Got a story. The other day I can back from some exercise and made a protein shake. Took a moment to consider why. It's a routine from when I trained more and at the time it was worth doing to improve nutrition and recovery. On this specific day, I just went to the gym to move my body. But the habit of the shake remain, even through it likely did not contribute much. I say this because I've been eating and sleeping lousy and those are the first things to address. But this advanced optimization remained.
I remembered that I don't listen to music when I run. Because for me it is a crutch. If the music is hype, I run too fast. If it's the wrong song late in a run, I stop and maybe switch the song. Either way, it keeps me from pushing me. Pushing me that day and pushing me in general to grow.
AI is the ultimate crutch for me. It's pumping out exactly what I intended, I'm hype. It's missing the bit, I give up. The common denominator is that it's one more thing to keep me from thinking harder, pushing myself, basking in the discomfort. And I think that is the problem with tools. The more you lean on them, the more you atrophy in that realm.
There's something about friction that switches on the learning part of the brain. Education, more generally, growth is all about the right amount of friction or stimuli to promote growth. In the past, the friction was too high and on average people did not grow. But as the friction drops, the education doesn't proportionally increase. Because we like our crutches and cling to them.
To play devil's advocate, I use the socratic method in my chats and get a lot out of it, that I otherwise wouldn't have bothered to learn.
AI merely exposes the inherent flaws in how we do education (exams, especially failure) - which I personally and honestly don't have a solution for. Both the carrot and the stick rip out some of the gold that is human potential (while instilling some useful skills and behavior).
The knee-jerk reaction of banning AI is probably the wrong choice, people need to learn where the limits are, but I also wouldn't be able to suggest a framework where students aren't rewarded for skirting around being educated.
I’m also for banning it in the sphere of education. First off for much of education the student is not on a quest to discover novel things. But rather to learn known concepts. Meaning the answer is not the point. The exploration of the student is the point. Having shortcut machines to the answer are detrimental.
Secondly it’s unclear whether shortcuts to many “answers” are wholly beneficial in many cases outside the academic sphere. Hence scandals when for example video educators use AI in their material. Or the general push back on AI writing, website design, etc. These answers are being pushed back on.
But why would there be any difference between a teacher giving you the shortcut or AI giving you the shortcut?
AI can give you 1000s of exercises and explain in great detail exactly how to solve them, and what you're doing wrong, as well as where you suck, explore other ways of solving and where you should improve. Learning to integrate or solve differential equations with Gemini is 10x better than even the good teachers I've had. It beats bad teachers because it's just simply better than them, and it really increases what you can learn from good teachers.
And the same goes for critiquing writing, whether code or prose. I do prose like 1/10th of the time, code much more often, but the problem with my writing is usually relatively basic grammar and consistent style problems. Also using codex with "I've written X, break my code" usually gives ... I should say usually more insight than I was ready for.
In the case you described you aren't using AI as a shortcut to an answer but rather to emulate what a teacher would. If you were to then use AI to answer the questions it brings up for you you would have learned nothing in the process. Its like asking your teacher to solve the problems for you. You have an answer but also no growth.
But yes one could argue that using AI not as a means for answering a class's problems but rather to not generate more exercises could be valid. Though I doubt students have a lack of practice problems these days.
I find AI is a bad communicator. So it is quite awkward to take advice on prose from a system that generates bad prose, though that could be possible. Still it is unclear, one would have to judge what benefits arise.
Now my earlier point was not that AI is not useful in any way. It is useful with code. Though there is the larger conversation on developer enjoyment etc in the age of AI coding--where the speed of AI generated code is expected, and developers are more alienated from aspects that made coding enjoyable.
People who engage with problems, think about them, reason about them, can ask AI AFTER that work, to get fruitful results - as in: increase insight. I'll have to find the study I'm referencing (might be a second)...
The issue is that isn't anyone's default mode of interaction with these systems...
Some teachers ban calculator use in their tests, some explicitly check if your work is derived from answers you did not come up with (something you found on the internet etc.). Unless a class takes account of some tool, using it could be to the detriment of the thing you are there for—namely an education.
Seriously? Idk about you but we weren't allowed to use calculators until long after we had already learned how to do whatever the calculator did by hand, except in cases where that was never the point of the lesson to begin with (like calculating cosines). And even then, sparingly. The reason they eventually let people use it was that it had nothing to do with the lesson being taught, and it was a waste of time at that point to make us do arithmetic by hand. Not because it suddenly became OK to turn off your brain and become lazy.
