If people attend university for certification instead of education I think the battle is already lost. AI is an easy, but possibly high risk approach to gaining the certification without work, but the tried and tested approach is doing the bare minimum, cramming, then forgetting everything after graduation.
If you penalize people who use AI but in the process have learned the required information you make the problem even worse.
These problems are all because of a culture that favours the measurement over what is being measured.
The author mentions the notion of a "permanent underclass." What that looks like relative to your comment is this: more enterprising students (usually, but not always, from privileged upbringings) will learn how to develop the restraint to use AI as an assistant and a learning aid, while the vast majority use AI as an easy button.
Except it's actually worse than that because the big AI providers with billions on the line are actively encouraging their users to use their services as easy buttons. If they weren't, Googlebook wouldn't have a Magic Cursor/Clippy as a Service feature that suggests using Gemini when you jiggle your cursor over something, and Google Docs wouldn't insert a "Write for me?" CTA when you stop typing.
indeed. a certification is required for ~every non-manual-labor job. even if not listed as a requirement, you're almost certainly competing against people with degrees.
so, the primary function of going to school is to get a job, not for self enrichment.
What made University of California Santa Cruz so attractive to undergraduates in the 70s was no grading of performance and engagement but instead written evaluations. I loved being able to concentrate on great subjects with the help of often excellent teachers and TAs. Swathmore is noted for this culture too.
Do let us know where we can find a chemistry lab or biology lab or the facilities used for various other engineering undergraduates in a public library.
Having gone to university in Germany, there are glimpses of this ideal, but they're mostly faint memories enshrined on faded plaques around the campus. I did have an old geezer prof (90+ years old) that went to the very same university over half a century ago and showed us his diploma: greek, latin, humanities, ... for a technical diploma, no less!
I do still cringe a little when we get newjoiners fresh out of university proudly proclaiming "Yeah, no can do, we didn't learn that at university". Yes, obviously, university is not an apprenticeship. You learn how to learn and then apply that to unknown-to-you problems. Oddly enough ChatGPT seems to have brought a change to that mindset, but Im not sure if it did so for the better.
For US undergrad, with all the resources available online, going to university just for education doesn't justify the high tuition. Research is mostly limited to graduate school anyways.
This is not at all what most students go to undergrad for.
For some reason in discussions like this, things are always framed as if everyone is going to Yale.
Everyone I know including myself, went to college because that is just what you were supposed to do. That was my mom's dream for me to get a degree and I didn't want to disappoint my mom.
I am sure there were really smart kids in my high school that were figuring out post graduate network strategy but that was hardly the majority.
I even had a friend that dropped out of high school that ended up going to community college. He wasn't exactly figuring out the eigenvector centrality of the nodes of his general studies classmates.
Correct. Social pressure is what pushes most people there.
As a result of this social pressure, people do tend to go off the rails when there (having fun) very few give a fck about studying - who cares, there's no reward for getting anything beyond 60-70%.
Lol, nice to meet you. Left freshman year, got paid, got out. Questions welcome. College for tech jobs became useless after like 2008. Books are cheap. I'd challenge anyone in response to prove they got a degree in the 21st century.
This is mythology. I went to a very high ranked uni and this is not what any of my classmates were interested in either - it was all about getting drunk and pulling all nighters a few weeks before exams.
Career networks? Hahaha. Very few benefitted from that.
Once again, I went to a top 3 ranked uni on many league tables worldwide.
This is one of those things that seems true in theory but doesn't have examples of success. There are people who self-study advanced mechanics at home, but they're child prodigies and enjoy (typically) going to universities as a way of meeting others who share some interests.
This particular battle for learning was lost a long time ago. If university stopped providing an earnings boost from attending, 90% of students would quit tomorrow.
It doesn’t help that a lot of desirable fields are comically out of date at the academic instructional level anyway.
Would you honestly tell an aspiring software engineer that your typical computer science degree will teach them much about wielding computers in a cutting edge way?
If I were to list the top 5 things I got from university, knowledge wouldn’t make the cut and were I to do it again, I would certainly attend less class.
> If university stopped providing an earnings boost from attending, 90% of students would quit tomorrow.
Maybe 10–20% would quit for that reason. There would be more attrition if you could get common jobs (such as teacher or nurse) that currently require a degree but don't pay that well without formal education.
Most people don't care that much about money. Sure they would like to have more money, but it's not the primary factor that drives their major life decisions. People are generally more interested in stable careers that pay their bills and seem like something they could continue doing until retirement.
Anyone who thinks that a Computer Science degree is supposed to prepare them for a job as a Software Engineer has completely missed the point. It's like getting a Physics degree for a job as a Mechanical Engineer. There is some overlap but a huge difference in focus on theoretical versus practical topics.
My computer science degree did not cover much actual computer science.
You can argue about whether that is a good thing or a bad thing, but the ship sailed long ago. CS undergrad degrees are about training software engineers, not about training computer scientists.
More and more colleges are saying they will cancel your degrees if they retroactively detect AI cheating. This will be hysterical.
AI isn't the issue as much how AI is used. Passive use of any tech, including social media, and now AI is lazy and has poor outcomes.
Aligning AI use with the goals of all sides, and not just one side getting paid, or just one side graduating could look different.
Most people attend higher education to access opportunity to improve their lives, overwhelmingly for a career and earning.
The idea of higher education teaching "learning how to think" is perhaps a relic of the origin of some universities which didn't historically do STEM, and focused on things like liberal arts, which in turn often had the support of coming from a privileged background, or financial safety net.
STEM money and funding though, attached a lot of traditional post secondaries to do that as well.
It's perfectly acceptable to expect higher education of any kind to have you ready to grow and earn more in better suited opportunities. Not enough educational institutaions don't publish their % of students who graduate in the area that they started in, and also the % of graduates who find their next step, career wise, etc, in 6-12 months of education.
You mean the same colleges that used AI proctoring during 2020 that had a lot of horror stories of false positives, like a guy passing in a window in the back of the student or another student that was not looking 100% of the time to the screen?
It is very difficult for them to prove the allegation of AI cheating. If you didn't cheat, you should fight. You have legal rights, and you don't have to roll over. I represent students who have been unfairly accused of AI cheating. Properly defended, it is not that easy for the school to win.
I have found the whole honor system of academics to be just another stick to whip students with when convenient. I am not opposed to any sort of honor system or academic integrity when there is actual honor and integrity, but the system needs to apply to both sides. None of this "rules for thee, but not for me."
> These problems are all because of a culture that favours the measurement over what is being measured.
Spot on. I am teenager going to college soon and I feel like the same way about the education system (and in extension, the job market but I suppose that the job market might be more understanding probably over all of it), part of my comment was as follows.
I do feel a bit like coding/a lot of fun-ness out of life is also like this, quantified, measured, transactional (posting for social-media?) [as I wonder if I am writing this comment for hackernews karma or relevant discussion talking points..]
This feels to me the most irreversible consequence because it might be hard for the generation (myself included) to see value in non-measurable things as everything has to be measured and transactional-ized.
(...) I would like for humanity to be more nuanced and less measured but more varied (grey rather than black or white) but I feel like that there is enough noise on the internet that maybe even this ends up becoming noise and I am not sure if anyone who might benefit from reading this actually does end up reading it.
A culture rooted in φιλοσοφία (greek, philosophia in Latin). So yes, I meant that literally. There were times, already 2500 years ago, where people wanted to study to become wiser.
It's great when you're a wealthy noble and have time to do luxury things.
For almost all of history, higher education has been a luxury good for the rich, including the Greek city states. There have been a few exceptions, most notably European countries with tax funded schools, but even those are primarily pumping out degrees used for chasing jobs.
The point of modernity was to broaden access and opportunity for everyone, not just the rich.
What we got instead was a regression to aristocracy, where critical social resources of all kinds were enclosed and captured by capital and financialisation.
The result is a lot of very broken systems, education being just one.
Universities became primarily about administration of property and income, and the educational element has become a form of marketing to attract money (and bodies) so the rest could function.
And now we're on the edge of the next stage, which isn't "What is education for?" but "What are humans for?"
We used to know. Or at least we used to believe we knew.
I'll throw my two cents in, this still exists today. The difference between now and then is that now a college degree is seen as a requirement for a higher quality of life. (And not in a eudaimonia way)
Many people are going to college primarily to make more money in their adult life, the actual learning is secondary. If you're already well-off or just don't care, you can still get the education for its own sake.
The issue is that we've created a perverse incentive to get a college degree.
Over a decade ago, my orientation at UChicago included the traditional "Aims of Education" address. They packed the whole first-year class into the chapel to explain, at length, that this education will not be "useful."
You're not supposed to make more money, or be happier, or really become anything other than a better version of yourself.
UChicago should be pretty uniquely positioned to address the problem of AI writ large. They already require a full year of each philosophy, literature, and history (all through primary sources). This "Core" should already be fairly AI-proof, given they are primarily small-group, discussion-driven courses; oral exams, in-class essays, or even graded discussions should be straightforward adaptations.
And yet, the university shifted towards professionalism before AI ("training a mind for the workforce" rather than "the good life").
Already, this transition did what the author observes AI is doing. I would hardly believe someone who cheats through an econ/stats major is less educated - if only through osmosis - than someone who honestly completes Business Economics.
And so I wonder - if the damage of AI is primarily instrumental to the broader trend of hyper-professionalism, what damage has it actually done?
If we automate away the signal to companies "yes, I can do stats for you," does that free students to focus more on the _less_ professional aspects of education?
Sure, it undercuts credentialism, making the "piece of paper" near worthless - but if our aim of education is just to "be better," should that not give us hope?
If the signal to corporations disappears the wage premium to a college education disappears and the students disappear along with the tuition paying the professors’ salaries.
Maybe - to your point, if we think of happiness as like “living one’s own purpose fully”, then yes, it does very directly.
But I was referring to happiness more generally as enjoyment, joy, satisfaction - that type of thing.
And in that case - there are plenty of ways being better = less happy. Eg if I were to sacrifice myself to save my family, then that’s the best version of me, but I’d be pretty dang unhappy about it.
Oddly enough my reaction to this is that it's a broader societal problem as opposed to an A.I. problem.
Why shouldn't universities switch to examinations where no technology (apart from say calculators) are allowed; and this is strictly enforced? This was certainly the norm when I went to university.
I agree that A.I. trivializes (or changes how you approach) a lot of take home work; but people who wanted to cheat could more or less always do so for that to some degree. I guess it makes it easier to do so; however my expectation would be a greater reliance or weighting on in person examinations as a response; as opposed to a normalization of cheating.
