What a fantastic article. I got it from a link on sanityunleashed substack. I have always spoken out against AI but have never been able to articulate my thoughts as well as you do. Thank you for giving me something I can copy and share with others. I have always told people that judgement cannot be computerized and use two examples from my personal experience:
1. The game of bridge. Unlike chess where it is possible to program almost every possible position or scenario, in bridge there are far too many different situations in bidding and playing the cards that require context and judgement. They have been making computerized forms of bridge for many years but all have been quite mediocre in the quality of decisions and they have not shown any significant improvement over time.
2. Computerized electrocardiogram (ECG) interpretation. I spent over 40 years in the practice of cardiology and my group did all the ECG interpretations for a large community hospital in southern California. The hospital had to pay us out of their billing receipts. Eventually they got a computerized system that printed out a report but they were so inaccurate that they had to then pay us to "overread" the computer generated diagnosis in order to avoid lawsuits based on misdiagnosis. Every three or four years some medical equipment manufacturer would come out with a new system and they would try it again with the same result. The problem for the computer was that there are a wide varieties of different configurations in both normal and abnormal ECGs and the hospital couldn't take the risk of allowing the computer reading to be the official chart diagnosis. Much like the bridge example, they could make minor improvements here and there but ultimately reached the limit of their ability to accurately diagnose due to the amount of judgement that went into each reading.
You are very welcome. I like the bridge example because many peoples' response to pointing out the limitations of AI is "Well the computer can beat the world's chess champion". Unfortunately very few young people take up bridge anymore. But, for those who know the game, it really gets the message across. The ECG example works better with other doctors if they are familiar with graphic interpretation. I am very pleased to have discovered your website and am looking forward to reading your earlier articles.
Talking about games, you did not mention Go. I am a 4 kyu player and thoroughly enjoyed the documentary about the historic defeat of Lee Sedol by the AI program AlphaGo and the evolution to Alpha zero by a self-thought algorithm.
Thanks for the suggestions and the links to your articles. I had the same reaction as you to the Popping the AI Bubble article. I thought it was one of the best I had ever seen. I have now read your first link (AI is no match...) and think it ranks right up there with this one. I especially liked your redefinition of AI as Computational Intelligence or Algorhythmic Intelligence. Those terms really describe the limitations of the tool. I can tell from your references to Hayek, Kuhn, et al that we are like-minded and l look forward to reading more of your thoughts and descriptions.
Well done! You have the unusual ability to view a topic from every perspective and then eloquently yet succinctly describe each facet. And your analysis is always comprehensive and accurate. It's a pleasure to read your articles and experience the intelligence behind them. That's why I subscribe.
I completely share your critical viewpoint. I notice that AI searches on topics where I am an expert yield obvious erroneous results, driven by results deemed superior by programmers who lack the experience and knowledge to make informed decisions (or buy surreptitious influences aimed at profits and control. And yet, for searches on topics about which I know little, I readily accept whatever I read. And where previously I could skim through dozens of web results to form my own conclusions, the few results returned (by Google) are mostly those that support the information in the AI summary. I'm not sure people realize that AI doesn't just summarize information, it censors it and restricts it.
What a fantastic article. Right from the start, I recoiled, rebelled and refused to use the term AI because I knew it was just a bigger computer. Yet the hard sell has many people so transfixed that, as with so many things sold through the media, they just can't see it and I, as usual, are the one labelled odd. I deduced several years ago that the computers would likely share the woke non-morality of their programmers.
Also, silicon technology will never be capable of the same processing capacity, cost efficiency, resilience and memory density as human brains.
So building and maintaining a fleet of humanoid robots is likely to be more expensive and less effective than hiring people for general tasks for the foreseeable future.
The only thing they are good at is being bulletproof, so let's hope that in the future war is fought by machines and human settlements become demilitarised.
Additionally, I'd like to point out that while many grammar guides claim that this: "Quote," with the comma inside the quotes is the correct form, they are actually mistaken.
This is probably a typo that has accidentally been canonised. Since the text inside the quote marks is meant to be separate from the paragraph where it is mentioned, and thus has its own context, it doesn't make sense for a comma that references the outside context to be inside it.
To clarify with an example, what happens if we ask a question that ends in a quoted statement such as, "It is clearly absurd?"