Skipping to learn long division by hand because you'll always have a calculator (or better a smartphone at hand) is like going to a gym with a forklift. It is faster to lift the weights, but it kinda defeats the purpose.
Learning to do long division by hand trains the mind. It prepares you to understand polynomial division. It teaches you the deep connection of division and subtraction. So, mastering it should be part of the curriculum. As soon as it is truly mastered, it should be allowed to use a calculator.
I do, in fact, not always have a calculator on my person and sometimes need to do basic arithmetic by hand or in my head. The teachers who told us that were 100% correct.
Most people structure their lives in such a way to always have that calculator with them if they expect to need it, or failing that, arrange to postpone the division so they can do it when they have the calculator again.
Difference is AI involves offloading most thinking. AI may be cool for already intelligent, but it is not going to form intelligent brains leaning on that from a young age.
I don't mean "AI" like generating quizzes or whatever. But people fooling themselves that LLMs are a "bicycle for the mind" with little kids are delusional.
Seriously. Experimenting on children should be considered a crime against humanity, these people are sick and always show how they will betray us with their anti-human beliefs.
I am being level headed, I do not think it's good to let advertising companies or for-profit entities that are based on not getting consent.
We have seen how damaging these tech companies already are to society and children.
Why should we continue trusting an industry that has shown no inclination to well being of children and not merely trying to pursue more profits in the form of corporate welfare.
Absolutely not the same at all and shame on you for saying so.
The biggest tell is how none of these tech executives ever subject their own children to these supposedly "great tools." Because they don't help children, they're made to extract profits by whatever harms as necessary.
The push for AI adoption in K-8 is the one that strikes me as wrong. The Alpha School and other initiatives are pushing AI on groups for which we have dubious measures of impact. Ostensibly, they are doing this to avoid pushback from traditional curriculum requirements in high school and university - but those kids are going to end up there eventually.
Why would I believe that someone who can't use a mouse/computer is ready to be taught by AI?
> Why would I believe that someone who can't use a mouse/computer is ready to be taught by AI?
The human element is critical for teaching, in my opinion. (Great) teachers are able to smooth over some of my criticisms of education. I also believe we are very far off from it being able to do even a bad job at teaching. So I'm not suggesting that at all.
What I'm talking about is teaching students how to use it, especially when not to use it, and how to get better results from it. I don't think AI is going to replace us any time soon, but people who use AI will replace those who don't.
> AI merely exposes the inherent flaws in how we do education ... The knee-jerk reaction of banning AI is probably the wrong choice, people need to learn where the limits are
But you can't do that without a certain level of maturity.
> but I also wouldn't be able to suggest a framework where students aren't rewarded for skirting around being educated.
Well, then "knee-jerk" ban AI in schools... until you can suggest such a framework, and test it to show it works, and roll it out at a society-wide level.
Until then, you're just handing kids flamethrowers in a highly flammable environment, and saying "we'll just figure out the problems later."
I reckon it's a similar issue to wikipedia when that became a resource available to us in school. Being taught how to use AI as a resource as opposed to 'the source' is the critical distinction. Which I guess is also the same issue when it comes to professionals and how they use AI
Ultimately it's the same issue across all technological tools. If the tools (AI, wikipedia, calculators, etc) can be used to complete work (or homework) without understanding the underlying fundamental concepts of the work being done, then you're hurting your education and/or professional development
> AI merely exposes the inherent flaws in how we do education (exams, especially failure) [...] but I also wouldn't be able to suggest a framework where students aren't rewarded for skirting around being educated
As someone who graduated from college in the 2000s, I've been surprised to hear from recently-graduated colleagues that doing exams in person is no longer the default. That strikes me as the obvious first line solution for (at least) the STEM subjects that lend themselves to exams centered around technical problem-solving: exams in person, no phones or similarly capable devices, cheaters get a failing grade with no possibility for appeal.
The conventional wisdom when I was in school was that classes where homework isn't graded aren't really any easier for most students, because the homework is where you learn the skills, and if you don't do it then you'll fail the exams. Frankly, the same will be true either way for students who use LLMs for all the exercises: they won't learn the skills, and they will fail the exams. This is a working-as-intended outcome.
