The idea of “ad machinam” is fascinating. It reminds me that arguments should be judged by their merits, not by assumptions about who—or what—produced them. More importantly, your essay pushes the conversation back to language itself, which may be the deeper question.
Just an added remark or two about your nice Aeon essay yesterday...
FWIW, I'm pretty sure that a lot of the confusion and anger over AI stems from the fact that most of us in the general public are completely incompetent at thinking about thinking. We aren't prepared for it by education or custom and we're rarely rewarded for it with economic incentives. Worse, because we carry both personal and class-based sensitivities about our intellectual abilities, any serious discussion risks giving unintended personal offence.
Consequently, this crucial "AI moment" in history finds us broadly unwilling and unprepared to grapple with questions about what constitutes reason and creativity and how they ought to be evaluated and valued. It's an excruciating irony and unfortunately it coincides with a political crisis involving the roles of reasoning and evidence in public life.
I've always been surprised by the broad indifference to cognitive psychology. You could spend half a lifetime reading the painstaking demonstrations of how often and how badly even the brightest among us go wrong in our thinking. I was first led to that material when, as a youngster working in a neuroscience lab, I often saw my talented and eminent elders fight like badgers in a sack over the meanings of eccentric data points. How likely is it, I asked myself, that convoluted arguments over ill-defined philosophical terms ("Sein" and "Dasein", anybody?) are likely to be resolved satisfactorily when brilliant and disciplined minds can't make sense of simple empirical statements that have specific physical referents?
Many of my friends in the humanities still have not, I think, fully appreciated the seriousness of the problems in that domain. And to be even-handed about it, a lot of the folks I know in science and engineering are at sea in a different but no less disturbing way, confronting deep issues of consciousness and volition as if they were new problems arriving without hundreds of years of philosophical baggage.
It's a mess that's well beyond me to untangle. In a different era all this would only be of interest only to handful of specialists, but I'm very much afraid that the social, economic, and political pressures converging right now will end in real damage to real people.
Mike, it's so interesting to hear you ruminate on these experiences. I've found myself thinking much more about logical fallacies and cognitive prejudices than previously partly because the distortions you mention have become much more visible to me. It would be foolish to believe that we can disentangle them, but I do take heart in the fact that more teachers in school and in higher education are waking up for the importance of these things. I'm a believer in attempts to foster debate and argumentation training. That's what I'm currently working on. I'd be glad for any pointers you can give me.
Thanks for your kind reply, Martin. I'm flattered that you'd imagine I might have pointers for you but in truth I can't imagine a better path than the one you're already exploring with your writing course. Teaching students to make use of LLMs as critics and assistants seems like the best possible use of the technology for fostering intellectual humility and a collaborative spirit. Even in a place like Harvard there never enough bright, willing, and energetic human interlocutors to satisfy every need in every field.
My own recent vibe coding experience has been a wonderful personal reminder about the difficulty and rewards of exploring a new field. After years of idle curiosity about why students rise or fall in Oregon's public education system I used OpenAI's Codex to build an interactive dashboard for exploring school achievement test results. By linking published test results to census-derived socioeconomic info across the state I was able to look into how childrens' academic fates are associated with parental education levels, income, class sizes, school spending, and so on.
Hardly an original concept, but being able to engage deeply with the raw data without a tame statistician and a roomful of eager graduate assistants was a revelation. No data transformation, representation, or statistical experiment was out of practical reach. The limit was my own inventiveness and even that was amplified by collaboration with the very able model behind the scenes, GPT 5.5 xhigh. I've no doubt that all kinds of research will flower as these models mature and folks learn to exploit them.
Of course few stories are entirely positive and the outcome of mine is still uncertain. Though my explorations uncovered some interesting things, the response from the handful of educators and reporters I've contacted has ranged from indifference clear through to disbelief and even outright hostility. It simply hadn't occurred to me that what I did might be seen as a provocative challenge to established consensus instead of a simple product of curiosity. It would be criminal understatement to observe that this society has a complicated relationship with evidence about matters of public interest ;-)
Thanks again for your interest. You've been more than kind, and I have to add that I was impressed by your patient and open response to some shockingly hostile comments in that Aeon thread. If you are in fact a "technobarbarian" you're surely the very best kind imaginable.
Reading this and having a follow up conversation with a friend in China, it got me thinking about how a citizen's trust in the ability of their government to function well, and to effectively protect the public, may also play a role in A.I. hesitancy?