Here it is obvious that the question mark should not be inside the quote as it completely changes the meaning of the quote and leaves the sentence without a terminating punctuation mark.
This is one of many significant mistakes that enormous numbers of people are making and have come to think is correct.
There are also problems with accountancy, programming, science, medicine, etc.
The recent decline in competency is truly staggering.
For a frightening window into just how far we have fallen, we can look at the short test created by binet simon in 1905 to identify children who had special needs.
According to them, every 10 year old should be able to correctly recall a 6 digit number heard once, and correctly reconstruct a design seen for 10 seconds.
Let me first endorse Brent Naseath's thanks at the top of the comments for your thought-expanding work, Lily.
The hype surrounding AI is fueled by the fact that it's tantalizingly a forever black box. Very likely there can never be a reliable human light source by which it can be illuminated. Here's Eric Schmidt, at the leading edge of AI development, painting an optimistically bright future for the AI winners. He barely mentions what may ironically become AI's (and our) undoing: its need for exponential supplies of energy to feed our insatiable craving to grasp the unobtainable; to know, and therefore control, everything - https://www.youtube.com/watch?v=8_2yFCm5sSM .
Absolutely brilliant and articulate analysis of what we refer to as artificial intelligence. Profoundly appropriate and essential information.
I'm including my response to a Substack post by Stephen Fry (AI: A Means to an End or A Means to Our End?/Sept. 15, 2024):
To the realistic and perceptive among us (yourself as well, I suspect, Stephen) it would be reasonable (if not obvious} to conclude that we have created a technology (monster?) that has already reached a point in its development and capabilities that is well beyond our ability to control. This is not, as you have already suggested, because there are not steps and strategies {potentially) available which could effectively control the reckless and dangerous proliferation of Ai in a manner that could ensure our safety from its ever-expanding depredations, but specifically because of our irrefutable and deeply- imbedded shortcomings as human societies. The social forces with the power, control and influence required to implement effective strategies to manage and utilize Ai safely, positively and beneficially are themselves fatally infected with a virus which will prevent them from marshalling the requisite cooperative and collaborative effort that could prove successful - and that virus is greed. Total, unmitigated, malignant, suicidal greed. By the time that humanity recognizes the urgency of the threat that improperly managed and controlled Ai ultimately poses - and belatedly, desperately seeks to regain control - it will sadly be far too late.
I wrote several posts about AI, you can fin in them many of the themes you are addressing here.
That is what I like about Substack. Realizing that the ideas are around us using us to give them life in various forms. This is the harmonic resonance of intelligence. 😉
I just started another series where I will address it again. There is no intelligence without life.
This is such a good top-down/bottom-up perspective that it is hard to quote individual lines that capture its overall essence. This was a pleasure to read and captured every key point and then some. Nicely done!
True. AI is he new thought police, ubiquitous, inescapable, a real Ministry of Truth from the novel 1984. It's already that now, and only needs widespread adoption to realize Orwell's nightmare.
This is by far the best post I read on Substack in the past two years. There is not a single sentence in it I would disagree with. I was so appalled by the AI hype that I wrote several posts on it, but nothin in my own post approached succinct elegance of this one. BRAVO!
An example that comes to mind: our dependency on GPS to navigate.
Using an old school paper map (or a good quality digital one) to get where you want to go requires a lot of frontal lobe functions - "executive functions" as they are called. You have to make a decision about your destination, then you have to explore alternative ways to get there, and choose the best one. Google maps telling you about traffic is useful - but depending on Google to tell you the best route is going to lead you into a pothole one day. If you choose your own path, you have a better idea of what to do if you run into a blocked bridge, or something else unexpected. You have an overview of the whole picture, not from reams of data, but from being able to extract the essential elements, and to decide what is most important.
Yes, I'm old school, and I grew up with paper maps... but the idea of taking direction from a disembodied voice has always struck me as quite sinister!
It won't just stop with GPS guidance though - eventually the car itself will be controlled by an unknown entity, and you will lose agency completely. And other aspects of your life will similarly be controlled and regulated.
But GPS guidance is something that we are being trained to follow blindly, have been now for many years, and I think it is just one of many preparatory steps.
Hi Lily, another too good text for another beautiful piece of crap.