So why not start there? It's how school has ~always worked, so there's no question of feasibility. The only question is whether our will to give committed students a real education is stronger than our desire to productize education as a rent extraction scheme tailored to a nepotistic and shallowly credentials-obsessed society where having 'B.Sc.' on your resume sorts you one bucket higher in the algorithmic hiring queue to flip burgers at your local McDonald's.
---
It's less obvious how to fix the humanities subjects where take-home writing assignments carry most of a student's grade, though I'm tempted to think that even a relatively simple and tightly scoped writing exercise given as an in-person exam could conclusively weed out most of the students who've been relying on LLMs to do the assignments for them.
The big problem with education is that all the work is done for grades instead of aiming to deliver value to others. It's important for students to practice but if you raise the bar high enough then AI is just one tool that helps you but doesn't have all the answers. If AI can do all the homework we have to ask ourselves how much value the homework had in the first place.
The word 'graduation', regarding education, somehow lost its original meaning -- to measure peers against one another -- its sinister edge can be easily visualized by imagining a class of graduates stuffed into giant graduated cylinders clearly showing their achievements with etched gradations along their respective tubes. It is difficult for society to consider this process as the, or major contributor to the, problem of education rather than AI. I tend toward the belief that we place too much value upon these highly dubious and artificial systems of academic assessment; AI is providing a long overdue opportunity to evaluate the excesses and weaknesses of institutional education.
Haidt is one of those political conservatives who calls themselves a centrist or moderate. He and the host David Brooks are both in the Epstein files, the latter described in the files as one of the biggest supporters of the Epstein backed EDGE foundation (with Haidt name dropped in the same email).
Speaking personally, YMMV, but I hear either name and I'm immediately skeptical.
what would be wrong with using AI only in /learning mode? I've used it this way before and found it even more useful than my memory of some teachers I had. If it was locked to this mode and you could not just generate answers I don't see the harm.
> I want to understand this more deeply. So instead of you just telling me
everything, lay out the problem and a question for me to think about, and
I'll try to answer. Even if I answer wrong, you should resist giving me the
answer, and instead keep digging with more questions, so that I eventually
arrive at the answer myself.
I use this prompt quite a bit. You get a lot of mileage out of it. But I have a pretty good idea of what I don't understand and am very focused on what I'm after. If you don't know what you don't know, then it's a mess. Even more if you don't know how to express what you're trying to understand.
In education, meaning primary and secondary, a narrative is used to provide motivation and create an interest and curiosity for the subject. I already have that, so it's much easier than digging through textbooks. The AI doesn't do that now.
AI in learning mode has the same shortcomings as virtual classroom. The physical presence of a teacher in the room. Sets the tone.
If we had AI when I was a kid, and you sat me down in front of a computer with AI, and told me to learn about sub-Saharan Africa and write a report manually, of course I would try to automate it. But it would be so much easier for me to learn with, because I have a learning disability, and my own style of learning new information.
Academia relies on teaching everyone the same way, which advantages some kids and disadvantages others. AI could potentially help more kids. To say nothing of increasing access to education, which so often depends on how wealthy the school district is.
>The criticism cited in Wikipedia is totally unsubstantiated.
???
Odgers, Candice L. (March 29, 2024). "The great rewiring: is social media really behind an epidemic of teenage mental illness?". Nature. 628 (8006): 29–30. Bibcode:2024Natur.628...29O. doi:10.1038/d41586-024-00902-2. Archived from the original on September 10, 2024. Retrieved April 21, 2024.
And if you go to the article that has citations as well.
>While I doubt it's a boost in anything but productivity
If you look at the last chart it shows that students that spend the greatest amount of time using AI gets higher exam scores. It's not just "completes homework quicker".
>the negatives are simply too great to ignore.
That's... exactly what I said? I'm not sure how anyone can read "For most kids, AI is a detriment, but for a minority (smartest? most self motivated?) it's a boost" and get the impression that the negatives are fine because a tiny minority are getting better learning.
As someone who's read all of his books and experienced the generations being studied I'd say he's been spot on and is extremely knowledgeable. One study that does not propose alternative views and cites "selling a lot of books" comes off to me as someone with an agenda. No thanks.
>One study that does not propose alternative views
Why is proposing "alternative views" required for a critique? If we're discussing what caused diseases pre-germ theory, does that mean we can't point out the flaws in miasma theory because no "alternative views"?
>and cites "selling a lot of books" comes off to me as someone with an agenda. No thanks.