Regardless, I do think some 'control group' (i like that term better than creative resistance) who engage with A.I. less than or in a different way than others, should exist -- just to be able to see the forest slightly further away from the trees, as we leap into the future.
Patrick, a very judicious point. I think you're right that corroded trust in the government is fostering fear and paranoia. And I also like your idea about deliberate abstemiousness. I think there's good reason to think that parts of education should exist entirely without AI, and the logic can be extended to other areas. I agree with those who say that people of a certain age, who had the benefits of an "analog" education, are now in the best position to use AI without succumbing to its pernicious effects.
Many writers seem preoccupied with proving that AI cannot write. What interests me far more is that it can use language at all. The existence of generative AI may tell us less about machines than about the nature of language itself.
I am not a fan of blood sports so will avoid reading people’s attacks on your brave essay. All the teeth-gnashing is unnerving. What a relief when our country’s secure (read:tenured) thinkers hold firm and make space for innovation, progressive thinking, and using new tools with potential to increase access and inclusion. Indeed, iron clad security appears necessary to withstand the vitriol and accusations. Your essay’s title references Hamlet. Here’s wishing for a less tragic ending for the truth speaker.
Very kind of you to say (I also want to shout out to the underpaid content moderators who have been working overtime on various social media platforms to take down the more extreme responses.)
I'm certainly glad I have tenure, but I actually don't think that tenured professors have a better track record at speaking out on uncomfortable issues than people (both inside and outside universities) without such protections. Perhaps this is because you have to be a conformist to get tenure in the first place? So, a second shout out to people addressing uncomfortable issues without tenure.
Martin, similar accusations were made by readers regarding the Pope's encyclical against AI...that it must have been written by AI. So it goes both ways.
Good point. In one case (me), the logic seems to be: you are an AI stooge, so you must be using it. In the other (Pope): you’re secretly using it so you are a hypocrite.
Having said that, the analysis of the encyclical I have seen convinced me that at least select passages do bear all the tell-tale signs of AI. What does that mean? Probably that some assistant used it and that no one caught it. I don’t think the whole thing was written by AI. Above all it, doesn’t mean that its arguments are wrong.
Fascinating and useful for teachers in addition to writers and readers. Just wondering why no citation of the head full of elements graphic at the top of your sub stack article?
The idea of “ad machinam” is fascinating. It reminds me that arguments should be judged by their merits, not by assumptions about who—or what—produced them. More importantly, your essay pushes the conversation back to language itself, which may be the deeper question.
Just an added remark or two about your nice Aeon essay yesterday...
FWIW, I'm pretty sure that a lot of the confusion and anger over AI stems from the fact that most of us in the general public are completely incompetent at thinking about thinking. We aren't prepared for it by education or custom and we're rarely rewarded for it with economic incentives. Worse, because we carry both personal and class-based sensitivities about our intellectual abilities, any serious discussion risks giving unintended personal offence.
Consequently, this crucial "AI moment" in history finds us broadly unwilling and unprepared to grapple with questions about what constitutes reason and creativity and how they ought to be evaluated and valued. It's an excruciating irony and unfortunately it coincides with a political crisis involving the roles of reasoning and evidence in public life.
I've always been surprised by the broad indifference to cognitive psychology. You could spend half a lifetime reading the painstaking demonstrations of how often and how badly even the brightest among us go wrong in our thinking. I was first led to that material when, as a youngster working in a neuroscience lab, I often saw my talented and eminent elders fight like badgers in a sack over the meanings of eccentric data points. How likely is it, I asked myself, that convoluted arguments over ill-defined philosophical terms ("Sein" and "Dasein", anybody?) are likely to be resolved satisfactorily when brilliant and disciplined minds can't make sense of simple empirical statements that have specific physical referents?
Many of my friends in the humanities still have not, I think, fully appreciated the seriousness of the problems in that domain. And to be even-handed about it, a lot of the folks I know in science and engineering are at sea in a different but no less disturbing way, confronting deep issues of consciousness and volition as if they were new problems arriving without hundreds of years of philosophical baggage.
It's a mess that's well beyond me to untangle. In a different era all this would only be of interest only to handful of specialists, but I'm very much afraid that the social, economic, and political pressures converging right now will end in real damage to real people.
Mike, it's so interesting to hear you ruminate on these experiences. I've found myself thinking much more about logical fallacies and cognitive prejudices than previously partly because the distortions you mention have become much more visible to me. It would be foolish to believe that we can disentangle them, but I do take heart in the fact that more teachers in school and in higher education are waking up for the importance of these things. I'm a believer in attempts to foster debate and argumentation training. That's what I'm currently working on. I'd be glad for any pointers you can give me.
