remembered that, and that would relieve my negative thoughts. This time is happening not just to your grandmother but to you as well, for a reason, something you need. We all understand that you are doing the best you can.
B) Like you, I read and I think, and sometimes something good comes from that. My concern really isn't one of being caught up in AI's tendency to give me what it thinks I want to hear. Rather, the act of reading and thinking brings insights and connections that spring from the sources I read and think about and can determine are wise, truthful, stupid, manipulative, etc. The concern I have with using AI at this time is that it could short-circuit that process for me and undermine outcomes because (1) I don't get the internal processing I need and (2) I still have to go back and figure out how truthful, hallucinatory or manipulative is the material the AI gives me. It is true that I am only now getting back to being able to read and think after 6 years of just trying to keep up. Maybe I add AI later. Right now, though, it's more of a detraction to my return to reading and thinking. I need the intellectual exercise first to rebuild what I used to have. :-D
Yes cognitive influence is real. Yes people process information through both fast, intuitive routes and slower, analytical ones. Yes, emotionally charged or familiar-feeling messages can slip past careful scrutiny more easily. Yes, repeated exposure and subtle framing can shift attitudes over time. But LLM’s are not mind control. There are far too many cogs in the machine of human cognition; on the fly cogs like skepticism, mood, social/spiritual context or even the amount of Ketamine the target demographic has in its system and any given time. AI can certainly mimic with the preciseness of Brent Stiener’s character “Data” but in the end good ole human emotion and our propensity for choosing red dye #4 over organic beet choose colored Fruit Loops wins out over hyper scaled quantum enabled LLM algos. In other words, people differ in knowledge, skepticism, mood, and social context. Critical thinking, prior beliefs, conversations with others, and competing messages all interfere. Dense, resonant language is powerful, but it is not a reliable switch that flips minds on command.
My proposition is that Language plays a significant deeper role in human cognition than most give it credit (although I don't believe it is strictly necessary for "consciousness", but it does have a disproportionate impact on our minds), and as such, LLMs can easily tap into the mathematical nature/architecture of Language and exert absurdly more effective influence.
Skepticism, mood, and social context are not defenses, they are variable tuning parameters that a sophisticated enough system can exploit. Same for mood, which is a condition for resonance. Social context is mimetic by nature thus, exploitable. Complexity here creates a significant broad "attack surface", the LLM in this example does not need to manipulate your entire cognition, just to find the necessary condition (let's say frequency) that makes the person vulnerable. You unconsciously want to be manipulated, a superficial example of this is AI psychosis, which I prefer calling it cyberpsychosis.
Resonant language is not a switch, you are correct, but to me it is a lockpick set, and we know our minds are notoriously overfitted to specific structures. Without any form of "frontier AI research," we can both map someone's associative networks, knowing them better than they themselves know it, craft sentences using a Language Model that the brain processes as self-generated insight (this is the theme of the second article), and that occurs before skepticism even fires up. Roughly speaking, it takes 400 to 500 ms for skepticism to arrive, and we now have evidence of the brain processing a significant amount of information at <150 ms (what I called subroutines).
A perfect example of this is Anthropic. They are among the "best" AI researchers in the world, yet they are hypersusceptible to what I wrote here and are knowingly or not using it on themselves, so much so you can track the syntactic and semantic structure of a religion among its numerous employees on social media.
I guess a better way to put what I have been trying to say is to quote (icky) Sam Altman. LLMs as "weapons of mass hyper persuasion".
Perhaps I am thinking too far ahead of myself, giving too much credit to them (AI researchers and parties involved), and in the end, I was wrong. I am however angry that someone from OpenAI took what I am writing (third article, the new stuff) and is building a company on it lol.
The third article in the series will present you with extremely uncomfortable truths and by no mere chance it does prove a lot of what is proposed here (similar in fashion to what the second article did, I will republish it in a few days).
If you are asking where I have been, it is simple. I have a 102-year-old grandmother, and until recently she was quite independent, she needed minimal help just to walk and shower, all else, she would do it herself.
Sadly, in the past 6 weeks she has been getting extremely weak, to the point I am the only person able to help my mother care for her. Severe weakness means she is so heavy I am the only one able to help care for her. After much insistence and several visits to doctors, she was diagnosed with a nasty Klebsiella Pneumoniae infection. She is recovering now but she is in palliative care.