Well clearly you only read the first two sentences (or first paragraph?), because it cites multiple studies:
>These are not just our data or my opinion. Several meta-analyses and systematic reviews converge on the same message2–5. An analysis done in 72 countries shows no consistent or measurable associations between well-being and the roll-out of social media globally6. Moreover, findings from the Adolescent Brain Cognitive Development study, the largest long-term study of adolescent brain development in the United States, has found no evidence of drastic changes associated with digital-technology use7. Haidt, a social psychologist at New York University, is a gifted storyteller, but his tale is currently one searching for evidence.
Not to mention "selling a lot of books" isn't even the critique, it's that it offers a seductive explanation:
"... because Jonathan Haidt is telling a scary story about children’s development that many parents are primed to believe"
> For most kids, AI is a detriment, but for a minority (smartest? most self motivated?) it's a boost.
Exactly. The educational system needs to be reformed to exalt and celebrate the minority boosted AI, and quickly shuttle those harmed into the trash. Our North Star needs to be AI technology and the returns of investors who support it. Everything else, including human beings, are unimportant and expendable. Tech first!
>Exactly. The educational system needs to be reformed to exalt and celebrate the minority boosted AI, and quickly shuttle those harmed into the trash.
I'm not sure how you got the impression that not being maximally rabid against AI (ie. recognizing there are benefits) implies I'm sort of AI bro that is willing to sacrifice the education of 80% (or whatever) of kids into the metaphorical woodchipper just so the remaining can get slightly better exam scores.
Haidt is first on my personal list of "people who don't really know what they're talking about but who I nonetheless mostly agree with." The "If Books Could Kill" show did a good episode about some of his works that indicate he has a fairly nonrigorous methodology and implements some iffy statistical controls to support what is probably a pre-conceived viewpoint. He doesn't have the best reputation in the academic space he operates within.
That being said, this stuff is hard to study, and I have no doubt that things like social media and AI are bad for a developing brain even if you can't arrive at a statistically satisfying conclusion. I can tell that this shit is bad for me, and I'm a fully grown man.
Wikipedia is fair to cite Odgers's review. But the nature of the available evidence means that Odgers would be overstating her case to say that Haidt's theory is conclusively wrong, and Haidt would be overstating his case if he claimed to be conclusively right.
You see, the effect on social media on socialization has two parts. The first is the direct effect on an individual caused by their own use of social media. The second is the effect on an individual's social environment when everyone around them is on social media. The first effect is difficult but basically feasible to study. The second effect is downright wicked. So dismissing evidence that hints at the second effect as being "mere correlation that disappears under scrutiny" is not valid reasoning when your process of scrutiny focuses on the first effect.
In social science you have to accept reasoning from incomplete data. You won't always have the answers. I don't think Haidt really stands above or below his peers in any category except notoriety. Of course anyone who has worked with scientists knows to be a little suspicious of a researcher who seems to enjoy being in the public eye. But as the public eye, we're still stuck with them.
Got a story. The other day I can back from some exercise and made a protein shake. Took a moment to consider why. It's a routine from when I trained more and at the time it was worth doing to improve nutrition and recovery. On this specific day, I just went to the gym to move my body. But the habit of the shake remain, even through it likely did not contribute much. I say this because I've been eating and sleeping lousy and those are the first things to address. But this advanced optimization remained.
I remembered that I don't listen to music when I run. Because for me it is a crutch. If the music is hype, I run too fast. If it's the wrong song late in a run, I stop and maybe switch the song. Either way, it keeps me from pushing me. Pushing me that day and pushing me in general to grow.
AI is the ultimate crutch for me. It's pumping out exactly what I intended, I'm hype. It's missing the bit, I give up. The common denominator is that it's one more thing to keep me from thinking harder, pushing myself, basking in the discomfort. And I think that is the problem with tools. The more you lean on them, the more you atrophy in that realm.
There's something about friction that switches on the learning part of the brain. Education, more generally, growth is all about the right amount of friction or stimuli to promote growth. In the past, the friction was too high and on average people did not grow. But as the friction drops, the education doesn't proportionally increase. Because we like our crutches and cling to them.
To play devil's advocate, I use the socratic method in my chats and get a lot out of it, that I otherwise wouldn't have bothered to learn.