One way in which A.I. could be seen as contributing to this is that it is devaluing the importance of what were seen as 'intellectual' pursuits; as we now have automation for them that is at the very least often surface level effective for undergraduate work.
As recently as 2015 when I attended a middling CS program, we had in-person timed exams where we had to write down DSA implementations on a blank sheet of paper in Java.
We were deducted points for trivial syntax mistakes.
If these stories I keep hearing are true, then university programs have really taken a nose dive recently. This isn’t a “back in my day” thing, but within the past 5 years.
The pace of the purported decline makes me question if some of these stories are sensationalist. But I don’t know, I keep hearing about them.
I'm not sure I understand your comment. Surely you don't think that the details of a particular programming language's syntax are an appropriate criteria for grading an exam? That seems crazy.
I would suggest the one exception to this would be courses explicitly designed to teach how to use AI, and how not to. But in that case, it's less "use AI to cheat on this course" and "AI is the tool this course is about."
Then it becomes, teach what? "To use AI", yes, and, then, to do what? Use it how? To make some software? Why? You are already taking software engineering classes to learn to make software. To write something? Why? You are already taking classes that ask you write things yourself. An AI class, to me at least, is akin to taking a class about how to pay someone to write your essay for you.
And if we are talking about the various AI strategies people have where they have LLMs talking to LLMs to come up with whatever gooblyguck, are the poor souls who've been asked to come up with the AI class for the department going to know any of these strategies themselves? Are these strategies even going to be sustainable going forward after VC is no longer subsidizing tokens?
Regardless of VC subsidies, the cost of compute always trends down over time. Whether you like it or not, LLMs will be a pervasive part of everyone's life forever (or at least until a better replacement comes along).
Cost of compute trending down is usually lost as the resulting software bloat that fills the empty space like a gas. We already see this with LLMs. Models get bigger and bigger in an arms race.
I don't think that is a safe assumption to make. Moore's law is not playing out any longer as it used to. Jensen Huang already called it dead 4 years ago.
The college system is creating the zombie underclass with AI or without it. The amount of money colleges charge combined with the text book thinking shapes people into thinking there are steps to success, there are right answers, and "getting a job" is the right way to go. Colleges don't teach independent thinking and that is the exact thinking we need in the era of Youtube and AI. You don't need college to teach you how to learn text book items anymore and I think that is scary to some.
> You don't need college to teach you how to learn text book items anymore and I think that is scary to some.
a) This is about universities, not "college"
b) The University teaches you critical thinking, not how to learn "textbook items". It's not vocational training for upper middle class. It's for building and developing citizens who can think critically.
In the United States, "college" and "university" are generally used interchangeably. I absolutely agree with your second point and the shift has been going on so long that people are genuinely baffled by the idea that college should not just be "white-collar vocational school"
Most jobs, including white collar jobs, don't require the skills and knowledge learned at a college level. For employment purposes, a college degree primarily acts as a signaling and filtering mechanism. It shows employers that you can grind your way through a moderately difficult task without giving up, and it allows hiring managers to reduce the number of applicants to a manageable level. When actual labor shortages occur, employers are quick to drop the fake degree requirements in order to fill critical roles.
Not really, you need cooperation with other people in this complex world to live. No necessarily a job. You could be self-employed or a member of a cooperative or an elected official.
But yeah, the capitalist default is to have a job, sure.
> And if I want to do something interesting I need the skills and knowledge which are learned at a college level.
Not really, no. You need the skills and knowledge and for some professions you do need the official certificate of education and for a subset of those that's actually warranted, because you cannot get your hands on the training other ways. Doctors kinda need the official system, self-taught appendectomy would not be ideal. English literature? Not so much.
You need a job to live - under the current economic system. a lot of the critiques of capitalism lead you straight into ai, and most economic critiques of AI are critiques of capitalism as it exists now.
You've needed a job to live for all of human history going back to whatever ancestor you like. That job is called survival. All modern life is doing is delegating that effort across many people such as to benefit from economies of scale. I don't need to fish for my sustenance because someone else is doing the fishing. I contribute in some other way that gives me credits to access the fishing take. People act like capitalism is some new evil but it is little different than the delegated labor you might find in a prehistorical tribe. The fundamental rules of this system are really the only way our species survives. The issue is sometimes compensation is uneven for the effort, but that is really it, not that the system itself is a failure. I'm not sure of any system that wouldn't be like this. What is communism but capitalism where compensation is more evenly distributed? It is still a game of delegating certain tasks to certain individuals such that one individual is not responsible for 100% of their own survival.
Survival didn't involve a CS job at FAANG until very recently.
Capitalism is not a new evil, you're absolutely right. It's just the system that, according to our current framework, came after feudalism, which came after something else.
the point is that actually seeing the changes in systems as consequences of massive technological change is likely directionally correct and it is not clear whether current capitalist logic survives the AI era.
What is a CS job at a FAANG in practice? Many hats. For example if you work at Netflix on content delivery, you are the modern version of someone involved in the boring logistics of theaters and plays. If you work at Google on search, you are the modern version of a librarian establishing ways to catalog information. If you work at say facebook, you are the modern version of a propagandist or even town crier perhaps. In all these cases, these workers were delegating their basic survival needs to other members of society. Ancient Rome even had its own slop bowl restaurants not all that different than what you see in SF today, and many ancient romans did not cook for themselves but instead traded credits gained from labor elsewhere to eat at these establishments. No different than lunch time in SF today.
The actual tasks being done have always been done even if the technologies used to do the work have changed and the work itself looks a little different. Our base needs have not changed. We are the same apes bound by dopamine. The functionality we require is pretty old. There is probably little we experience that is truly "new" compared to what one might experience in a classical civilization.
Capitalism is far more ancient than feudalism. Phoenicians were famous capitalists, likewise most any group of people involved in trade. Trade and accumulation of capital representing labor is ancient.
You're just confusing definitions. Capitalism as its commonly thought I would think just about what it says on the wikipedia page
"Capitalism in its modern form emerged from agrarianism in England, as well as mercantilist practices by European countries between the 16th and 18th centuries. The Industrial Revolution of the 18th century established capitalism as a dominant mode of production, characterized by factory work, and a complex division of labor. "
Trades can pay very well and frequently require nothing more than on the job training.
You think you need college for the same reason you equate "job" with survival. These are not universal truths, not even in capitalist hellscape America. It might be harder but it is in no way a requirement. Anyone who tells you different is lying to you.
It does feel like there is an easy solution to this:
Have tests.
Supervise said tests to make sure people don't cheat.
That's how it worked when i was in university. Admittedly maybe that is easier in the sciences than humanities, but still, it seems doable. Cheating isn't a new phenomenon it just got cheaper and easier.
The tests will be made by AI as well, because professors who spend less time on making tests can spend more time on research and get hired/promoted more. For the same reason, professors won't care who is using AI to cheat on the test that was made by AI. Maybe some people will care, but not enough to do anything about it.
(Cheating was already rampant in many classes 20 years ago when I was in college, I can't imagine what it's like now.)
This selects for people who excel at taking tests. Sure, there may not be a better alternative, but as an empirical measure I learned just about nothing in my test heavy college courses, as I was incentivized to cram for the exams and purged everything immediately after the final (and midterm).
OK but why? That will reward students who are good at taking exams, which has very little relevance to being a competent worker or well-rounded citizen or innovative leader or anything else that we need.
Overall it’s a very good read, really enjoyed the author voice and style
> I don’t think she was laughing two years later when I was TAing the class and we observed a fairly distinct gap of about 40 percentage points between the take-home test and the one administered in-person.
40pp is massive. Take homes are pretty much dead at that point. And not just in schools, but also for interviews. I don’t see how you can get a meaningful signal, it’s guaranteed they will be made using AI.
I had very few take-homes in school (discounting project courses which were very time consuming). One of the problems with take-homes in most interview situations is that, at least for developer jobs, the reality (whatever the stated guardrails are) is that this will be at least a weekend project. That said, I've spent multiple days traveling and doing related things for an on-site interview which I realize not everyone here would find acceptible.
Of course the take home interview assignments will be made using AI. Isn't that the point? For most white collar jobs we should hire candidates who can make effective use of AI.
Tangental, but is Leetcode dead too? Part of me could foresee Leetcode being believed to be a stronger indicator now, since it would be unlikely that someone that can answer those questions would know absolutely nothing about programming.
Then again, with LLMs, learning to answer those questions is probably easier now than ever.
I haven't applied for a job in a decade, so I am utterly clueless on the current landscape.
Clearly not all of them want to do so. My son just finished up his freshman year at college. A couple months ago he called me and told me about an interaction he had with his English composition professor. She asked the class how they were integrating AI into their writing process. When she got to him, he said he didn't use AI, as he wanted to use the assignments to learn to write better. Her response was, "Well, you're probably too advanced for this class then." This is my son, who has a dysgraphia diagnosis, being told he is "too advanced" for freshman English class because he doesn't use AI to do his writing.
That is fascinating to me. Especially in English composition. The flip side is if it's an adjunct prof making $4k per class (English typically pays poorly) then she's doing the hourly rate calculation and thinking that AI is going to help her with the students. So, a potential solution to this is for the universities to have the willingness to pay their professors and ask for no AI (at least in this type of class) in exchange.
It's going to be interesting to see where universities are in ten years. Higher education has for a long time been a bad proxy for internships and apprenticeships, because if you aren't teaching what the market is paying for then you don't get funding or students. Watching universities start to defend against this obvious and accelerating association is fascinating. There's an obvious decoupling between "learning how to do a job" and "higher education for fun because I'm rich" that we're only going to see getting crazier.
Same. Some of the developers I work with leave Claude's comments in the code: "Wait... let me try something else!" I don't mind they're using AI, but leaving junk around is a code smell. What else aren't you paying attention to?
Universities have the greatest discrepancy to me between how they're described from those who are no longer associated with them or who are weakly associated with them and the reality of those who are dependent on them as employees or as students. To listen to the alumni, every university was an institute of learning where all standards were held to the highest and everyone adhered to every rule to the letter.
To the majority of students, they seem quite laser focused on acquiring the degree with the right grades so that they can maximize their chance of a job after university (apart from the personal element of partying and having a boy/girl-friend etc.). The primary utility of the university to students is the credentialing, and secondarily the structure to the learning program, but otherwise the books themselves suffice to teach.