We can also remember that AI is a tool of control, just like the “virtualization” of society. As you say, AI participates in the atomization of intelligence. And if we want to go further in THEIR delirium, it is a necessary step for the control of virtual money, and the remote control of objects and people.
I quote Günther Anders in "The Obsolescence of Man" written in 1956 (visionary) :
Anyone who collaborates in human engineering finds nothing morally shocking in his activity: he sees nothing there “fantastic”, nothing “impossible” and nothing dehumanizing.[...]
As for the criticism addressed to human engineering for being “dehumanizing”, for him it is simply inadmissible. Be a instrument is his most ardent desire, the task he has imposed. These robots that click in the “cartoons” do not are not grotesque or frightening figures in his eyes but the incarnation, pleasantly disguised as a bogeyman, of his ideal and the duty of compliance that he imposed on himself.
- I say “robots” and I mean “brought into line”. “Robots”, because the real robots of today are not these “computing machines" assembled from dead things which have human form, but instruments composed partly of living men. “Bringing into line”, because the process in question here constitutes a variant of known behavior in the political domain under the name of “bringing into line by the systems of domination”. The variant we are talking about is more extreme, to the extent where man here seeks to “self-reify”, while man politically placed walking (even one who is totally deprived of freedom) only becomes a "thing" metaphorically (Author's Note) -
I would add that solitude, or exclusion (as was the case with the containment of the false covid pandemic) is a tool for the system. A large majority of people can't stand it; it’s insidious, and psychologically alters people.
This is the beginning of intrusive control, what we call "biopower".
And, to somehow self-regulate, to support, humans have no other choice than the one proposed, to become the object.
What a fantastic article. I got it from a link on sanityunleashed substack. I have always spoken out against AI but have never been able to articulate my thoughts as well as you do. Thank you for giving me something I can copy and share with others. I have always told people that judgement cannot be computerized and use two examples from my personal experience:
1. The game of bridge. Unlike chess where it is possible to program almost every possible position or scenario, in bridge there are far too many different situations in bidding and playing the cards that require context and judgement. They have been making computerized forms of bridge for many years but all have been quite mediocre in the quality of decisions and they have not shown any significant improvement over time.
2. Computerized electrocardiogram (ECG) interpretation. I spent over 40 years in the practice of cardiology and my group did all the ECG interpretations for a large community hospital in southern California. The hospital had to pay us out of their billing receipts. Eventually they got a computerized system that printed out a report but they were so inaccurate that they had to then pay us to "overread" the computer generated diagnosis in order to avoid lawsuits based on misdiagnosis. Every three or four years some medical equipment manufacturer would come out with a new system and they would try it again with the same result. The problem for the computer was that there are a wide varieties of different configurations in both normal and abnormal ECGs and the hospital couldn't take the risk of allowing the computer reading to be the official chart diagnosis. Much like the bridge example, they could make minor improvements here and there but ultimately reached the limit of their ability to accurately diagnose due to the amount of judgement that went into each reading.
Thank you for the compliment and your support. Also thank you for sharing these examples. Compliments the article quite well. Going to pin this.
You are very welcome. I like the bridge example because many peoples' response to pointing out the limitations of AI is "Well the computer can beat the world's chess champion". Unfortunately very few young people take up bridge anymore. But, for those who know the game, it really gets the message across. The ECG example works better with other doctors if they are familiar with graphic interpretation. I am very pleased to have discovered your website and am looking forward to reading your earlier articles.
Talking about games, you did not mention Go. I am a 4 kyu player and thoroughly enjoyed the documentary about the historic defeat of Lee Sedol by the AI program AlphaGo and the evolution to Alpha zero by a self-thought algorithm.
In one of my many posts on AI (https://zorkthehun.substack.com/p/ai-is-no-match-for-natural-stupidity) I mention this with a link to the experiment where alpha zero was defeated by amateur players with unconventional strategies.
It is worth checking out.
as is this: https://slideslive.com/39006680/adversarial-policies-beat-superhuman-go-ais
Thanks for the suggestions and the links to your articles. I had the same reaction as you to the Popping the AI Bubble article. I thought it was one of the best I had ever seen. I have now read your first link (AI is no match...) and think it ranks right up there with this one. I especially liked your redefinition of AI as Computational Intelligence or Algorhythmic Intelligence. Those terms really describe the limitations of the tool. I can tell from your references to Hayek, Kuhn, et al that we are like-minded and l look forward to reading more of your thoughts and descriptions.