AI merely exposes the inherent flaws in how we do education (exams, especially failure) - which I personally and honestly don't have a solution for. Both the carrot and the stick rip out some of the gold that is human potential (while instilling some useful skills and behavior).
The knee-jerk reaction of banning AI is probably the wrong choice, people need to learn where the limits are, but I also wouldn't be able to suggest a framework where students aren't rewarded for skirting around being educated.
This is objectively false. AI stunts and reduces the learning process. Banning it is 100% the right move.
I’m also for banning it in the sphere of education. First off for much of education the student is not on a quest to discover novel things. But rather to learn known concepts. Meaning the answer is not the point. The exploration of the student is the point. Having shortcut machines to the answer are detrimental.
Secondly it’s unclear whether shortcuts to many “answers” are wholly beneficial in many cases outside the academic sphere. Hence scandals when for example video educators use AI in their material. Or the general push back on AI writing, website design, etc. These answers are being pushed back on.
But why would there be any difference between a teacher giving you the shortcut or AI giving you the shortcut?
AI can give you 1000s of exercises and explain in great detail exactly how to solve them, and what you're doing wrong, as well as where you suck, explore other ways of solving and where you should improve. Learning to integrate or solve differential equations with Gemini is 10x better than even the good teachers I've had. It beats bad teachers because it's just simply better than them, and it really increases what you can learn from good teachers.
And the same goes for critiquing writing, whether code or prose. I do prose like 1/10th of the time, code much more often, but the problem with my writing is usually relatively basic grammar and consistent style problems. Also using codex with "I've written X, break my code" usually gives ... I should say usually more insight than I was ready for.
In the case you described you aren't using AI as a shortcut to an answer but rather to emulate what a teacher would. If you were to then use AI to answer the questions it brings up for you you would have learned nothing in the process. Its like asking your teacher to solve the problems for you. You have an answer but also no growth.
But yes one could argue that using AI not as a means for answering a class's problems but rather to not generate more exercises could be valid. Though I doubt students have a lack of practice problems these days.
I find AI is a bad communicator. So it is quite awkward to take advice on prose from a system that generates bad prose, though that could be possible. Still it is unclear, one would have to judge what benefits arise.
Now my earlier point was not that AI is not useful in any way. It is useful with code. Though there is the larger conversation on developer enjoyment etc in the age of AI coding--where the speed of AI generated code is expected, and developers are more alienated from aspects that made coding enjoyable.
> AI stunts and reduces the learning process
Mostly agree, but there is nuance to it.
People who engage with problems, think about them, reason about them, can ask AI AFTER that work, to get fruitful results - as in: increase insight. I'll have to find the study I'm referencing (might be a second)...
The issue is that isn't anyone's default mode of interaction with these systems...
replace AI with calculators, computers, the internet, _books_ (Plato) and you have the entire history of technology in education..
You are not going to put the genie back into the bottle.
Some teachers ban calculator use in their tests, some explicitly check if your work is derived from answers you did not come up with (something you found on the internet etc.). Unless a class takes account of some tool, using it could be to the detriment of the thing you are there for—namely an education.
> replace AI with calculators
Seriously? Idk about you but we weren't allowed to use calculators until long after we had already learned how to do whatever the calculator did by hand, except in cases where that was never the point of the lesson to begin with (like calculating cosines). And even then, sparingly. The reason they eventually let people use it was that it had nothing to do with the lesson being taught, and it was a waste of time at that point to make us do arithmetic by hand. Not because it suddenly became OK to turn off your brain and become lazy.
Idk about you but we weren't allowed to use calculators until long after we had already learned how to do whatever the calculator did by hand
Yes, I remember those days all too well. "You need to learn to perform long division by hand because you won't always have a calculator with you."
LOL, no.
It will play out pretty much the same way this time.
Skipping to learn long division by hand because you'll always have a calculator (or better a smartphone at hand) is like going to a gym with a forklift. It is faster to lift the weights, but it kinda defeats the purpose.
Learning to do long division by hand trains the mind. It prepares you to understand polynomial division. It teaches you the deep connection of division and subtraction. So, mastering it should be part of the curriculum. As soon as it is truly mastered, it should be allowed to use a calculator.
I do, in fact, not always have a calculator on my person and sometimes need to do basic arithmetic by hand or in my head. The teachers who told us that were 100% correct.
Most people structure their lives in such a way to always have that calculator with them if they expect to need it, or failing that, arrange to postpone the division so they can do it when they have the calculator again.