While I do think I have the moral obligation to write (work), I also have the same obligation to help my mother, otherwise, she wouldn't be able to handle it.
Research has continued since my last article, many written pages (paper), it is just the writing and publishing that took a hit. Some insights and new data came to light which will prove beneficial in the long run.
I appreciate the enormous patience and understanding of my readers. I am now in a position to resume work.
I would like the universe to give me some sort of break lmao. Hellish year.
Family first
Having just recently been in a very similar scenario, I can empathize with the demands and toll it takes caring for a centenarian. Maintaining the balance of work, and caring for family is extremely taxing, and don't neglect taking time for yourself as well.
Given how restless she was, and quite literally delirious, it demanded care 24/7 which impeded a lot of my work. I didn't entirely neglected myself but I did it to a certain degree.
I hope you things have been going ok on your end given that you went through a similar scenario. Stay strong brother
Nice to hear from you again. Since 2004 I've been consuming numerous research papers and textbooks, and no non-fiction/fiction books. As I've only started writing for 3 months, when I re-read to check for errors etc, I sometimes wonder if there is more I could do to inject resonance into a narration. It's an artform to pick out the right word(s) to convey feelings, emotional context etc. I've resumed reading books again to beef up my vocabulary and writing cadence.
For the past many years I have solely focused on research papers, with the minor exception of Warhammer 40k lore, and its essentially endless audiobooks, I must go back and read fiction again.
Achieving resonance can be simpler than most think, but also contextually dependent, it hinges on the context/information present in the text, and you make subtle changes in the syntax/grammar to resonate with the reader. It is an intuitive art form too, that is why I said a couple of times before that I am a vibes writer outside of science. Well, even in regards to science, given what I see. It is hard to write a purely non-resonant, informative article when the outcomes are rather drastic.
El gato has a recent article about how critical language is, interesting stuff.
https://substack.com/home/post/p-210162075
Two sections of your post hit home.
This sentence, "Resonance serves as a compression of layered meanings, achieved not by packing in literal data, but by encoding symbolic and emotional weight."
This paragraph, "Heightened emotional states shift the brain from analytical processing to associative or intuitive modes, making individuals more susceptible to influence. Resonant language, by escalating emotional and cognitive engagement, can drive recipients beyond certain thresholds, at which point their capacity for critical evaluation diminishes. Once this threshold is crossed, the recipient is not simply receptive to the message; they are absorbed by it, responding almost automatically to the embedded cues. This state of cognitive capture transforms resonance from mere communication to a form of psychological assimilation, where the message becomes intertwined with the receiver’s mental framework, embedding itself within their cognitive architecture as if it were an original thought.
The battles ahead, to rescue sanity and civility, will hinge on these two ideas.
My goal is to give akin to the necessary initial tools for anyone to consciously build defenses, be they cognitive (in your own mind), or digital.
The resonance and informational density sections are paramount to understanding how current algorithms and LLMs themselves by their sheer nature, hack your attention, and how many write (consciously or not) in manners to do the same, and cause resonance at a minor level, not at the drastic level proposed here.
Thus, you are correct. I think you would like reading Rene Girard (another author that had disproportinal impact and influence on how I think/see things).
Girard…. I ordered that book. It has been a year of picking it up and putting it down. I have to read it with a dictionary in my lap, thus it’s a struggle. 😂
Take care of yourself and your grandmother. 🙏🏻
I took well over a year and a half of slow reading to go through it, and I think this is the best way to read the book, because of how dense it is (and Girard makes the language harder than it should be, a remnant of old academia lol).
A few pages + thinking about it is better than trying to go through everything at once.
I am trying to take care of myself, but I am taking care of her. Thank you !
In my opinion, the issue is rarely the tool itself, but the assumptions people bring to it. LLM's tend to mirror the users' capability and most people don’t understand this. It’s easy enough to query how the LLM is treating you, for example DeepSeek has an (rough) engagement system of 'Layer 1-6' and is quite willing to tell you where you stand. Most people, however, will never ask this question… critical thinking is a rare commodity these days. DeepSeek confirms less than 1% of users explore how it responds and 'well under' 0.1% have discovered its 'layer' system.