Perhaps we should move more training to technical institutes and people can come out with the knowledge of how to operate this or that thing. The problem is that everyone will know that the smarter student has gone for the higher-end university. The credentialing then works not because of the program but because of the selection that the university can do. Okay, so the whole thing continues to make sense even if AI zombifies everything.
Back in my day, we did the learning in class, and the evaluation (homework, projects, but sometimes even exams) at home. Seems like AI just flips the script. Now, you can learn anything you want with a private tutor. The teacher can just upload the entire course material to a place where the LLM can rag over it and answer anybody's questions. Learn it on your own time. Then, use class time for evaluation. Do the "homework" in class. Take proctored tests without access to the internet, etc.
> The buildings, of course, will remain, to be observed and treated respectfully — like old cathedrals, mainline Protestant churches, and most of the European continent.
The universities can survive a wave of student ai laziness- blue books and pass fail homework. cfg. I can’t see how they survive a wave of professor ai laziness. Paying $200,000 for $40 of tokens is such a brutal way to be scammed and the resentment will get refreshed by loan payments every month for decades
I'm a big fan of the fact that the following quoted sentence has 25 words before the verb. An LLM would never. I probably wouldn't.
> Whatever your conception of the modern university, whether grand or grim, understanding the current landscape of campus-wide AI use, much less its intensification, should destroy it.
> Tying education to a capital-intensive and (likely soon to be) tightly regulated technology is one more step toward a different, frightening future. A world in which independent educational institutions are neutered and transformed by their reliance on a central authority into factories designed to train students according to the “needs of society” is not a new prospect — it has been the persistent dream of Fabians, technocrats, and engineers…
I hadn’t thought of this. Every school district and university tied into centralized AI inherently undermined its ability to decide how its kids are to be taught.
Universities will still act as gatekeepers of prestige and status. There is no AI alternative to the top-20 schools...I remember all the hype from 10-15 years ago about how online learning and "MIT courseware" would upend the universities or threaten credentialism, and nothing even close to that happened. As it turned out, the online version of MIT is not a substitute for the actual thing.
Schools will adapt, as they have already, by weighing grading more towards in-class quizzes and tests . I think the humanities will continue to struggle, but I see the AI boom making STEM more relevant, even if AI can automate a lot of code or math.
>As it turned out, the online version of MIT is not a substitute for the actual thing.
More precisely, the people motivated enough to actually do the online MIT version were often already on a high-performance trajectory, and for the people who were not, few people took the online credential seriously, despite whatever skills they acquired.
Educators absolutely can teach students to think. The scientific method is one example of a key mental tool which provides an organized, disciplined framework for thinking. If you read a lot of stuff written before the scientific revolution it's kind of a mess because the authors literally didn't know how to think. When they occasionally got things right it was mostly by accident.
On the other hand, it would be really good if universities stopped being gatekeepers of prestige and status. It seems like some of the biggest idiots in high visibility posts right now come from the ivy league..
I'm not sure if an ivy league education proves anything anymore other than that you're connected.
That’s because the value of the university is the assessment, not the education itself.
They are vouching for the intelligence, knowledge acquisition and work ethic of their graduates. If they lose that signal, they lose the ability to gate keep prestige and status.
One professor in one class has not fixed the problem. You can't generalize that to the whole university. Even for that one professor/class, it's not the case that the problem can't be fixed (it could be, quite easily), it's that the professor evidently cares so little about doing their job as an educator that they simply should no longer have a job in a university.
I was forced to get a cloud certification while working for a former employer. I took it remotely and had to leave the camera on the whole time. If you looked away even for a second, you'd get a warning. Very dystopian.
I'm a current student, who also happens to be a full-time professional who is "all in on AI", and I think most are missing the true opportunities AI opens up for education.
Because my student path is non-linear (vs just following a life script), I may be a bit weird / not the average student, but it's especially true for me that I'm very intentional about actually learning the things I sign up for classes to learn.
My point is that I'm not taking classes just for the motions or to create slop. With that context, here is how AI helped me very specifically in a recent linear algebra course:
1. I was able to prompt very specific questions, usually audits of my work, in ways that provided responses that were more like a socratic tutor and not a cheating parter. In this way I did not need to bother my professor as much or seek out a tutor, when I was stuck. But I also didnt shortcut my way to answers. I was intentionally limiting the AI assistance to finding small errors or jogging my memory about steps missed or next steps.
2. I vibe coded a note taking web application (started as a chrome plugin for notion) so that I could shortcode and pick math symbols while my other arm was full holding my newborn (yes I'm a dad too). This has since evolved into a full-on science writing platform that I love whether or not anyone else ever uses it (though I am trying to turn it into a business). Maybe I actually ended up adding more work to my math class but it added a layer to the learning (what math symbols are needed, what are typical patterns for this subject, etc) that I think helped with my overall absorbtion of the subject.
I dont know if #2 is transposable to other students or to other subjects but I imagine there is some version of a double major yet to be created that is Core Subject + "how to properly use AI to learn (including vibe coding tools to help yourself and other students)".
There are many other smaller ways AI can be used to help learning (flash cards, generated quizzes, etc) that are oft mentioned but that articles like this gloss over.
Having said that, I loved reading this (so well written it could not be AI despite the emdashes), and especially appreciate any mention of "The Whispering Earring", which is one of my spinning tops to remind me to remain vigilant of my cognitive health despite my almost complete embrace of AI.
I can't help but wonder if the fundamental problem is just that we spent decades pretending that a university degree was some sort of useful job training in the first place. As a professional software developer, I think my computer science degree is not actually all that important. Sure there are some relevant concepts, but they're ones you'd pick up on the job anyway.
I don't regret getting my degree (back in 2009), but I think requiring a person to have one is a dumb job requirement.
Frankly, we shouldn't have so many people going to university in the first place. There's a lot of people it's just utterly wasted on, and it drags down the entire apparatus as a result. In a sane society we'd have much more apprenticeships, vocational training, etc.
Even though the degree really didn't teach you what you'd do on the job, it was a signal to employers that 1. you were capable of learning stuff when needed, 2. you didn't give up on doing hard things, and 3. they wouldn't have to explain to you what a loop or a class is.
With rampant AI cheating it's no longer a guarantee of any of those.
I do not think I had a single university course that was not ultimately assessed through supervised written tests. That matters, because homework was never a reliable anti-cheating mechanism. Even before AI, you could just pay someone smarter to do your homework for you, or ask older students for their assignments (and sometimes even their test questions, there's even a pg essay on this [0]).
You can cheat as much as you want on homework, but it won't help you on supervised written tests. At some point you have to sit down, unaided, and show that you can solve the problems yourself. So I do not see how AI substantially weakens the signaling value of a degree, at least in systems where the degree is backed by in-person written assessment. It may make take-home coursework less meaningful, but that was already the weakest part of the signal.
I don't know, I went to a pretty well regarded CS program but some of the people I graduated with are definitely not people I'd want to work with. They might have hypothetically known what a loop is but I wouldn't expect them to put one to good use.
For Computer Science in particular: it's not supposed to be job training. CS is an education closer to math or science (in fact, at my university it belongs to the department of hard sciences and math). If you like that (and I sure did!) it'll be worth your time. If you're just looking for job training, you're looking in the wrong place.
My university CS program didn't even teach programming in any of the major classes, it was assumed you'd learn on your own or by doing one of the optional workshops.
There's a lot of stuff taught in academic CS that you simply won't learn on the job, or if you do, it won't be as rigorous and you'll be missing the fundamentals.
What program was this? It sounds much closer to what I wish my CS degree had been. In my CS program, the courses I actually enjoyed were the math-heavy, but always optional, like computability/decidability/complexity, cryptography, etc.
The mandatory "practical" courses were often much worse. For example, I studied relational algebra on my own, plus a few chapters from Kleppmann's Data-Intensive Applications book, and it was painful to realise how shallow it made the mandatory database course look.
I agree that CS should not be mere job training. I think many CS programs are neither rigorous enough to feel like math/science and prepare you for proper academic work, nor practical enough to be good vocational training. They sit in a bad middle ground, where academics teach industry-lite.
Industry-lite is a good way to describe it. I remember having a 300 level class which was supposed to be about real world application architecture but it was essentially just about making UML diagrams (because the professor happened to be on the board of whoever was in charge of UML.) Nobody serious (even at the time!) uses UML.
Zombification of USA people was already happening before AI. Not surprisingly, it's been one of the favorite cultural themes in the country's cultural produce for long years. Zombies + superheroes was not poised to produce great non-drooling non-moron americans.
> There is an extreme idealist view of education that might see the threat of AI as good precisely because it could transform those kids — the former connoisseurs of SparkNotes and Mathway, the ones snickering in lectures and inking formulas onto their palms before exams before the rise of generative AI — into zombies lurching and stumbling their way into the “permanent underclass” (as the tech bros say), leaving the elect few free to enjoy the benefits of a humanist education without all the noise and din.
In this world, what are the benefits of a humanist education? The only reason we care so much about education is that it's how to determine merit in meritocratic societies, and therefore a key part in how people gain social status. In a world where AI does all the knowledge work and robots do all the physical work, with an 'elect few' owning everything and everyone else in a 'permanent underclass', why do the elect few even need to keep the permanent underclass alive?
I think that universities just have to adapt to deal with slop, or think of new ways to challenge people to learn the essence of their studies. I wouldn’t want to be a uni teacher in these times though.
The (Western) educational system still does this for PhDs. Grades barely matter, and in most places you have two oral exams in your entire 5 years: your qualifying exam, and your final defense.
The reason it doesn't happen for the rest of the system is scaling. The US awards about 60k PhDs per year, compared to about 2M bachelors. There simply are not enough faculty and it is not realistic to hire enough (if there are even enough qualified people in existence)
And that's ignoring all of the problems with "not giving out grades" or "ending credentialism" - I guess people are supposed to just get hired on vibes?
Since everyone has a meaningless degree already the degree absolutely does not get you hired for over a decade already. Interview outcome, which sure you could call vibes
Degrees don't get you hired, they get you interviewed in your first couple of jobs.
I'm not going to interview some guy with no work experience and no credentials for an engineering position. I will gladly interview someone with no work experience but a relevant degree or bootcamp.
I agree with you that no-tech parts of universities would work - obviously you can't avoid tech when teaching some things like coding, but mostly I think it would be a good idea.
There are problems: Having students attend lectures is great but they have to work with the material and prove they understand it - how to do that without homework? I'm sure there are ways. Have them work in a building full of computers cut-off from the internet maybe, but how to keep them from using their phones?