Difference is AI involves offloading most thinking. AI may be cool for already intelligent, but it is not going to form intelligent brains leaning on that from a young age.
I don't mean "AI" like generating quizzes or whatever. But people fooling themselves that LLMs are a "bicycle for the mind" with little kids are delusional.
>Difference is AI involves offloading most thinking.
i dont really think ai should be in most schooling, but calculators and computers also offload thinking. books offload memory.
It doesn't have to be naked LLMs where you can just ask it for the answer in an educational setting.
Channeled correctly it could be a fantastic educational tool. And will be in my opinion.
Obviously the "give me the answer" style of AI should be banned for students. Do good educational AI tools exist yet?
Maybe if your prompt is "learn for me, make no mistakes"
Seriously. Experimenting on children should be considered a crime against humanity, these people are sick and always show how they will betray us with their anti-human beliefs.
Everything ever done in schools was once an experiment done on children. Please approach the conversation with a level head.
I am being level headed, I do not think it's good to let advertising companies or for-profit entities that are based on not getting consent.
We have seen how damaging these tech companies already are to society and children.
Why should we continue trusting an industry that has shown no inclination to well being of children and not merely trying to pursue more profits in the form of corporate welfare.
Absolutely not the same at all and shame on you for saying so.
The biggest tell is how none of these tech executives ever subject their own children to these supposedly "great tools." Because they don't help children, they're made to extract profits by whatever harms as necessary.
The push for AI adoption in K-8 is the one that strikes me as wrong. The Alpha School and other initiatives are pushing AI on groups for which we have dubious measures of impact. Ostensibly, they are doing this to avoid pushback from traditional curriculum requirements in high school and university - but those kids are going to end up there eventually.
Why would I believe that someone who can't use a mouse/computer is ready to be taught by AI?
> Why would I believe that someone who can't use a mouse/computer is ready to be taught by AI?
The human element is critical for teaching, in my opinion. (Great) teachers are able to smooth over some of my criticisms of education. I also believe we are very far off from it being able to do even a bad job at teaching. So I'm not suggesting that at all.
What I'm talking about is teaching students how to use it, especially when not to use it, and how to get better results from it. I don't think AI is going to replace us any time soon, but people who use AI will replace those who don't.
> AI merely exposes the inherent flaws in how we do education ... The knee-jerk reaction of banning AI is probably the wrong choice, people need to learn where the limits are
But you can't do that without a certain level of maturity.
> but I also wouldn't be able to suggest a framework where students aren't rewarded for skirting around being educated.
Well, then "knee-jerk" ban AI in schools... until you can suggest such a framework, and test it to show it works, and roll it out at a society-wide level.
Until then, you're just handing kids flamethrowers in a highly flammable environment, and saying "we'll just figure out the problems later."
I reckon it's a similar issue to wikipedia when that became a resource available to us in school. Being taught how to use AI as a resource as opposed to 'the source' is the critical distinction. Which I guess is also the same issue when it comes to professionals and how they use AI
The problem is not using AI to get some easy answers, it's using it to do all homework, and cheat on exams.
Agreed.
Ultimately it's the same issue across all technological tools. If the tools (AI, wikipedia, calculators, etc) can be used to complete work (or homework) without understanding the underlying fundamental concepts of the work being done, then you're hurting your education and/or professional development
> AI merely exposes the inherent flaws in how we do education (exams, especially failure) [...] but I also wouldn't be able to suggest a framework where students aren't rewarded for skirting around being educated
As someone who graduated from college in the 2000s, I've been surprised to hear from recently-graduated colleagues that doing exams in person is no longer the default. That strikes me as the obvious first line solution for (at least) the STEM subjects that lend themselves to exams centered around technical problem-solving: exams in person, no phones or similarly capable devices, cheaters get a failing grade with no possibility for appeal.
The conventional wisdom when I was in school was that classes where homework isn't graded aren't really any easier for most students, because the homework is where you learn the skills, and if you don't do it then you'll fail the exams. Frankly, the same will be true either way for students who use LLMs for all the exercises: they won't learn the skills, and they will fail the exams. This is a working-as-intended outcome.
So why not start there? It's how school has ~always worked, so there's no question of feasibility. The only question is whether our will to give committed students a real education is stronger than our desire to productize education as a rent extraction scheme tailored to a nepotistic and shallowly credentials-obsessed society where having 'B.Sc.' on your resume sorts you one bucket higher in the algorithmic hiring queue to flip burgers at your local McDonald's.