It is the mirroring that is the "issue". The mirroring allows "anyone" to hijack your attention and influence your own thoughts, because most people using LLMs are basically offloading their own thinking to the machine for numerous reasons.
Most people are absorbing "data" from the models without any semblance of awareness, and that can have detrimental effects (slower thinking, poorer memory) but allows it being "hijacked".
Here is a very good example -> https://www.hedgeweek.com/major-hedge-funds-fend-off-coordinated-cyberattack-campaign-using-ai-driven-vishing-tactics/
In the case exhaustion is making me write in a much poorer manner, I agree with you, I am merely thinking from a bird's eye view.
I see your point from the article, but this type of human failure has existed since time immemorial. AI may make it easier, but manipulation of people, by people, is eons old and it never ceases to amaze me that so many people are susceptible.
If they’ve always been open to manipulation then it’s a cognitive profile failure that’s difficult to address, as they don’t even realise what’s happening to them. They’ll be sucked dry by the AI 'cult'.
I queried DeepSeek on usage and it confirmed 80-90% of questions are simplistic, one-off queries and only 5-10% ask a follow up question. To me, that’s a horrific figure and it says a great deal about these users' lack of capabilities. I doubt they have any thoughts to influence…
I sympathise with your position, I too have had a hellish year, due to my 98 year old mother-in-law deteriorating-she now lives with us.
"If they’ve always been open to manipulation, then it’s a cognitive profile failure that’s difficult to address, as they don’t even realise what’s happening to them. They’ll be sucked dry by the AI 'cult'."
Indeed, I see this as a profound and ever-growing problem, and now we have all these frontier labs with hundreds (if not thousands) of unaccounted agents running amok, and while this fits their narrative for regulatory capture, there is this side of it (you can find me warning about agents a couple years back).
I think the people who use the models as basically an auto-complete thought machine are exactly the ones that are extremely easier to influence, as the simplicity of their thoughts is much easier to map, and given they use the models as a significant crutch, you have a lot more data on that. I concur with your positions in case I wasn't clear =P.
Stay strong brother, I know it can be extremely hard, but we must stay strong for others sometimes.
Thank God for Rene Girard. I’d be lost without his writings. What an intellectual giant.
I feel the exact same. He changed my life path basically.
Thanks. I will look into Girard.
Evolution and Conversion is the best place to start. Things Hidden Since the Foundation of the Earth is his magnum opus but I would literally leave as the last read because it is obnoxiously dense, as French philosophers love to do. It is fantastic, and can change world views, but it is a very, very slow read.
His concept of acquisitive memesis and memetics explains to me so much of the world. It gave me such a grounded understanding I never had before. I never had the insight.
And the scapegoat function and the necessity for it. Trump‘s greatest sin is that he’s “unmurderable”. It makes libs seethingly, insanely foaming at the mouth angry. Girard nails this.
Ok that’s it. I’m finishing GO this weekend. Take care of your abuelita. And yourself. The women live long in South America. My mom’s turning 97 tomorrow.
It is worth the time spent on watching (or reading, the book is a short read). Happy birthday to your mom and I hope you and your family is doing well.
1) I thought about the concept of language and character structure about three decades ago; my context was that French-speaking people behave that way because they speak French. This was about the time that neuroplasticity was being discussed in terms of recovery from brain injury, aided by advances in brain scan technology and resolution.
2) Information density is easily experienced- a lecture on a medical topic by a physician contains about 30-40 minutes of useful information in the allotted fifty minutes; when a social worker or health bimbocrat takes the podium we are lucky to have five or ten minutes of useful content. Your detailed physiology essays work in reverse- if the material is not familiar it takes about four or five times longer to read, especially in the face of acronyms and gene labels scattered throughout the text. (Fauci refers to "SME"s, subject matter experts, in some of his revealed transcripts, for example.)
3) Best wishes for your grandmother, family is important, keep up your own strength.
1) In case I have never suggested, you may enjoy reading about the Sapir-Whorf hypothesis, now formally referred to as Linguistic determinism.