Another option is just severe comprehensive testing in heavily inviglated rooms long after they finished the class involving the material to prove they know it. Perhaps you could do this for the first few years of knowledge in a discipline and then assume the student actually is serious and take the leash off after they passed the tests. I know some disciplines already do this kind of thing, even before AI. Basically everyone has to pass a bar-exam type thing, even if they're studying art - but things like art can't really be condensed into an exam and it would certainly restrict and narrow what can be taught and learned, that's a big problem in my mind. Also what if there are new ideas in the study of physics and they can't really be taught because the exam is too difficult to change quickly? What if there's a big split in the philosophy of buisness, but the exam only asks about one side of the split? What if you have an ingenious professor who wishes to talk about a new branch of philosophy he's created - not on the exam though.
Edit: I guess if professors designed their own exams, instead of some distant exam-comittee it would alleviate most of my concerns about them.
For coding you can actually teach students on commodore 64s. It’s actually better because they have a BASIC shell and assembly language. Most importantly, no internet. :)
Actually, give them internet why not. But they have to use a 56k modem. Mwhaaha
Tests. Many of my university courses only graded on tests. They strongly encouraged you to do the homework to better understand the material, but didn't consider homework completion when calculating your grade.
Consider that universities are educating adults who are -often- paying to be there. If we assume competent course design and instruction, if an adult chooses to not work on the material until they understand it, then the only person they're harming is themselves... which -as an adult- is a thing that they're usually fully entitled to do.
I liked the classes I had with a "reverse" format: the "homework" was done in-class, checked for correctness but not part of the grade, and the lecture was a recording watched at home.
You give exams in person, in class, on blue books, no phones. This part isn't hard. Instructors have been doing it for generations. It's only in the post COVID era that some have moved to having exams take home and on Canvas or similar platforms. This is great for instructors -- less work! but I am not convinced it actually helps students.
The part that is more difficult is take-home work, and I think the solution is that instead of being something that you turn in for credit, it needs to move to being more of a chance to practice for in-person exams.
What about essays? I've taught classes where students had to write essays in class, in person. On paper, with a pen (this may no longer be allowed on many campuses because of access and perceived fairness reasons, which IMO is a shame, but it is what it is). I think the traditional assignment of "write a 15 page paper on XYZ" is probably done. Instead students will have to prepare to write an essay in class by reading the source material (books, papers, etc) and converse with AIs that are hopefully not hallucinating, to get an understanding of the material and then come to class and be prepared to write about it.
Assignments, sure. But if tests/exams are proctored in-person with pen and paper, the students may quickly pivot to traditional learning methods if they want to pass their courses.
Oh how I hated having to write in longhand throughout my schooling. It was always so slow and painful no matter what I tried in terms of pens or techniques. Typing is much faster and more fluent. In the real world no one writes longhand any more: it's all keyboards, swiping, and dictation.
It would actually be interesting to see what people do attempting to transcribe AI generated material to paper. At the very least it's another layer of learning in writing it out.
Engineering at my EU uni, homework and coursework were at most a tiny part of the total grade, and never enough on their own to pass. If they were relatively bigger projects, you'll pass an interview or similar review after delivering it. This all were just nudges study and to check ourselves and they were seen as a "gift" of the Bologne Process (restructuring/standardizing of unis in the EU).
The only thing that mattered were the exams, be it pen and paper or coding/electronics labs, in person and proctored. No matter how much slop I could have access to back in the day I would have failed the same subjects I did.
Yes but AI doesn't prevent you from doing that learning. What is makes harder is the old broken ways of credentialing your learning. The way you do the learning has no need to change.
I personally feel like the software engineering profession may have to start moving more towards an apprenticeship model than a theoretical CS-gradate-then-work model.
Internship / coop programs at places like Waterloo already look a bit like this.
Slop made by students is one thing, but slop generated by facilities and fed at extreme premium to students just asks a question "why someone would pay for this instead of buying some LLM tokens, taking curriculum and teaching themselves".
If we want to teach students to use AI, it should just be a separate course, not shoving it in every possible nook and cranny to the point it is teacher AI talking with student AI with light supervision from both AI handlers
> The prevalence of AI use on college campuses, particularly at “elite” universities, is a cancer on our culture that threatens to turn a generation of promising young Americans into a class of drooling morons...
How exactly is it that you think Enzensberger is arguing against nwhnwh's point. Will an excerpt from Discourse Networks 1800/1900 be far behind, be still my heart
Right. I got my bachelors degree more than a decade ago, and they did a good enough job of it without AI. College was the first time in my life that I'd ever heard, "You cannot aspire to [ambition]," and I heard it there quite a few times.
Higher education needs reform more intensive than a simple defense against LLMs (as does the legal system and profession, as does the software engineering field, as does the field of psychology/psychiatry, as does-).
The prose is a bit too purple and tortured for that, IMO. Stock Opus 4.7 or 5.5 Pro is a more disciplined writer.
And, anyway, the point the article is trying to make is obvious. What's absolutely not obvious, and what it sheds very little light on, is what the University is going to look like in 10 years. Not what it should look like, but what it is most likely to look like.
You may be right. But that will mean the university becomes increasingly irrelevant and disconnected from the reality that students experience after they graduate.
It is easy to change the system prompt to make the AI talk with a different voice. It is remarkably hard (at least for Claude, I haven't experimented as much with GPT) to get it to not use so many em-dashes like this essay does.
There's no way. Just the first paragraph alone is enough to convince me; it's too well-written and melodious to be AI, with too much original thought:
Today, the demonic vice of the old is not that they are hard and demanding on the youth — instead they do not demand enough from us, and they cannot quite believe that we have not lived up to the little they have demanded. They think too well of our generation.
Without defending the quality of the rest of the essay, it's a great start. LLMs today could never match it.
Style is the wrong diagnostic. Purple prose and em-dashes can be prompted in or out. The harder question is whether the reasoning was committed or generated. A distinctive voice tells you nothing about whether the person actually worked through the argument or had it produced for them. Which is sort of the point the essay is making about students.
This is such an ignorant trope. The last few places I worked ALL used em-dashes as part of house style and I will continue to use them. It's extremely common (and arguably the LLMs do it because it is extremely common).
As noted in my own parent comment: I loved reading this (it's too well written it could not be AI despite the emdashes), and especially appreciate any mention of "The Whispering Earring", which is one of my spinning tops to remind me to remain vigilant of my cognitive health despite my almost complete embrace of AI.
AI generated or not, I concur. I rally want to know what Universities will look like in 10 years time.
What will be taught there that cannot be taught by an AI (whatever form or interface it has).
Will Universities still be centers of knowledge and exploration? or will that be more disseminated through society, and so Universities not so important?
What courses will exist? Are those vastly different from today's courses?
> AI generated or not, I concur. I rally want to know what Universities will look like in 10 years time. What will be taught there that cannot be taught by an AI (whatever form or interface it has).
Computer-assisted instruction been amazing unsuccessful. Why is that?
If people attend university for certification instead of education I think the battle is already lost. AI is an easy, but possibly high risk approach to gaining the certification without work, but the tried and tested approach is doing the bare minimum, cramming, then forgetting everything after graduation.
If you penalize people who use AI but in the process have learned the required information you make the problem even worse.
These problems are all because of a culture that favours the measurement over what is being measured.
The author mentions the notion of a "permanent underclass." What that looks like relative to your comment is this: more enterprising students (usually, but not always, from privileged upbringings) will learn how to develop the restraint to use AI as an assistant and a learning aid, while the vast majority use AI as an easy button.
Except it's actually worse than that because the big AI providers with billions on the line are actively encouraging their users to use their services as easy buttons. If they weren't, Googlebook wouldn't have a Magic Cursor/Clippy as a Service feature that suggests using Gemini when you jiggle your cursor over something, and Google Docs wouldn't insert a "Write for me?" CTA when you stop typing.
>If people attend university for certification instead of education I think the battle is already lost
It's been a lost battle for decades, then.
indeed. a certification is required for ~every non-manual-labor job. even if not listed as a requirement, you're almost certainly competing against people with degrees.
so, the primary function of going to school is to get a job, not for self enrichment.
Manual labor jobs require certifications in today's world as well.
It's a shit show everywhere.
All time.
What made University of California Santa Cruz so attractive to undergraduates in the 70s was no grading of performance and engagement but instead written evaluations. I loved being able to concentrate on great subjects with the help of often excellent teachers and TAs. Swathmore is noted for this culture too.
The purpose of the university has always been certification.
“You wasted $150,000 on an education you coulda got for $1.50 in late fees at the public library."
- Will Hunting
Do let us know where we can find a chemistry lab or biology lab or the facilities used for various other engineering undergraduates in a public library.
Not always [1]
Having gone to university in Germany, there are glimpses of this ideal, but they're mostly faint memories enshrined on faded plaques around the campus. I did have an old geezer prof (90+ years old) that went to the very same university over half a century ago and showed us his diploma: greek, latin, humanities, ... for a technical diploma, no less!
I do still cringe a little when we get newjoiners fresh out of university proudly proclaiming "Yeah, no can do, we didn't learn that at university". Yes, obviously, university is not an apprenticeship. You learn how to learn and then apply that to unknown-to-you problems. Oddly enough ChatGPT seems to have brought a change to that mindset, but Im not sure if it did so for the better.
[1] https://en.wikipedia.org/wiki/Humboldtian_model_of_higher_ed...
For US undergrad, with all the resources available online, going to university just for education doesn't justify the high tuition. Research is mostly limited to graduate school anyways.
The course material isn’t really what your undergrad tuition pays for.
You’re paying to be surrounded by smart kids that will ensmarten you too and for access to the school’s career networks.
This is not at all what most students go to undergrad for.
For some reason in discussions like this, things are always framed as if everyone is going to Yale.
Everyone I know including myself, went to college because that is just what you were supposed to do. That was my mom's dream for me to get a degree and I didn't want to disappoint my mom.
I am sure there were really smart kids in my high school that were figuring out post graduate network strategy but that was hardly the majority.
I even had a friend that dropped out of high school that ended up going to community college. He wasn't exactly figuring out the eigenvector centrality of the nodes of his general studies classmates.
Correct. Social pressure is what pushes most people there.
As a result of this social pressure, people do tend to go off the rails when there (having fun) very few give a fck about studying - who cares, there's no reward for getting anything beyond 60-70%.
>You’re paying to be surrounded by smart kids that will ensmarten you too and for access to the school’s career networks.
If you're paying and it isn't for a new campus building with your family's name on it, then you're not going to have access to those things at all.
