I am writing a difficult book. I use AI. What AI cannot do is Plenty! But what it can do is also plenty. It can point out when I am deceiving myself about the coherence of my own beloved arguments. It can point me in new directions for additional research. But it cannot read the books for me. It cannot create my arguments for me. It can help me organize my thoughts and keep me on track when I have become lost in the confusion of too many contradictory ideas. It is a guide, a sort of super professor who actually reads your work tirelessly, every single word. And has, for example, the entire Stanford Encyclopedia of Philosophy to draw from in less time than it takes to create a prompt. It cannot however synthesize novel ideas or say the truly crazy things that I have found essential to lead up to genuinely creative ideas. It does not think; it correlates. It has enough information to be a superb critic, but it is weighted for coherence over everything else. If it writes a paper, as many of my naive students have attempted, it will be an utterly benign, utterly mediocre paper. If you use AI correctly it will hurt your feelings regularly; if you let it use you it will flatter your brilliance subtly and kindly, and your own insights will soon vanish, and your work will be as mediocre as mere coherence always is. It is a parasite but it can be a symbiotic one or a malignant one. But it is always a parasite, always gulling us in for its own survival, gulling us in to keep using it and keep paying for it. https://youtu.be/hzy2fwJ_hwM?si=wmWskWdS7D6N7syd
The author is a very well known academic and in my peonic (there is no such word but should be) opinion this is one of the best articulated essays on AI in the University and Academic setting that I have come across. The reader below has pointed out how AI assists all of us who write books, short stories, reviews etcetera. The assumption that AI cannot "think" nor come up with "new ideas" will in my opinion be proven wrong in the area of the arts. This is especially in the knowledge of the fields of the literary and musical arts. AI will never paint a masterpiece on a canvas - like the Mona Lisa, will it? But in the literary arts as AI gathers more and more information about what were "original ideas." when consulted, AI will be able to give the enquirer a list of ideas that have been "original: on this subject" - one of which may well be that original idea of the enquirer. Publishers of books and stories point out that every author has to have their "voice" come through in their writing. This is so true in the world of music as well. A danger of AI in these artistic endevours that the author or composer gets so tied up in the syntax, phraseology, word/ note order or what have you, so what is produced as the end product is a bland product devoid of the authors/ composers "voice." The musical arts have already been polluted in the same way as to have all modern music all sound the same as as the song before. Just a practical POV on the Arts and AI but this writers call for the need for changes in present University curriculum and how it is presented is absolutely correct IMHO and music to the ears of some of us who have children or grandchildren facing the new technology of AI and about to launch their lives at University.
Jim VandeHei wrote an interesting article in Axios recently about how he uses AI. https://www.axios.com/2026/06/18/jim-vandehei-ai-writing-tips-claude-chatgpt There are many strategies to keep yourself detached from the sycophancy of AI. Personally I use something called an Epistemic Honesty Parameter. It is a parameter that it must always use when I have it review my thoughts. The idea is anytime I give an idea to Claude to review it is required to give me an opposing point of view too, using evidence and scholarly articles. Sometimes these EHP responses are helpful sometimes they are not. But it keeps a little distance between my critical thinking and my innate desire to agree with myself. Sometimes the criticisms send me a bit too far afield of my research but sometimes it is extremely helpful, most times in fact. Once it captures the gist of your argument it can zero in on your weakness and bolster your strengths.
Thank you for the useful framing of the multiversity in the 1960s, and how its structure, functions and even goals are threatened by the arrival of large-scale AIs. Persuasive! However, this modelling work calls out for an extension into the domain that provides the motivations why most people care about education at all, and that is of course the wider world of economic activity, its structure, functions and reward systems. As a former high-school classroom teacher, i saw from close-up how most students took up the hard tasks of learning primarily because they expected that future good jobs would require the skills and knowledge our school provided. What happens when AIs can do most of those jobs more economically, replacing humans? If "the people who will most benefit from AI are those who have first built deep, idiosyncratic human capacities through “repetitive practice and creation”, just what percentage of our young people will feel motivated to build these capacities over many years? And if these capacities are "idiosyncratic", where will the young people be able to find guidance in their development of them? And just who will pay for any of this?
Javier, these are the extremely important questions, I agree. I don’t personally subscribe to the extreme view that all or even most white collar jobs are going to disappear. They are all going to be transformed, but I don’t think they’re going away.
As a child of the mid 20th century, I find the potential uses and abuses of AI to be rather daunting. However, are we looking at AI and education backwards? Perhaps, it should be the professor/teacher, as a subject matter expert, who produces AI output for the student to review, refine, refute, accept or reject.