---
It's less obvious how to fix the humanities subjects where take-home writing assignments carry most of a student's grade, though I'm tempted to think that even a relatively simple and tightly scoped writing exercise given as an in-person exam could conclusively weed out most of the students who've been relying on LLMs to do the assignments for them.
The big problem with education is that all the work is done for grades instead of aiming to deliver value to others. It's important for students to practice but if you raise the bar high enough then AI is just one tool that helps you but doesn't have all the answers. If AI can do all the homework we have to ask ourselves how much value the homework had in the first place.
> If AI can do all the homework we have to ask ourselves how much value the homework had in the first place.
That's just inane. The existence of the nailgun does not mean learning how to swing a hammer has no value.
[dead]
The word 'graduation', regarding education, somehow lost its original meaning -- to measure peers against one another -- its sinister edge can be easily visualized by imagining a class of graduates stuffed into giant graduated cylinders clearly showing their achievements with etched gradations along their respective tubes. It is difficult for society to consider this process as the, or major contributor to the, problem of education rather than AI. I tend toward the belief that we place too much value upon these highly dubious and artificial systems of academic assessment; AI is providing a long overdue opportunity to evaluate the excesses and weaknesses of institutional education.
This video is 40 minutes. Could someone summarize Haidt’s view?
Haidt generally has very interesting opinions but I don’t have time to listen to this this morning unfortunately.
Ironically you can now have Gemini summarise YouTube videos for you within the YouTube app/website.
Haidt is one of those political conservatives who calls themselves a centrist or moderate. He and the host David Brooks are both in the Epstein files, the latter described in the files as one of the biggest supporters of the Epstein backed EDGE foundation (with Haidt name dropped in the same email).
Speaking personally, YMMV, but I hear either name and I'm immediately skeptical.
[dead]
[dead]
Just like other "inferior replacements", many people aren't getting the real thing at all and the fake version is better than nothing.
what would be wrong with using AI only in /learning mode? I've used it this way before and found it even more useful than my memory of some teachers I had. If it was locked to this mode and you could not just generate answers I don't see the harm.
> I want to understand this more deeply. So instead of you just telling me everything, lay out the problem and a question for me to think about, and I'll try to answer. Even if I answer wrong, you should resist giving me the answer, and instead keep digging with more questions, so that I eventually arrive at the answer myself.
I use this prompt quite a bit. You get a lot of mileage out of it. But I have a pretty good idea of what I don't understand and am very focused on what I'm after. If you don't know what you don't know, then it's a mess. Even more if you don't know how to express what you're trying to understand.
In education, meaning primary and secondary, a narrative is used to provide motivation and create an interest and curiosity for the subject. I already have that, so it's much easier than digging through textbooks. The AI doesn't do that now.
AI in learning mode has the same shortcomings as virtual classroom. The physical presence of a teacher in the room. Sets the tone.
If we had AI when I was a kid, and you sat me down in front of a computer with AI, and told me to learn about sub-Saharan Africa and write a report manually, of course I would try to automate it. But it would be so much easier for me to learn with, because I have a learning disability, and my own style of learning new information.
Academia relies on teaching everyone the same way, which advantages some kids and disadvantages others. AI could potentially help more kids. To say nothing of increasing access to education, which so often depends on how wealthy the school district is.
No, it isn't
[dead]
Haidt doesn't have the greatest track record[1] when it comes to diagnosing society's ills, but in this case he does appear to be on the right side:
https://infographics.economist.com/2026/20260822_WOC963/
https://infographics.economist.com/2026/20260822_WOC965/
https://infographics.economist.com/2026/20260822_WOC839/
For most kids, AI is a detriment, but for a minority (smartest? most self motivated?) it's a boost.
[1] https://en.wikipedia.org/wiki/The_Anxious_Generation#Academi...
Name me someone who has a great track record when it comes to such topics. The criticism cited in Wikipedia is totally unsubstantiated.
> For most kids, AI is a detriment, but for a minority (smartest? most self motivated?) it's a boost.
While I doubt it's a boost in anything but productivity, the negatives are simply too great to ignore.
>The criticism cited in Wikipedia is totally unsubstantiated.
???