Well done! You have the unusual ability to view a topic from every perspective and then eloquently yet succinctly describe each facet. And your analysis is always comprehensive and accurate. It's a pleasure to read your articles and experience the intelligence behind them. That's why I subscribe.
I completely share your critical viewpoint. I notice that AI searches on topics where I am an expert yield obvious erroneous results, driven by results deemed superior by programmers who lack the experience and knowledge to make informed decisions (or buy surreptitious influences aimed at profits and control. And yet, for searches on topics about which I know little, I readily accept whatever I read. And where previously I could skim through dozens of web results to form my own conclusions, the few results returned (by Google) are mostly those that support the information in the AI summary. I'm not sure people realize that AI doesn't just summarize information, it censors it and restricts it.
An excellent and thought provoking article. Thank you for your insights.
What a fantastic article. Right from the start, I recoiled, rebelled and refused to use the term AI because I knew it was just a bigger computer. Yet the hard sell has many people so transfixed that, as with so many things sold through the media, they just can't see it and I, as usual, are the one labelled odd. I deduced several years ago that the computers would likely share the woke non-morality of their programmers.
Also, silicon technology will never be capable of the same processing capacity, cost efficiency, resilience and memory density as human brains.
So building and maintaining a fleet of humanoid robots is likely to be more expensive and less effective than hiring people for general tasks for the foreseeable future.
The only thing they are good at is being bulletproof, so let's hope that in the future war is fought by machines and human settlements become demilitarised.
Additionally, I'd like to point out that while many grammar guides claim that this: "Quote," with the comma inside the quotes is the correct form, they are actually mistaken.
This is probably a typo that has accidentally been canonised. Since the text inside the quote marks is meant to be separate from the paragraph where it is mentioned, and thus has its own context, it doesn't make sense for a comma that references the outside context to be inside it.
To clarify with an example, what happens if we ask a question that ends in a quoted statement such as, "It is clearly absurd?"
Here it is obvious that the question mark should not be inside the quote as it completely changes the meaning of the quote and leaves the sentence without a terminating punctuation mark.
This is one of many significant mistakes that enormous numbers of people are making and have come to think is correct.
There are also problems with accountancy, programming, science, medicine, etc.
The recent decline in competency is truly staggering.
For a frightening window into just how far we have fallen, we can look at the short test created by binet simon in 1905 to identify children who had special needs.
According to them, every 10 year old should be able to correctly recall a 6 digit number heard once, and correctly reconstruct a design seen for 10 seconds.
Let me first endorse Brent Naseath's thanks at the top of the comments for your thought-expanding work, Lily.
The hype surrounding AI is fueled by the fact that it's tantalizingly a forever black box. Very likely there can never be a reliable human light source by which it can be illuminated. Here's Eric Schmidt, at the leading edge of AI development, painting an optimistically bright future for the AI winners. He barely mentions what may ironically become AI's (and our) undoing: its need for exponential supplies of energy to feed our insatiable craving to grasp the unobtainable; to know, and therefore control, everything - https://www.youtube.com/watch?v=8_2yFCm5sSM .
Absolutely brilliant and articulate analysis of what we refer to as artificial intelligence. Profoundly appropriate and essential information.
I'm including my response to a Substack post by Stephen Fry (AI: A Means to an End or A Means to Our End?/Sept. 15, 2024):
To the realistic and perceptive among us (yourself as well, I suspect, Stephen) it would be reasonable (if not obvious} to conclude that we have created a technology (monster?) that has already reached a point in its development and capabilities that is well beyond our ability to control. This is not, as you have already suggested, because there are not steps and strategies {potentially) available which could effectively control the reckless and dangerous proliferation of Ai in a manner that could ensure our safety from its ever-expanding depredations, but specifically because of our irrefutable and deeply- imbedded shortcomings as human societies. The social forces with the power, control and influence required to implement effective strategies to manage and utilize Ai safely, positively and beneficially are themselves fatally infected with a virus which will prevent them from marshalling the requisite cooperative and collaborative effort that could prove successful - and that virus is greed. Total, unmitigated, malignant, suicidal greed. By the time that humanity recognizes the urgency of the threat that improperly managed and controlled Ai ultimately poses - and belatedly, desperately seeks to regain control - it will sadly be far too late.