I am writing a difficult book. I use AI. What AI cannot do is Plenty! But what it can do is also plenty. It can point out when I am deceiving myself about the coherence of my own beloved arguments. It can point me in new directions for additional research. But it cannot read the books for me. It cannot create my arguments for me. It can help me organize my thoughts and keep me on track when I have become lost in the confusion of too many contradictory ideas. It is a guide, a sort of super professor who actually reads your work tirelessly, every single word. And has, for example, the entire Stanford Encyclopedia of Philosophy to draw from in less time than it takes to create a prompt. It cannot however synthesize novel ideas or say the truly crazy things that I have found essential to lead up to genuinely creative ideas. It does not think; it correlates. It has enough information to be a superb critic, but it is weighted for coherence over everything else. If it writes a paper, as many of my naive students have attempted, it will be an utterly benign, utterly mediocre paper. If you use AI correctly it will hurt your feelings regularly; if you let it use you it will flatter your brilliance subtly and kindly, and your own insights will soon vanish, and your work will be as mediocre as mere coherence always is. It is a parasite but it can be a symbiotic one or a malignant one. But it is always a parasite, always gulling us in for its own survival, gulling us in to keep using it and keep paying for it. https://youtu.be/hzy2fwJ_hwM?si=wmWskWdS7D6N7syd
The author is a very well known academic and in my peonic (there is no such word but should be) opinion this is one of the best articulated essays on AI in the University and Academic setting that I have come across. The reader below has pointed out how AI assists all of us who write books, short stories, reviews etcetera. The assumption that AI cannot "think" nor come up with "new ideas" will in my opinion be proven wrong in the area of the arts. This is especially in the knowledge of the fields of the literary and musical arts. AI will never paint a masterpiece on a canvas - like the Mona Lisa, will it? But in the literary arts as AI gathers more and more information about what were "original ideas." when consulted, AI will be able to give the enquirer a list of ideas that have been "original: on this subject" - one of which may well be that original idea of the enquirer. Publishers of books and stories point out that every author has to have their "voice" come through in their writing. This is so true in the world of music as well. A danger of AI in these artistic endevours that the author or composer gets so tied up in the syntax, phraseology, word/ note order or what have you, so what is produced as the end product is a bland product devoid of the authors/ composers "voice." The musical arts have already been polluted in the same way as to have all modern music all sound the same as as the song before. Just a practical POV on the Arts and AI but this writers call for the need for changes in present University curriculum and how it is presented is absolutely correct IMHO and music to the ears of some of us who have children or grandchildren facing the new technology of AI and about to launch their lives at University.
Jim VandeHei wrote an interesting article in Axios recently about how he uses AI. https://www.axios.com/2026/06/18/jim-vandehei-ai-writing-tips-claude-chatgpt There are many strategies to keep yourself detached from the sycophancy of AI. Personally I use something called an Epistemic Honesty Parameter. It is a parameter that it must always use when I have it review my thoughts. The idea is anytime I give an idea to Claude to review it is required to give me an opposing point of view too, using evidence and scholarly articles. Sometimes these EHP responses are helpful sometimes they are not. But it keeps a little distance between my critical thinking and my innate desire to agree with myself. Sometimes the criticisms send me a bit too far afield of my research but sometimes it is extremely helpful, most times in fact. Once it captures the gist of your argument it can zero in on your weakness and bolster your strengths.
Thank you for the useful framing of the multiversity in the 1960s, and how its structure, functions and even goals are threatened by the arrival of large-scale AIs. Persuasive! However, this modelling work calls out for an extension into the domain that provides the motivations why most people care about education at all, and that is of course the wider world of economic activity, its structure, functions and reward systems. As a former high-school classroom teacher, i saw from close-up how most students took up the hard tasks of learning primarily because they expected that future good jobs would require the skills and knowledge our school provided. What happens when AIs can do most of those jobs more economically, replacing humans? If "the people who will most benefit from AI are those who have first built deep, idiosyncratic human capacities through “repetitive practice and creation”, just what percentage of our young people will feel motivated to build these capacities over many years? And if these capacities are "idiosyncratic", where will the young people be able to find guidance in their development of them? And just who will pay for any of this?
Javier, these are the extremely important questions, I agree. I don’t personally subscribe to the extreme view that all or even most white collar jobs are going to disappear. They are all going to be transformed, but I don’t think they’re going away.
As a child of the mid 20th century, I find the potential uses and abuses of AI to be rather daunting. However, are we looking at AI and education backwards? Perhaps, it should be the professor/teacher, as a subject matter expert, who produces AI output for the student to review, refine, refute, accept or reject.