Odgers, Candice L. (March 29, 2024). "The great rewiring: is social media really behind an epidemic of teenage mental illness?". Nature. 628 (8006): 29–30. Bibcode:2024Natur.628...29O. doi:10.1038/d41586-024-00902-2. Archived from the original on September 10, 2024. Retrieved April 21, 2024.
And if you go to the article that has citations as well.
>While I doubt it's a boost in anything but productivity
If you look at the last chart it shows that students that spend the greatest amount of time using AI gets higher exam scores. It's not just "completes homework quicker".
>the negatives are simply too great to ignore.
That's... exactly what I said? I'm not sure how anyone can read "For most kids, AI is a detriment, but for a minority (smartest? most self motivated?) it's a boost" and get the impression that the negatives are fine because a tiny minority are getting better learning.
As someone who's read all of his books and experienced the generations being studied I'd say he's been spot on and is extremely knowledgeable. One study that does not propose alternative views and cites "selling a lot of books" comes off to me as someone with an agenda. No thanks.
>One study that does not propose alternative views
Why is proposing "alternative views" required for a critique? If we're discussing what caused diseases pre-germ theory, does that mean we can't point out the flaws in miasma theory because no "alternative views"?
>and cites "selling a lot of books" comes off to me as someone with an agenda. No thanks.
Well clearly you only read the first two sentences (or first paragraph?), because it cites multiple studies:
>These are not just our data or my opinion. Several meta-analyses and systematic reviews converge on the same message2–5. An analysis done in 72 countries shows no consistent or measurable associations between well-being and the roll-out of social media globally6. Moreover, findings from the Adolescent Brain Cognitive Development study, the largest long-term study of adolescent brain development in the United States, has found no evidence of drastic changes associated with digital-technology use7. Haidt, a social psychologist at New York University, is a gifted storyteller, but his tale is currently one searching for evidence.
Not to mention "selling a lot of books" isn't even the critique, it's that it offers a seductive explanation:
"... because Jonathan Haidt is telling a scary story about children’s development that many parents are primed to believe"
> For most kids, AI is a detriment, but for a minority (smartest? most self motivated?) it's a boost.
Exactly. The educational system needs to be reformed to exalt and celebrate the minority boosted AI, and quickly shuttle those harmed into the trash. Our North Star needs to be AI technology and the returns of investors who support it. Everything else, including human beings, are unimportant and expendable. Tech first!
>Exactly. The educational system needs to be reformed to exalt and celebrate the minority boosted AI, and quickly shuttle those harmed into the trash.
I'm not sure how you got the impression that not being maximally rabid against AI (ie. recognizing there are benefits) implies I'm sort of AI bro that is willing to sacrifice the education of 80% (or whatever) of kids into the metaphorical woodchipper just so the remaining can get slightly better exam scores.
It's a common trope on HN to misinterpret a commenter's intention and then take it to the extreme. You're not alone.
Or I'm just responding to a piece of information with my AI bro filter on (in the Instagram sense).
Haidt is first on my personal list of "people who don't really know what they're talking about but who I nonetheless mostly agree with." The "If Books Could Kill" show did a good episode about some of his works that indicate he has a fairly nonrigorous methodology and implements some iffy statistical controls to support what is probably a pre-conceived viewpoint. He doesn't have the best reputation in the academic space he operates within.
That being said, this stuff is hard to study, and I have no doubt that things like social media and AI are bad for a developing brain even if you can't arrive at a statistically satisfying conclusion. I can tell that this shit is bad for me, and I'm a fully grown man.
[dead]
Wikipedia is fair to cite Odgers's review. But the nature of the available evidence means that Odgers would be overstating her case to say that Haidt's theory is conclusively wrong, and Haidt would be overstating his case if he claimed to be conclusively right.
You see, the effect on social media on socialization has two parts. The first is the direct effect on an individual caused by their own use of social media. The second is the effect on an individual's social environment when everyone around them is on social media. The first effect is difficult but basically feasible to study. The second effect is downright wicked. So dismissing evidence that hints at the second effect as being "mere correlation that disappears under scrutiny" is not valid reasoning when your process of scrutiny focuses on the first effect.
In social science you have to accept reasoning from incomplete data. You won't always have the answers. I don't think Haidt really stands above or below his peers in any category except notoriety. Of course anyone who has worked with scientists knows to be a little suspicious of a researcher who seems to enjoy being in the public eye. But as the public eye, we're still stuck with them.