Like a breath of fresh air. Thanks. This harari character strikes me as a sly heartless plagiarist . Functioning much like the modus operandi of "ai".
Good post!
YNH is a pompous fraud. I compared him to Humpty Dumpty in
https://zorkthehun.substack.com/p/the-not-so-smart-books-of-yuval-noah I also wrote several
I wrote several posts about AI, you can fin in them many of the themes you are addressing here.
That is what I like about Substack. Realizing that the ideas are around us using us to give them life in various forms. This is the harmonic resonance of intelligence. 😉
I just started another series where I will address it again. There is no intelligence without life.
This is such a good top-down/bottom-up perspective that it is hard to quote individual lines that capture its overall essence. This was a pleasure to read and captured every key point and then some. Nicely done!
True. AI is he new thought police, ubiquitous, inescapable, a real Ministry of Truth from the novel 1984. It's already that now, and only needs widespread adoption to realize Orwell's nightmare.
This is by far the best post I read on Substack in the past two years. There is not a single sentence in it I would disagree with. I was so appalled by the AI hype that I wrote several posts on it, but nothin in my own post approached succinct elegance of this one. BRAVO!
Great Article on the scourge of AI.
“If you can't dazzle them with brilliance, baffle them with bullshit.” - W.C. Fields
Because that is what AI is bullshit. Siri or Alexa will disagree with that statement- now you know.
An example that comes to mind: our dependency on GPS to navigate.
Using an old school paper map (or a good quality digital one) to get where you want to go requires a lot of frontal lobe functions - "executive functions" as they are called. You have to make a decision about your destination, then you have to explore alternative ways to get there, and choose the best one. Google maps telling you about traffic is useful - but depending on Google to tell you the best route is going to lead you into a pothole one day. If you choose your own path, you have a better idea of what to do if you run into a blocked bridge, or something else unexpected. You have an overview of the whole picture, not from reams of data, but from being able to extract the essential elements, and to decide what is most important.
Yes, I'm old school, and I grew up with paper maps... but the idea of taking direction from a disembodied voice has always struck me as quite sinister!
It won't just stop with GPS guidance though - eventually the car itself will be controlled by an unknown entity, and you will lose agency completely. And other aspects of your life will similarly be controlled and regulated.
But GPS guidance is something that we are being trained to follow blindly, have been now for many years, and I think it is just one of many preparatory steps.
Hi Lily, another too good text for another beautiful piece of crap.
We can also remember that AI is a tool of control, just like the “virtualization” of society. As you say, AI participates in the atomization of intelligence. And if we want to go further in THEIR delirium, it is a necessary step for the control of virtual money, and the remote control of objects and people.
I quote Günther Anders in "The Obsolescence of Man" written in 1956 (visionary) :
Anyone who collaborates in human engineering finds nothing morally shocking in his activity: he sees nothing there “fantastic”, nothing “impossible” and nothing dehumanizing.[...]
As for the criticism addressed to human engineering for being “dehumanizing”, for him it is simply inadmissible. Be a instrument is his most ardent desire, the task he has imposed. These robots that click in the “cartoons” do not are not grotesque or frightening figures in his eyes but the incarnation, pleasantly disguised as a bogeyman, of his ideal and the duty of compliance that he imposed on himself.
- I say “robots” and I mean “brought into line”. “Robots”, because the real robots of today are not these “computing machines" assembled from dead things which have human form, but instruments composed partly of living men. “Bringing into line”, because the process in question here constitutes a variant of known behavior in the political domain under the name of “bringing into line by the systems of domination”. The variant we are talking about is more extreme, to the extent where man here seeks to “self-reify”, while man politically placed walking (even one who is totally deprived of freedom) only becomes a "thing" metaphorically (Author's Note) -
I would add that solitude, or exclusion (as was the case with the containment of the false covid pandemic) is a tool for the system. A large majority of people can't stand it; it’s insidious, and psychologically alters people.
This is the beginning of intrusive control, what we call "biopower".
And, to somehow self-regulate, to support, humans have no other choice than the one proposed, to become the object.
Thank you Lilybit! You are an incredible writer. ✍🏼✨☀️🥰