Suuuuure, you're going to be able to access some sort of 'career' network. You'll be able to find amazing and high-paying jobs such as... um... uh...
Abbreviated:
- Surrounded by other kids who have access to the same internet you do
- Access to the school's career networks (LinkedIn)
I will challenge anyone who thinks they can get an equivalent quality of undergrad education and access to career services remotely.
Lol, nice to meet you. Left freshman year, got paid, got out. Questions welcome. College for tech jobs became useless after like 2008. Books are cheap. I'd challenge anyone in response to prove they got a degree in the 21st century.
This is mythology. I went to a very high ranked uni and this is not what any of my classmates were interested in either - it was all about getting drunk and pulling all nighters a few weeks before exams.
Career networks? Hahaha. Very few benefitted from that.
Once again, I went to a top 3 ranked uni on many league tables worldwide.
This is one of those things that seems true in theory but doesn't have examples of success. There are people who self-study advanced mechanics at home, but they're child prodigies and enjoy (typically) going to universities as a way of meeting others who share some interests.
Is learning something while you plagiarising someone’s work to submit as your own coursework should also be an excuse to allow plagiarism?
This particular battle for learning was lost a long time ago. If university stopped providing an earnings boost from attending, 90% of students would quit tomorrow.
It doesn’t help that a lot of desirable fields are comically out of date at the academic instructional level anyway.
Would you honestly tell an aspiring software engineer that your typical computer science degree will teach them much about wielding computers in a cutting edge way?
If I were to list the top 5 things I got from university, knowledge wouldn’t make the cut and were I to do it again, I would certainly attend less class.
> If university stopped providing an earnings boost from attending, 90% of students would quit tomorrow.
Maybe 10–20% would quit for that reason. There would be more attrition if you could get common jobs (such as teacher or nurse) that currently require a degree but don't pay that well without formal education.
Most people don't care that much about money. Sure they would like to have more money, but it's not the primary factor that drives their major life decisions. People are generally more interested in stable careers that pay their bills and seem like something they could continue doing until retirement.
Anyone who thinks that a Computer Science degree is supposed to prepare them for a job as a Software Engineer has completely missed the point. It's like getting a Physics degree for a job as a Mechanical Engineer. There is some overlap but a huge difference in focus on theoretical versus practical topics.
My computer science degree did not cover much actual computer science.
You can argue about whether that is a good thing or a bad thing, but the ship sailed long ago. CS undergrad degrees are about training software engineers, not about training computer scientists.
More and more colleges are saying they will cancel your degrees if they retroactively detect AI cheating. This will be hysterical.
AI isn't the issue as much how AI is used. Passive use of any tech, including social media, and now AI is lazy and has poor outcomes.
Aligning AI use with the goals of all sides, and not just one side getting paid, or just one side graduating could look different.
Most people attend higher education to access opportunity to improve their lives, overwhelmingly for a career and earning.
The idea of higher education teaching "learning how to think" is perhaps a relic of the origin of some universities which didn't historically do STEM, and focused on things like liberal arts, which in turn often had the support of coming from a privileged background, or financial safety net.
STEM money and funding though, attached a lot of traditional post secondaries to do that as well.
It's perfectly acceptable to expect higher education of any kind to have you ready to grow and earn more in better suited opportunities. Not enough educational institutaions don't publish their % of students who graduate in the area that they started in, and also the % of graduates who find their next step, career wise, etc, in 6-12 months of education.
You mean the same colleges that used AI proctoring during 2020 that had a lot of horror stories of false positives, like a guy passing in a window in the back of the student or another student that was not looking 100% of the time to the screen?
It is very difficult for them to prove the allegation of AI cheating. If you didn't cheat, you should fight. You have legal rights, and you don't have to roll over. I represent students who have been unfairly accused of AI cheating. Properly defended, it is not that easy for the school to win.
I have found the whole honor system of academics to be just another stick to whip students with when convenient. I am not opposed to any sort of honor system or academic integrity when there is actual honor and integrity, but the system needs to apply to both sides. None of this "rules for thee, but not for me."
>These problems are all because of a culture that favours the measurement over what is being measured.
Hear, hear!
> These problems are all because of a culture that favours the measurement over what is being measured.
Spot on. I am teenager going to college soon and I feel like the same way about the education system (and in extension, the job market but I suppose that the job market might be more understanding probably over all of it), part of my comment was as follows.
I do feel a bit like coding/a lot of fun-ness out of life is also like this, quantified, measured, transactional (posting for social-media?) [as I wonder if I am writing this comment for hackernews karma or relevant discussion talking points..]
This feels to me the most irreversible consequence because it might be hard for the generation (myself included) to see value in non-measurable things as everything has to be measured and transactional-ized.
(...) I would like for humanity to be more nuanced and less measured but more varied (grey rather than black or white) but I feel like that there is enough noise on the internet that maybe even this ends up becoming noise and I am not sure if anyone who might benefit from reading this actually does end up reading it.
From one of my comments that I had written sometime ago: https://news.ycombinator.com/item?id=47559013
> These problems are all because of a culture that favours the measurement over what is being measured.
What other kind of culture is there? A culture of not measuring?
A culture rooted in φιλοσοφία (greek, philosophia in Latin). So yes, I meant that literally. There were times, already 2500 years ago, where people wanted to study to become wiser.
It's great when you're a wealthy noble and have time to do luxury things.
For almost all of history, higher education has been a luxury good for the rich, including the Greek city states. There have been a few exceptions, most notably European countries with tax funded schools, but even those are primarily pumping out degrees used for chasing jobs.
The point of modernity was to broaden access and opportunity for everyone, not just the rich.
What we got instead was a regression to aristocracy, where critical social resources of all kinds were enclosed and captured by capital and financialisation.
The result is a lot of very broken systems, education being just one.
Universities became primarily about administration of property and income, and the educational element has become a form of marketing to attract money (and bodies) so the rest could function.
And now we're on the edge of the next stage, which isn't "What is education for?" but "What are humans for?"
We used to know. Or at least we used to believe we knew.
Now we don't any more.
I'll throw my two cents in, this still exists today. The difference between now and then is that now a college degree is seen as a requirement for a higher quality of life. (And not in a eudaimonia way)
Many people are going to college primarily to make more money in their adult life, the actual learning is secondary. If you're already well-off or just don't care, you can still get the education for its own sake.
The issue is that we've created a perverse incentive to get a college degree.
Read the last four words of the sentence you quoted. You'll find your answer.
Over a decade ago, my orientation at UChicago included the traditional "Aims of Education" address. They packed the whole first-year class into the chapel to explain, at length, that this education will not be "useful."
You're not supposed to make more money, or be happier, or really become anything other than a better version of yourself.
I wonder if they still do this.
Expanding this thought...
UChicago should be pretty uniquely positioned to address the problem of AI writ large. They already require a full year of each philosophy, literature, and history (all through primary sources). This "Core" should already be fairly AI-proof, given they are primarily small-group, discussion-driven courses; oral exams, in-class essays, or even graded discussions should be straightforward adaptations.
And yet, the university shifted towards professionalism before AI ("training a mind for the workforce" rather than "the good life").
Already, this transition did what the author observes AI is doing. I would hardly believe someone who cheats through an econ/stats major is less educated - if only through osmosis - than someone who honestly completes Business Economics.
And so I wonder - if the damage of AI is primarily instrumental to the broader trend of hyper-professionalism, what damage has it actually done?
If we automate away the signal to companies "yes, I can do stats for you," does that free students to focus more on the _less_ professional aspects of education?
Sure, it undercuts credentialism, making the "piece of paper" near worthless - but if our aim of education is just to "be better," should that not give us hope?
If the signal to corporations disappears the wage premium to a college education disappears and the students disappear along with the tuition paying the professors’ salaries.
Maybe? Elite colleges have been around a lot longer than the professional credentialism of the last 20 years, no?
The point was still the credentials.
Otherwise what signal does “prominent family and graduated from Harvard” have over just “prominent family”?
Surely some better versions of yourself would make more money or be happier?
Maybe - to your point, if we think of happiness as like “living one’s own purpose fully”, then yes, it does very directly.
But I was referring to happiness more generally as enjoyment, joy, satisfaction - that type of thing.
And in that case - there are plenty of ways being better = less happy. Eg if I were to sacrifice myself to save my family, then that’s the best version of me, but I’d be pretty dang unhappy about it.
Oddly enough my reaction to this is that it's a broader societal problem as opposed to an A.I. problem.
Why shouldn't universities switch to examinations where no technology (apart from say calculators) are allowed; and this is strictly enforced? This was certainly the norm when I went to university.
I agree that A.I. trivializes (or changes how you approach) a lot of take home work; but people who wanted to cheat could more or less always do so for that to some degree. I guess it makes it easier to do so; however my expectation would be a greater reliance or weighting on in person examinations as a response; as opposed to a normalization of cheating.
One way in which A.I. could be seen as contributing to this is that it is devaluing the importance of what were seen as 'intellectual' pursuits; as we now have automation for them that is at the very least often surface level effective for undergraduate work.
i always found it funny when people complaied about white board coding tests. Back in school, I had to write C code in those little blue books. :)
EDIT: I meant writing in blue books before this era of copying words out of the claude app on your phone
As recently as 2015 when I attended a middling CS program, we had in-person timed exams where we had to write down DSA implementations on a blank sheet of paper in Java.
We were deducted points for trivial syntax mistakes.
If these stories I keep hearing are true, then university programs have really taken a nose dive recently. This isn’t a “back in my day” thing, but within the past 5 years.
The pace of the purported decline makes me question if some of these stories are sensationalist. But I don’t know, I keep hearing about them.
I'm not sure I understand your comment. Surely you don't think that the details of a particular programming language's syntax are an appropriate criteria for grading an exam? That seems crazy.
I would suggest the one exception to this would be courses explicitly designed to teach how to use AI, and how not to. But in that case, it's less "use AI to cheat on this course" and "AI is the tool this course is about."
Otherwise your suggestion makes sense.
Then it becomes, teach what? "To use AI", yes, and, then, to do what? Use it how? To make some software? Why? You are already taking software engineering classes to learn to make software. To write something? Why? You are already taking classes that ask you write things yourself. An AI class, to me at least, is akin to taking a class about how to pay someone to write your essay for you.
And if we are talking about the various AI strategies people have where they have LLMs talking to LLMs to come up with whatever gooblyguck, are the poor souls who've been asked to come up with the AI class for the department going to know any of these strategies themselves? Are these strategies even going to be sustainable going forward after VC is no longer subsidizing tokens?