Thanks for your kind reply, Martin. I'm flattered that you'd imagine I might have pointers for you but in truth I can't imagine a better path than the one you're already exploring with your writing course. Teaching students to make use of LLMs as critics and assistants seems like the best possible use of the technology for fostering intellectual humility and a collaborative spirit. Even in a place like Harvard there never enough bright, willing, and energetic human interlocutors to satisfy every need in every field.
My own recent vibe coding experience has been a wonderful personal reminder about the difficulty and rewards of exploring a new field. After years of idle curiosity about why students rise or fall in Oregon's public education system I used OpenAI's Codex to build an interactive dashboard for exploring school achievement test results. By linking published test results to census-derived socioeconomic info across the state I was able to look into how childrens' academic fates are associated with parental education levels, income, class sizes, school spending, and so on.
Hardly an original concept, but being able to engage deeply with the raw data without a tame statistician and a roomful of eager graduate assistants was a revelation. No data transformation, representation, or statistical experiment was out of practical reach. The limit was my own inventiveness and even that was amplified by collaboration with the very able model behind the scenes, GPT 5.5 xhigh. I've no doubt that all kinds of research will flower as these models mature and folks learn to exploit them.
Of course few stories are entirely positive and the outcome of mine is still uncertain. Though my explorations uncovered some interesting things, the response from the handful of educators and reporters I've contacted has ranged from indifference clear through to disbelief and even outright hostility. It simply hadn't occurred to me that what I did might be seen as a provocative challenge to established consensus instead of a simple product of curiosity. It would be criminal understatement to observe that this society has a complicated relationship with evidence about matters of public interest ;-)
Thanks again for your interest. You've been more than kind, and I have to add that I was impressed by your patient and open response to some shockingly hostile comments in that Aeon thread. If you are in fact a "technobarbarian" you're surely the very best kind imaginable.
Reading this and having a follow up conversation with a friend in China, it got me thinking about how a citizen's trust in the ability of their government to function well, and to effectively protect the public, may also play a role in A.I. hesitancy?
Regardless, I do think some 'control group' (i like that term better than creative resistance) who engage with A.I. less than or in a different way than others, should exist -- just to be able to see the forest slightly further away from the trees, as we leap into the future.
Patrick, a very judicious point. I think you're right that corroded trust in the government is fostering fear and paranoia. And I also like your idea about deliberate abstemiousness. I think there's good reason to think that parts of education should exist entirely without AI, and the logic can be extended to other areas. I agree with those who say that people of a certain age, who had the benefits of an "analog" education, are now in the best position to use AI without succumbing to its pernicious effects.
Many writers seem preoccupied with proving that AI cannot write. What interests me far more is that it can use language at all. The existence of generative AI may tell us less about machines than about the nature of language itself.
I am not a fan of blood sports so will avoid reading people’s attacks on your brave essay. All the teeth-gnashing is unnerving. What a relief when our country’s secure (read:tenured) thinkers hold firm and make space for innovation, progressive thinking, and using new tools with potential to increase access and inclusion. Indeed, iron clad security appears necessary to withstand the vitriol and accusations. Your essay’s title references Hamlet. Here’s wishing for a less tragic ending for the truth speaker.
Very kind of you to say (I also want to shout out to the underpaid content moderators who have been working overtime on various social media platforms to take down the more extreme responses.)
I'm certainly glad I have tenure, but I actually don't think that tenured professors have a better track record at speaking out on uncomfortable issues than people (both inside and outside universities) without such protections. Perhaps this is because you have to be a conformist to get tenure in the first place? So, a second shout out to people addressing uncomfortable issues without tenure.
Martin, similar accusations were made by readers regarding the Pope's encyclical against AI...that it must have been written by AI. So it goes both ways.
https://www.reddit.com/r/slatestarcodex/comments/1toa872/claude_author_of_the_humanitas_evidence_that_the/
Good point. In one case (me), the logic seems to be: you are an AI stooge, so you must be using it. In the other (Pope): you’re secretly using it so you are a hypocrite.
Having said that, the analysis of the encyclical I have seen convinced me that at least select passages do bear all the tell-tale signs of AI. What does that mean? Probably that some assistant used it and that no one caught it. I don’t think the whole thing was written by AI. Above all it, doesn’t mean that its arguments are wrong.
Fascinating and useful for teachers in addition to writers and readers. Just wondering why no citation of the head full of elements graphic at the top of your sub stack article?