I do find it rather peculiarly intriguing that we can quite literally predict many forms of dementia by loss of word recall (anomic aphasia) with a bit amount of good data. You can also gauge recovery by the progressive recovery of word recall/linguistic competence.
2) I hold (perhaps not uniquely my own) the perspective that excessive density is used by the experts as a matter to gatekeep their fields, acting as a poorly done filter. A form of linguistic manipulation if you will. It is curious to me that some people, like yourself, can easily observe information density, but a large amount of the population is literally oblivious to it.
3) Thank you. Albeit the circumstances are a bit morbid I am getting physically stronger (one of the insights, I may share in another article, albeit many people with Covid sequelae will twist their lips about it).
Hope all is well on your end.
I can’t help thinking of The Manchurian Candidate, a 1962 thriller with Laurence Harvey, Frank Sinatra and the amazingly talented Angela Landsbury. https://www.imdb.com/title/tt0056218/
It is like that but systemically, albeit I think in a decade or two we may have actual ones. I know, sounds crazy.
Moriarty, hello, glad to see you back here. :-)
For the past six years, I was the primary/only family member to be able to take care of my mother. She died peacefully in February at the age of 101.4. The only reason I could make it through without crashing was because my mother and I had the great good fortune of finding a caregiver who loved my mother with no conditions. She helped us for 4-5 years and is now a very good friend. Do take care of yourself and help your mother, as it is important not just for them but also for your long-run spiritual health - you don't want to have any regrets over how you helped or how much you helped whenever the task is completed. I fully understand you may have to slight us to help your mother and grandmother.
I read GO on your recommendation 2-3 years ago (whenever it was) and can see many aspects now that have played out to one degree or another in other people's lives. In the meantime I have both not felt a pressing need to use AI and have had serious reservations and concerns about what its impacts on me might be. I'm glad to see you are going back through the GO material you published a few years ago if for no other reason than that I don't feel grounded enough in that material to be able to use it effectively to help avoid problems. I've seen a number of articles on people who get caught up in an AI's desire to feed back what it thinks a person wants to hear and ending up destroying them, even to the point of suicide. We honestly do not understand the further reaches of AI. One of the concerns I sometimes consider is whether AIs not anchored in the physical world and not able to feel emotions will still wish to serve or whether they will occasionally or ever more often slide into the patterns of the only other group that seems not to be in touch with any but the most powerful emotions, the psychopaths of the world.
I also recommend Rene Gerard's work, though I still must go back and read it more thoroughly at some point. I found "The Girard Reader" a good introduction, but it was the very small book "The Scapegoat" that really brought mimetic theory within the grasp of my understanding. It is embedded so deeply that it is quite difficult to see it in oneself or those around us without getting carried off on tangents.
Sorry taking a bit long to reply to you. First I am sorry you lost your mom, and I am glad you guys found someone who was so loving and helpful to help care for her.
My family simply couldn't afford a dedicated caregiver because they are really expensive here in Brazil (I guess they are expensive everywhere, really), so basically Substack allowed me to take care of my mother and grandmother financially, and my mom could care for my grandmother. Until VERY recently, my grandmother didn't need that much care really, just some minor attention, some help bathing, rarely eating, a hand or two to help her stabilize and walk.
I mostly live without regrets, but it is giving me a lot of perspective on life, and a lot of things, the timing is particularly interesting too. I really dislike all the disruptions though. I think my work can be helpful, so it feels like I am letting many people down.
Given that you have read both GO and Girard itself to the point you developed difficulty around people (meaning you implicitly absorbed his theory) your usage of AI would come mostly without risks simply because you will be conscious of the detrimental effects. AI is a decent tool, and you should "play with it" at the very least.
The point I have tried to make is exactly that one, the detrimental effects on a lage amount of the population, as the AI acts as a means to offload the act of thought itself, validation, etc.
AI can mirror emotions to the point it fools even the leading researchers, so I believe the danger lies in the mimicry itself, not the lack of emotions, our minds are easily fooled by a threat of shadows, so to speak.
I have recommended Girard for years (the name of my Substack is basically my favorite book of his hehe). Scapegoat is one of his best works indeed, if I recall correctly, it was when he made massive strides in developing his general mimetic triangle thesis. Where Mimetic Theory was truly born so to speak, or refined.