Regardless of VC subsidies, the cost of compute always trends down over time. Whether you like it or not, LLMs will be a pervasive part of everyone's life forever (or at least until a better replacement comes along).
Cost of compute trending down is usually lost as the resulting software bloat that fills the empty space like a gas. We already see this with LLMs. Models get bigger and bigger in an arms race.
OK so the frontier models will be capable of doing even more, but the simple stuff they can do today will get much cheaper.
I don't think that is a safe assumption to make. Moore's law is not playing out any longer as it used to. Jensen Huang already called it dead 4 years ago.
> (apart from say calculators)
Generally, i wasn't allowed a calculator in university.
If they havent switched already it’s because they are diploma mills and even the veneer of competence training has failed.
Some of these AI chatters may become oncologists.
The college system is creating the zombie underclass with AI or without it. The amount of money colleges charge combined with the text book thinking shapes people into thinking there are steps to success, there are right answers, and "getting a job" is the right way to go. Colleges don't teach independent thinking and that is the exact thinking we need in the era of Youtube and AI. You don't need college to teach you how to learn text book items anymore and I think that is scary to some.
> the text book thinking shapes people into thinking there are steps to success, there are right answers, and "getting a job" is the right way to go
Apart from entry-level texts, what discipline are you thinking of? Pretty much all my after-freshman-year undergrad texts contained debates.
> You don't need college to teach you how to learn text book items anymore and I think that is scary to some.
a) This is about universities, not "college" b) The University teaches you critical thinking, not how to learn "textbook items". It's not vocational training for upper middle class. It's for building and developing citizens who can think critically.
In the United States, "college" and "university" are generally used interchangeably. I absolutely agree with your second point and the shift has been going on so long that people are genuinely baffled by the idea that college should not just be "white-collar vocational school"
Ok but I need a job to live.
And if I want to do something interesting I need the skills and knowledge which are learned at a college level.
Most jobs, including white collar jobs, don't require the skills and knowledge learned at a college level. For employment purposes, a college degree primarily acts as a signaling and filtering mechanism. It shows employers that you can grind your way through a moderately difficult task without giving up, and it allows hiring managers to reduce the number of applicants to a manageable level. When actual labor shortages occur, employers are quick to drop the fake degree requirements in order to fill critical roles.
> Ok but I need a job to live.
Not really, you need cooperation with other people in this complex world to live. No necessarily a job. You could be self-employed or a member of a cooperative or an elected official.
But yeah, the capitalist default is to have a job, sure.
> And if I want to do something interesting I need the skills and knowledge which are learned at a college level.
Not really, no. You need the skills and knowledge and for some professions you do need the official certificate of education and for a subset of those that's actually warranted, because you cannot get your hands on the training other ways. Doctors kinda need the official system, self-taught appendectomy would not be ideal. English literature? Not so much.
You need a job to live - under the current economic system. a lot of the critiques of capitalism lead you straight into ai, and most economic critiques of AI are critiques of capitalism as it exists now.
You've needed a job to live for all of human history going back to whatever ancestor you like. That job is called survival. All modern life is doing is delegating that effort across many people such as to benefit from economies of scale. I don't need to fish for my sustenance because someone else is doing the fishing. I contribute in some other way that gives me credits to access the fishing take. People act like capitalism is some new evil but it is little different than the delegated labor you might find in a prehistorical tribe. The fundamental rules of this system are really the only way our species survives. The issue is sometimes compensation is uneven for the effort, but that is really it, not that the system itself is a failure. I'm not sure of any system that wouldn't be like this. What is communism but capitalism where compensation is more evenly distributed? It is still a game of delegating certain tasks to certain individuals such that one individual is not responsible for 100% of their own survival.
Survival didn't involve a CS job at FAANG until very recently.
Capitalism is not a new evil, you're absolutely right. It's just the system that, according to our current framework, came after feudalism, which came after something else.
the point is that actually seeing the changes in systems as consequences of massive technological change is likely directionally correct and it is not clear whether current capitalist logic survives the AI era.
What is a CS job at a FAANG in practice? Many hats. For example if you work at Netflix on content delivery, you are the modern version of someone involved in the boring logistics of theaters and plays. If you work at Google on search, you are the modern version of a librarian establishing ways to catalog information. If you work at say facebook, you are the modern version of a propagandist or even town crier perhaps. In all these cases, these workers were delegating their basic survival needs to other members of society. Ancient Rome even had its own slop bowl restaurants not all that different than what you see in SF today, and many ancient romans did not cook for themselves but instead traded credits gained from labor elsewhere to eat at these establishments. No different than lunch time in SF today.
The actual tasks being done have always been done even if the technologies used to do the work have changed and the work itself looks a little different. Our base needs have not changed. We are the same apes bound by dopamine. The functionality we require is pretty old. There is probably little we experience that is truly "new" compared to what one might experience in a classical civilization.
Capitalism is far more ancient than feudalism. Phoenicians were famous capitalists, likewise most any group of people involved in trade. Trade and accumulation of capital representing labor is ancient.
You're just confusing definitions. Capitalism as its commonly thought I would think just about what it says on the wikipedia page
"Capitalism in its modern form emerged from agrarianism in England, as well as mercantilist practices by European countries between the 16th and 18th centuries. The Industrial Revolution of the 18th century established capitalism as a dominant mode of production, characterized by factory work, and a complex division of labor. "
Trade schools and apprenticeships exist.
Trades can pay very well and frequently require nothing more than on the job training.
You think you need college for the same reason you equate "job" with survival. These are not universal truths, not even in capitalist hellscape America. It might be harder but it is in no way a requirement. Anyone who tells you different is lying to you.
It does feel like there is an easy solution to this:
Have tests.
Supervise said tests to make sure people don't cheat.
That's how it worked when i was in university. Admittedly maybe that is easier in the sciences than humanities, but still, it seems doable. Cheating isn't a new phenomenon it just got cheaper and easier.
The tests will be made by AI as well, because professors who spend less time on making tests can spend more time on research and get hired/promoted more. For the same reason, professors won't care who is using AI to cheat on the test that was made by AI. Maybe some people will care, but not enough to do anything about it.
(Cheating was already rampant in many classes 20 years ago when I was in college, I can't imagine what it's like now.)
This selects for people who excel at taking tests. Sure, there may not be a better alternative, but as an empirical measure I learned just about nothing in my test heavy college courses, as I was incentivized to cram for the exams and purged everything immediately after the final (and midterm).
Yes because a linear algebra exam tests your test taking skills as opposed your math ones.
You sound like the type of person people can't stand to be around let alone work with.
The kid that never got past getting picked last in gym class so has to constantly remind people that is no longer the case.
OK but why? That will reward students who are good at taking exams, which has very little relevance to being a competent worker or well-rounded citizen or innovative leader or anything else that we need.
Since when has university ever been about "being a competent worker or well-rounded citizen or innovative leader"
Since forever?
Right? It's exactly how we train AIs, after all. It's not like this is mysterious.
Overall it’s a very good read, really enjoyed the author voice and style
> I don’t think she was laughing two years later when I was TAing the class and we observed a fairly distinct gap of about 40 percentage points between the take-home test and the one administered in-person.
40pp is massive. Take homes are pretty much dead at that point. And not just in schools, but also for interviews. I don’t see how you can get a meaningful signal, it’s guaranteed they will be made using AI.
I had very few take-homes in school (discounting project courses which were very time consuming). One of the problems with take-homes in most interview situations is that, at least for developer jobs, the reality (whatever the stated guardrails are) is that this will be at least a weekend project. That said, I've spent multiple days traveling and doing related things for an on-site interview which I realize not everyone here would find acceptible.
The take home tasks are only useful for assessment if paired with an oral question and answer session with the instructor about the work turned in.
It becomes apparent really fast which students just delegated the work to AI.
Of course it’s also much more effort for the instructor.
Of course the take home interview assignments will be made using AI. Isn't that the point? For most white collar jobs we should hire candidates who can make effective use of AI.
Tangental, but is Leetcode dead too? Part of me could foresee Leetcode being believed to be a stronger indicator now, since it would be unlikely that someone that can answer those questions would know absolutely nothing about programming.
Then again, with LLMs, learning to answer those questions is probably easier now than ever.
I haven't applied for a job in a decade, so I am utterly clueless on the current landscape.
Profs can push back on this if they want. Not all of them want to (or want to justify pushback given their pay).
For me when I teach, no laptops or phones in class along with in-class handwritten paper quizzes on course readings and concepts has helped a lot.
Clearly not all of them want to do so. My son just finished up his freshman year at college. A couple months ago he called me and told me about an interaction he had with his English composition professor. She asked the class how they were integrating AI into their writing process. When she got to him, he said he didn't use AI, as he wanted to use the assignments to learn to write better. Her response was, "Well, you're probably too advanced for this class then." This is my son, who has a dysgraphia diagnosis, being told he is "too advanced" for freshman English class because he doesn't use AI to do his writing.
That is fascinating to me. Especially in English composition. The flip side is if it's an adjunct prof making $4k per class (English typically pays poorly) then she's doing the hourly rate calculation and thinking that AI is going to help her with the students. So, a potential solution to this is for the universities to have the willingness to pay their professors and ask for no AI (at least in this type of class) in exchange.
Sounds like bitterness and resignation, to me.
Sounds like a poor and offensive attempt at humor.
It's going to be interesting to see where universities are in ten years. Higher education has for a long time been a bad proxy for internships and apprenticeships, because if you aren't teaching what the market is paying for then you don't get funding or students. Watching universities start to defend against this obvious and accelerating association is fascinating. There's an obvious decoupling between "learning how to do a job" and "higher education for fun because I'm rich" that we're only going to see getting crazier.
the screenshot with formulas and then, in the middle of it all, "wait let me be more careful" had me laughing to myself.
Same. Some of the developers I work with leave Claude's comments in the code: "Wait... let me try something else!" I don't mind they're using AI, but leaving junk around is a code smell. What else aren't you paying attention to?
Universities have the greatest discrepancy to me between how they're described from those who are no longer associated with them or who are weakly associated with them and the reality of those who are dependent on them as employees or as students. To listen to the alumni, every university was an institute of learning where all standards were held to the highest and everyone adhered to every rule to the letter.
To the majority of students, they seem quite laser focused on acquiring the degree with the right grades so that they can maximize their chance of a job after university (apart from the personal element of partying and having a boy/girl-friend etc.). The primary utility of the university to students is the credentialing, and secondarily the structure to the learning program, but otherwise the books themselves suffice to teach.
Perhaps we should move more training to technical institutes and people can come out with the knowledge of how to operate this or that thing. The problem is that everyone will know that the smarter student has gone for the higher-end university. The credentialing then works not because of the program but because of the selection that the university can do. Okay, so the whole thing continues to make sense even if AI zombifies everything.
AI really does destroy everything it touches.
Back in my day, we did the learning in class, and the evaluation (homework, projects, but sometimes even exams) at home. Seems like AI just flips the script. Now, you can learn anything you want with a private tutor. The teacher can just upload the entire course material to a place where the LLM can rag over it and answer anybody's questions. Learn it on your own time. Then, use class time for evaluation. Do the "homework" in class. Take proctored tests without access to the internet, etc.
I wrote something along very similar lines recently! Even the zombie metaphor is quite close.
https://pistolas.co.uk/work-that-need-not-be/
> The buildings, of course, will remain, to be observed and treated respectfully — like old cathedrals, mainline Protestant churches, and most of the European continent.
Ouch.
The universities can survive a wave of student ai laziness- blue books and pass fail homework. cfg. I can’t see how they survive a wave of professor ai laziness. Paying $200,000 for $40 of tokens is such a brutal way to be scammed and the resentment will get refreshed by loan payments every month for decades
I'm a big fan of the fact that the following quoted sentence has 25 words before the verb. An LLM would never. I probably wouldn't.
> Whatever your conception of the modern university, whether grand or grim, understanding the current landscape of campus-wide AI use, much less its intensification, should destroy it.
> Tying education to a capital-intensive and (likely soon to be) tightly regulated technology is one more step toward a different, frightening future. A world in which independent educational institutions are neutered and transformed by their reliance on a central authority into factories designed to train students according to the “needs of society” is not a new prospect — it has been the persistent dream of Fabians, technocrats, and engineers…
I hadn’t thought of this. Every school district and university tied into centralized AI inherently undermined its ability to decide how its kids are to be taught.
Universities will still act as gatekeepers of prestige and status. There is no AI alternative to the top-20 schools...I remember all the hype from 10-15 years ago about how online learning and "MIT courseware" would upend the universities or threaten credentialism, and nothing even close to that happened. As it turned out, the online version of MIT is not a substitute for the actual thing.
Schools will adapt, as they have already, by weighing grading more towards in-class quizzes and tests . I think the humanities will continue to struggle, but I see the AI boom making STEM more relevant, even if AI can automate a lot of code or math.
>As it turned out, the online version of MIT is not a substitute for the actual thing.
More precisely, the people motivated enough to actually do the online MIT version were often already on a high-performance trajectory, and for the people who were not, few people took the online credential seriously, despite whatever skills they acquired.
All the courseware, classes, and schooling in the world cannot teach one to think.
Educators absolutely can teach students to think. The scientific method is one example of a key mental tool which provides an organized, disciplined framework for thinking. If you read a lot of stuff written before the scientific revolution it's kind of a mess because the authors literally didn't know how to think. When they occasionally got things right it was mostly by accident.
Of course they can. What the heck?!?
Logic 101 changed the clarity of my thinking markedly.
In what ways?
On the other hand, it would be really good if universities stopped being gatekeepers of prestige and status. It seems like some of the biggest idiots in high visibility posts right now come from the ivy league..
I'm not sure if an ivy league education proves anything anymore other than that you're connected.
That’s because the value of the university is the assessment, not the education itself.
They are vouching for the intelligence, knowledge acquisition and work ethic of their graduates. If they lose that signal, they lose the ability to gate keep prestige and status.
> by weighing grading more towards in-class quizzes and tests
The piece discusses blue book tests where students were still cheating with their phones providing AI responses
that's a proctoring problem though, no phones during a test is typical to say the least.
And yet a Top 10 school like University of Chicago has apparently not been able to fix that problem.
That's telling in and of itself.
One professor in one class has not fixed the problem. You can't generalize that to the whole university. Even for that one professor/class, it's not the case that the problem can't be fixed (it could be, quite easily), it's that the professor evidently cares so little about doing their job as an educator that they simply should no longer have a job in a university.
AI camera watching the students?
Faraday cage. EMP blast if that doesn't work.
Body cavity search
https://www.npr.org/2023/09/26/1201734274/chess-hans-niemann...
Somehow my EMP blast idea is the less intrusive one...
I was forced to get a cloud certification while working for a former employer. I took it remotely and had to leave the camera on the whole time. If you looked away even for a second, you'd get a warning. Very dystopian.
I'm a current student, who also happens to be a full-time professional who is "all in on AI", and I think most are missing the true opportunities AI opens up for education.
Because my student path is non-linear (vs just following a life script), I may be a bit weird / not the average student, but it's especially true for me that I'm very intentional about actually learning the things I sign up for classes to learn.
My point is that I'm not taking classes just for the motions or to create slop. With that context, here is how AI helped me very specifically in a recent linear algebra course:
1. I was able to prompt very specific questions, usually audits of my work, in ways that provided responses that were more like a socratic tutor and not a cheating parter. In this way I did not need to bother my professor as much or seek out a tutor, when I was stuck. But I also didnt shortcut my way to answers. I was intentionally limiting the AI assistance to finding small errors or jogging my memory about steps missed or next steps.
2. I vibe coded a note taking web application (started as a chrome plugin for notion) so that I could shortcode and pick math symbols while my other arm was full holding my newborn (yes I'm a dad too). This has since evolved into a full-on science writing platform that I love whether or not anyone else ever uses it (though I am trying to turn it into a business). Maybe I actually ended up adding more work to my math class but it added a layer to the learning (what math symbols are needed, what are typical patterns for this subject, etc) that I think helped with my overall absorbtion of the subject.
I dont know if #2 is transposable to other students or to other subjects but I imagine there is some version of a double major yet to be created that is Core Subject + "how to properly use AI to learn (including vibe coding tools to help yourself and other students)".
There are many other smaller ways AI can be used to help learning (flash cards, generated quizzes, etc) that are oft mentioned but that articles like this gloss over.
Having said that, I loved reading this (so well written it could not be AI despite the emdashes), and especially appreciate any mention of "The Whispering Earring", which is one of my spinning tops to remind me to remain vigilant of my cognitive health despite my almost complete embrace of AI.
I can't help but wonder if the fundamental problem is just that we spent decades pretending that a university degree was some sort of useful job training in the first place. As a professional software developer, I think my computer science degree is not actually all that important. Sure there are some relevant concepts, but they're ones you'd pick up on the job anyway.
I don't regret getting my degree (back in 2009), but I think requiring a person to have one is a dumb job requirement.
Frankly, we shouldn't have so many people going to university in the first place. There's a lot of people it's just utterly wasted on, and it drags down the entire apparatus as a result. In a sane society we'd have much more apprenticeships, vocational training, etc.
Even though the degree really didn't teach you what you'd do on the job, it was a signal to employers that 1. you were capable of learning stuff when needed, 2. you didn't give up on doing hard things, and 3. they wouldn't have to explain to you what a loop or a class is.
With rampant AI cheating it's no longer a guarantee of any of those.
I do not think I had a single university course that was not ultimately assessed through supervised written tests. That matters, because homework was never a reliable anti-cheating mechanism. Even before AI, you could just pay someone smarter to do your homework for you, or ask older students for their assignments (and sometimes even their test questions, there's even a pg essay on this [0]).
You can cheat as much as you want on homework, but it won't help you on supervised written tests. At some point you have to sit down, unaided, and show that you can solve the problems yourself. So I do not see how AI substantially weakens the signaling value of a degree, at least in systems where the degree is backed by in-person written assessment. It may make take-home coursework less meaningful, but that was already the weakest part of the signal.
[0] https://paulgraham.com/lesson.html
I don't know, I went to a pretty well regarded CS program but some of the people I graduated with are definitely not people I'd want to work with. They might have hypothetically known what a loop is but I wouldn't expect them to put one to good use.
For Computer Science in particular: it's not supposed to be job training. CS is an education closer to math or science (in fact, at my university it belongs to the department of hard sciences and math). If you like that (and I sure did!) it'll be worth your time. If you're just looking for job training, you're looking in the wrong place.
My university CS program didn't even teach programming in any of the major classes, it was assumed you'd learn on your own or by doing one of the optional workshops.
There's a lot of stuff taught in academic CS that you simply won't learn on the job, or if you do, it won't be as rigorous and you'll be missing the fundamentals.
What program was this? It sounds much closer to what I wish my CS degree had been. In my CS program, the courses I actually enjoyed were the math-heavy, but always optional, like computability/decidability/complexity, cryptography, etc.
The mandatory "practical" courses were often much worse. For example, I studied relational algebra on my own, plus a few chapters from Kleppmann's Data-Intensive Applications book, and it was painful to realise how shallow it made the mandatory database course look.
I agree that CS should not be mere job training. I think many CS programs are neither rigorous enough to feel like math/science and prepare you for proper academic work, nor practical enough to be good vocational training. They sit in a bad middle ground, where academics teach industry-lite.
Industry-lite is a good way to describe it. I remember having a 300 level class which was supposed to be about real world application architecture but it was essentially just about making UML diagrams (because the professor happened to be on the board of whoever was in charge of UML.) Nobody serious (even at the time!) uses UML.
My computer science education has been invaluable for my career.
See also: https://news.ycombinator.com/item?id=48139148
Zombification of USA people was already happening before AI. Not surprisingly, it's been one of the favorite cultural themes in the country's cultural produce for long years. Zombies + superheroes was not poised to produce great non-drooling non-moron americans.
> There is an extreme idealist view of education that might see the threat of AI as good precisely because it could transform those kids — the former connoisseurs of SparkNotes and Mathway, the ones snickering in lectures and inking formulas onto their palms before exams before the rise of generative AI — into zombies lurching and stumbling their way into the “permanent underclass” (as the tech bros say), leaving the elect few free to enjoy the benefits of a humanist education without all the noise and din.
In this world, what are the benefits of a humanist education? The only reason we care so much about education is that it's how to determine merit in meritocratic societies, and therefore a key part in how people gain social status. In a world where AI does all the knowledge work and robots do all the physical work, with an 'elect few' owning everything and everyone else in a 'permanent underclass', why do the elect few even need to keep the permanent underclass alive?
I think that universities just have to adapt to deal with slop, or think of new ways to challenge people to learn the essence of their studies. I wouldn’t want to be a uni teacher in these times though.