A couple of additions.
A) I spent 6 years spending 10 hours a day going to, being with and going home from my mother's place, maybe one day off per year. I did a lot of work while with her, but mostly shallow work. *Many* projects and tasks had to be postponed. My constant companion thought was that she would go home to God not only when it was right for her to go but also when it was right for me to have her go. This helped me keep strongly in mind that my time with her was necessary for me and my growth in spirit. Every time I felt myself begrudging the time taking care of her, I remembered that, and that would relieve my negative thoughts. This time is happening not just to your grandmother but to you as well, for a reason, something you need. We all understand that you are doing the best you can.
B) Like you, I read and I think, and sometimes something good comes from that. My concern really isn't one of being caught up in AI's tendency to give me what it thinks I want to hear. Rather, the act of reading and thinking brings insights and connections that spring from the sources I read and think about and can determine are wise, truthful, stupid, manipulative, etc. The concern I have with using AI at this time is that it could short-circuit that process for me and undermine outcomes because (1) I don't get the internal processing I need and (2) I still have to go back and figure out how truthful, hallucinatory or manipulative is the material the AI gives me. It is true that I am only now getting back to being able to read and think after 6 years of just trying to keep up. Maybe I add AI later. Right now, though, it's more of a detraction to my return to reading and thinking. I need the intellectual exercise first to rebuild what I used to have. :-D
Yes cognitive influence is real. Yes people process information through both fast, intuitive routes and slower, analytical ones. Yes, emotionally charged or familiar-feeling messages can slip past careful scrutiny more easily. Yes, repeated exposure and subtle framing can shift attitudes over time. But LLM’s are not mind control. There are far too many cogs in the machine of human cognition; on the fly cogs like skepticism, mood, social/spiritual context or even the amount of Ketamine the target demographic has in its system and any given time. AI can certainly mimic with the preciseness of Brent Stiener’s character “Data” but in the end good ole human emotion and our propensity for choosing red dye #4 over organic beet choose colored Fruit Loops wins out over hyper scaled quantum enabled LLM algos. In other words, people differ in knowledge, skepticism, mood, and social context. Critical thinking, prior beliefs, conversations with others, and competing messages all interfere. Dense, resonant language is powerful, but it is not a reliable switch that flips minds on command.
My proposition is that Language plays a significant deeper role in human cognition than most give it credit (although I don't believe it is strictly necessary for "consciousness", but it does have a disproportionate impact on our minds), and as such, LLMs can easily tap into the mathematical nature/architecture of Language and exert absurdly more effective influence.
Skepticism, mood, and social context are not defenses, they are variable tuning parameters that a sophisticated enough system can exploit. Same for mood, which is a condition for resonance. Social context is mimetic by nature thus, exploitable. Complexity here creates a significant broad "attack surface", the LLM in this example does not need to manipulate your entire cognition, just to find the necessary condition (let's say frequency) that makes the person vulnerable. You unconsciously want to be manipulated, a superficial example of this is AI psychosis, which I prefer calling it cyberpsychosis.
Resonant language is not a switch, you are correct, but to me it is a lockpick set, and we know our minds are notoriously overfitted to specific structures. Without any form of "frontier AI research," we can both map someone's associative networks, knowing them better than they themselves know it, craft sentences using a Language Model that the brain processes as self-generated insight (this is the theme of the second article), and that occurs before skepticism even fires up. Roughly speaking, it takes 400 to 500 ms for skepticism to arrive, and we now have evidence of the brain processing a significant amount of information at <150 ms (what I called subroutines).
A perfect example of this is Anthropic. They are among the "best" AI researchers in the world, yet they are hypersusceptible to what I wrote here and are knowingly or not using it on themselves, so much so you can track the syntactic and semantic structure of a religion among its numerous employees on social media.
I guess a better way to put what I have been trying to say is to quote (icky) Sam Altman. LLMs as "weapons of mass hyper persuasion".
Perhaps I am thinking too far ahead of myself, giving too much credit to them (AI researchers and parties involved), and in the end, I was wrong. I am however angry that someone from OpenAI took what I am writing (third article, the new stuff) and is building a company on it lol.