It's not hard, just unpopular. End credentialism, stop giving out grades or administering exams.
The (Western) educational system still does this for PhDs. Grades barely matter, and in most places you have two oral exams in your entire 5 years: your qualifying exam, and your final defense.
The reason it doesn't happen for the rest of the system is scaling. The US awards about 60k PhDs per year, compared to about 2M bachelors. There simply are not enough faculty and it is not realistic to hire enough (if there are even enough qualified people in existence)
And that's ignoring all of the problems with "not giving out grades" or "ending credentialism" - I guess people are supposed to just get hired on vibes?
Since everyone has a meaningless degree already the degree absolutely does not get you hired for over a decade already. Interview outcome, which sure you could call vibes
Degrees don't get you hired, they get you interviewed in your first couple of jobs.
I'm not going to interview some guy with no work experience and no credentials for an engineering position. I will gladly interview someone with no work experience but a relevant degree or bootcamp.
The solution is obvious. Teaching must be no-tech—just go back to 1950s.
The other problem of course is attention span due to social-media erosion.
The big tech has really done a number on society already and they’re just getting started.
I agree with you that no-tech parts of universities would work - obviously you can't avoid tech when teaching some things like coding, but mostly I think it would be a good idea.
There are problems: Having students attend lectures is great but they have to work with the material and prove they understand it - how to do that without homework? I'm sure there are ways. Have them work in a building full of computers cut-off from the internet maybe, but how to keep them from using their phones?
Another option is just severe comprehensive testing in heavily inviglated rooms long after they finished the class involving the material to prove they know it. Perhaps you could do this for the first few years of knowledge in a discipline and then assume the student actually is serious and take the leash off after they passed the tests. I know some disciplines already do this kind of thing, even before AI. Basically everyone has to pass a bar-exam type thing, even if they're studying art - but things like art can't really be condensed into an exam and it would certainly restrict and narrow what can be taught and learned, that's a big problem in my mind. Also what if there are new ideas in the study of physics and they can't really be taught because the exam is too difficult to change quickly? What if there's a big split in the philosophy of buisness, but the exam only asks about one side of the split? What if you have an ingenious professor who wishes to talk about a new branch of philosophy he's created - not on the exam though.
Edit: I guess if professors designed their own exams, instead of some distant exam-comittee it would alleviate most of my concerns about them.
For coding you can actually teach students on commodore 64s. It’s actually better because they have a BASIC shell and assembly language. Most importantly, no internet. :)
Actually, give them internet why not. But they have to use a 56k modem. Mwhaaha
I would actually love taking a class like that.
> ...how to do that without homework?
Tests. Many of my university courses only graded on tests. They strongly encouraged you to do the homework to better understand the material, but didn't consider homework completion when calculating your grade.
Consider that universities are educating adults who are -often- paying to be there. If we assume competent course design and instruction, if an adult chooses to not work on the material until they understand it, then the only person they're harming is themselves... which -as an adult- is a thing that they're usually fully entitled to do.
I liked the classes I had with a "reverse" format: the "homework" was done in-class, checked for correctness but not part of the grade, and the lecture was a recording watched at home.
Huh. I never took a class like that, but that seems like a great idea.
but how would you do that? what about homework and coursework? students will just transcribe claude slop on paper and submit that.
You give exams in person, in class, on blue books, no phones. This part isn't hard. Instructors have been doing it for generations. It's only in the post COVID era that some have moved to having exams take home and on Canvas or similar platforms. This is great for instructors -- less work! but I am not convinced it actually helps students.
The part that is more difficult is take-home work, and I think the solution is that instead of being something that you turn in for credit, it needs to move to being more of a chance to practice for in-person exams.
What about essays? I've taught classes where students had to write essays in class, in person. On paper, with a pen (this may no longer be allowed on many campuses because of access and perceived fairness reasons, which IMO is a shame, but it is what it is). I think the traditional assignment of "write a 15 page paper on XYZ" is probably done. Instead students will have to prepare to write an essay in class by reading the source material (books, papers, etc) and converse with AIs that are hopefully not hallucinating, to get an understanding of the material and then come to class and be prepared to write about it.
It's a new world, but one we can adapt to.
Assignments, sure. But if tests/exams are proctored in-person with pen and paper, the students may quickly pivot to traditional learning methods if they want to pass their courses.
I thought it was widely agreed that most homework was literal busywork
Do it on the classroom or it doesn't count
Requiring them to write it in longhand at least removes the instant gratification. I think that will work for some students.
Oh how I hated having to write in longhand throughout my schooling. It was always so slow and painful no matter what I tried in terms of pens or techniques. Typing is much faster and more fluent. In the real world no one writes longhand any more: it's all keyboards, swiping, and dictation.
It would be fun trying to draw all those emojis in longhand. You could probably do the bullet points as tiny manicules.
It would actually be interesting to see what people do attempting to transcribe AI generated material to paper. At the very least it's another layer of learning in writing it out.
I dunno think outside the box.
One option… They can do homework just test them every week in class. Homework doesn’t count for grade anymore. But test questions based upon homework.
Another… kids do reading at home in textbook, then work together in class to finish. Adjust hours accordingly.
There’s a very interesting problem space here though, to “disrupt” education by going back in time and applying a modern spin on education.
Engineering at my EU uni, homework and coursework were at most a tiny part of the total grade, and never enough on their own to pass. If they were relatively bigger projects, you'll pass an interview or similar review after delivering it. This all were just nudges study and to check ourselves and they were seen as a "gift" of the Bologne Process (restructuring/standardizing of unis in the EU).
The only thing that mattered were the exams, be it pen and paper or coding/electronics labs, in person and proctored. No matter how much slop I could have access to back in the day I would have failed the same subjects I did.
In-person tests and workshops, including oral exams.
Like we'd been doing for literally hundreds of years.
This is already being done. I teach computer science at bachelor-level and all exams are in-person. We talk through the code.
More in class, in discussion, and less "assignments"
Unfortunately that's way more expensive to do.
for STEM topics, I feel like some amount of "personal study time" is kind of needed to really grok stuff, at least for a percentage of students.
I studied maths, and spending time alone trying to solve problems and redoing the proofs from memory was important for my learning.
I don't think I'd have learned as much had those moments been replaced with more in class discussion.
Yes but AI doesn't prevent you from doing that learning. What is makes harder is the old broken ways of credentialing your learning. The way you do the learning has no need to change.
I personally feel like the software engineering profession may have to start moving more towards an apprenticeship model than a theoretical CS-gradate-then-work model.
Internship / coop programs at places like Waterloo already look a bit like this.
Yes you would. Trust me, you would
Slop made by students is one thing, but slop generated by facilities and fed at extreme premium to students just asks a question "why someone would pay for this instead of buying some LLM tokens, taking curriculum and teaching themselves".
If we want to teach students to use AI, it should just be a separate course, not shoving it in every possible nook and cranny to the point it is teacher AI talking with student AI with light supervision from both AI handlers
> The prevalence of AI use on college campuses, particularly at “elite” universities, is a cancer on our culture that threatens to turn a generation of promising young Americans into a class of drooling morons...
Modern education is like that, even before AI. Check this https://www.jstor.org/stable/25006902
> Check this
Evidence of people complaining about a thing isn’t evidence of the thing per se.
How exactly is it that you think Enzensberger is arguing against nwhnwh's point. Will an excerpt from Discourse Networks 1800/1900 be far behind, be still my heart
Right. I got my bachelors degree more than a decade ago, and they did a good enough job of it without AI. College was the first time in my life that I'd ever heard, "You cannot aspire to [ambition]," and I heard it there quite a few times.
Higher education needs reform more intensive than a simple defense against LLMs (as does the legal system and profession, as does the software engineering field, as does the field of psychology/psychiatry, as does-).
This whole piece is AI generated.
The prose is a bit too purple and tortured for that, IMO. Stock Opus 4.7 or 5.5 Pro is a more disciplined writer.
And, anyway, the point the article is trying to make is obvious. What's absolutely not obvious, and what it sheds very little light on, is what the University is going to look like in 10 years. Not what it should look like, but what it is most likely to look like.
> what the University is going to look like in 10 years
Mostly like they look like now, probably. With slightly more strictly enforced rules around exam.
I fail to see why it won't be like that.
You may be right. But that will mean the university becomes increasingly irrelevant and disconnected from the reality that students experience after they graduate.
I read a lot of AI prose three days and this bears none of the hallmarks. If this is AI, if really live to see the prompt.
I'm confident this is human.
It is easy to change the system prompt to make the AI talk with a different voice. It is remarkably hard (at least for Claude, I haven't experimented as much with GPT) to get it to not use so many em-dashes like this essay does.
There's no way. Just the first paragraph alone is enough to convince me; it's too well-written and melodious to be AI, with too much original thought:
Today, the demonic vice of the old is not that they are hard and demanding on the youth — instead they do not demand enough from us, and they cannot quite believe that we have not lived up to the little they have demanded. They think too well of our generation.
Without defending the quality of the rest of the essay, it's a great start. LLMs today could never match it.
Style is the wrong diagnostic. Purple prose and em-dashes can be prompted in or out. The harder question is whether the reasoning was committed or generated. A distinctive voice tells you nothing about whether the person actually worked through the argument or had it produced for them. Which is sort of the point the essay is making about students.
This is such an ignorant trope. The last few places I worked ALL used em-dashes as part of house style and I will continue to use them. It's extremely common (and arguably the LLMs do it because it is extremely common).
I can tell you with 100% certainty this is just how UChicago students write
It sounds to me like how I'd imagine a Philosophy student at the University of Chicago would write.
As noted in my own parent comment: I loved reading this (it's too well written it could not be AI despite the emdashes), and especially appreciate any mention of "The Whispering Earring", which is one of my spinning tops to remind me to remain vigilant of my cognitive health despite my almost complete embrace of AI.
this comment is ai generated
AI generated or not, I concur. I rally want to know what Universities will look like in 10 years time. What will be taught there that cannot be taught by an AI (whatever form or interface it has).
Will Universities still be centers of knowledge and exploration? or will that be more disseminated through society, and so Universities not so important?
What courses will exist? Are those vastly different from today's courses?
> AI generated or not, I concur. I rally want to know what Universities will look like in 10 years time. What will be taught there that cannot be taught by an AI (whatever form or interface it has).
Computer-assisted instruction been amazing unsuccessful. Why is that?
Kinda glad to see it as universities have made a mockery of education and learning for decades; hoping AI just replaces them altogether