A lot of the purposes in education for which the use of AI would be considered "cheating" involve writing assignments of one sort or another, so I don't know why most of these education scenarios don't simply redirect the incentive.
For example, in an English class with a lot of essay-writing assignments, the assignments could simply be worth 0% of the final mark. There would still be deadlines as usual, and they would be marked as usual, but the students would be free to do them or not as they pleased. The catch would be that the *proctored, for-credit* exams would demand that they write similar essays, which would then be graded based on the knowledge/skills the students would have been expected to gain if they'd done the assignments.
Advantages:
- No more issues with cheating.
- Students get to manage (or learn to manage) their own time and priorities, as is expected of adults, without being whipped as much with the cane of class grades.
- The advanced students who can already write clearly, concisely and convincingly (or whatever the objectives are of the writing exercises) don't have to waste time with unneeded assignments.
- If students skip the assignments, learn to write on their own time using ChatGPT and friends, and can demonstrate their skills in exam conditions, then it's a win-win.
This all requires that whoever is in charge of the class have clear and testable learning goals in mind -- which, alas, they all-too-often do not.
A lot of students, even at the college level, don't think that far ahead and make bad decisions because of short term thinking.
Look at any list of advice for new college students and almost every one of them includes "go to class". Simply attending class is way easier than homework and yet, when there's no short term consequences for not doing it, plenty of students will just not do it.
Cheating is another great example. Cheating in college is rampant because kids don't want to do the work they're assigned. I don't understand the logic behind the idea that if you tell all the kids currently using ChatGPT to write their essays, "Hey, you don't actually have to write that essay at all" that you think they will somehow choose to write it anyway. They're already choosing to ignore the long term benefits of homework even when there are short term consequences, so I don't see how removing those short term consequences will make things better.
If you tell kids there are no immediate consequences for not doing homework, many of them just won't do it and they will fail because they haven't learned anything.
Maybe you're okay with that. Honestly, I'm not actually trying to convince you that it's a bad idea. I just think if your proposal is based on the idea that kids will choose to something boring that they don't have to do in the short term because it benefits them in the long term you're overestimating a lot of kids (and adults for that matter).
There's a section in Zen And The Art Of Motorcycle Maintenance (1974) where Persig takes this all the way to the final conclusion that there should be no grades at University, and no degree at the end, and then and only then will everyone who goes there actually be learning-motivated.
So. I teach at a university and I do give an "assignment" exactly like this.
In a few of my classes, I have final projects that teams work on. I also have presentations. I used to require them of all students; and quickly learned this is a good way to waste valuable time.
Now, all my presentations are completely optional for NO CREDIT. You don't get penalized if you don't do them, and perhaps more importantly, I give ZERO EXTRA CREDIT for doing them.
As you can imagine, every single presentation I've gotten from this has been absolutely worth it.
The problem is that the motivation from above (i.e., administration, state legislatures, employers, etc.) is no longer really about learning. We could have an entirely learning-motivated university right now and it would be considered a bad thing by many powerful people because it's not aimed at "preparing people for the workforce" (in part simply by providing that degree).
> I don't understand the logic behind the idea that if you tell all the kids currently using ChatGPT to write their essays, "Hey, you don't actually have to write that essay at all" that you think they will somehow choose to write it anyway.
I unironically believe if you tell all the kids they don't have to write the essay at all, much more will choose to write it.
Kids cheat not just because they're lazy. Cheating makes people feel smart. The fact you can get credits by doing very little while others work their asses off is rewarding and self-validating.
The big issue of exam-only approach is that a one-hour exam is not enough to evaluate a student's performance, unless your educational goal is just to make students memorize stuff by rote. I'd consider a 3-hour open-book exam bare minimal. But if every class does that it'll be too exhausting.
Do Universities no longer do that? All of my finals were 3hrs. There was a special schedule during finals week with 3 slots per day. The time of your final exam was based on when the first lecture session of a class took place. Really sucked to get an 8-11 AM slot when your classes never started before 11.
Fun prank: set all of the clocks in your dorm neighbor’s room to different wrong times. Guy across the hall knew we were messing with him, trusted his watch - which had the correct time, but wrong alarm time. Realized he had a problem when he had hot water in the shower and no one was around. He was only 45 min late to the exam. Good times.
This, in general, seems like a great thing. The goal of a university should be to produce premium students, and nothing's better than a trial by fire.
We actually had this exact thing at my university. One sophomore level weed out class was a "self paced" electrical engineering class. It was called self paced because you were given a textbook and were free to work through it at your own pace. But to finish the class by the end of the semester you had to average 2 chapters completed per week, and completing a chapter not only included finishing a problem set and taking a test which you had to score 90%+ on (and were required to finish another problem set and retake it otherwise), but on occasion also demoing some skill in the lab.
It was brutal, but one of the most educational classes I've ever taken - and obviously not just because of what I learned about electrical engineering. Of course it seems modern universities have just become profit-driven degree treadmills. Weeding out students? That's reducing profit! And yeah looking back at my uni's page it seems this class is no longer self paced. Lol. And that's at a top 10 school. The enshittification of education.
Part of the issue is with the purpose as you describe it. Sure, at top 10 schools, a trial by fire would result in much needed “growing up” as the gifted but undisciplined (speaking for myself and many users of this site) students find their way to more durable motivations. But at the vast majority of schools, a trial by fire would end with a lot of students burned.
Perhaps that begs the question, if those kids can’t handle self-directed education, why are we putting them there in the first place, but that’s definitely a grey area, and there are hundreds of thousands of students who are smart enough to do well in higher education and skilled work, but weren’t disciplined enough to handle what you’re describing as freshmen.
Weeding out as I've seen it is a class that requires a certain level of commitment and ability to either plan your work or tough it out that a high school just can't really prepare anyone for. So in a way the student isn't a "weed" but their motivation or maturity might be and they're free to retake the class once they know that university will require them to put in more work than high school.
If they can't put in the work then completing a thesis and graduating is going to be very hard and that happens the last year of uni so better to set the expectations early with a "weed out" class.
On the class topic, I suspect that attendance was more impactful for students pre-internet as the alternative was to wade through the library piecing together material.
With lecture notes/slides available online, well prepared books and study forums readily available - in-person attendance can feel archaic.
We may be experiencing a similar dynamic in education with AI. In a world where we can create individualized curriculum’s for each student encompassing the entire tree of knowledge - Perhaps it’s time to rethink how we educate students rather than push them into lecture halls designed for the Middle Ages.
People thrive under regularity, and young people (especially) tend not to understand that. Similarly, being able to focus on a single thing is a kind of super-power, while multi-tasking generally hurts performance on tasks.
Going to class (and paying attention) means that you've got a regular period of focus on the class topic. That combination of regularity and focus translates into long-term learning and better performance.
> the assignments could simply be worth 0% [..] that the proctored, for-credit exams would demand that they write similar essays.
We run university programs at my company, and arrived at this bit of insight as well. That said, some of your points are incorrect or incomplete:
- You can't build systems assuming responsible individuals. These systems are guaranteed to fail. Instead, assume individuals are mould-able, and build a system which nurtures discipline towards goals. This works.
- There are still issues with cheating, but it's more of an older way of thinking, that we developed methods to reset.
- Advanced students need to be given more challenging assignments - quantum of assignments should be the same no matter the capability of students. This solution was unworkable until GenAI came about.
Looked from a pure individual skill-building perspective your ideas are alluring, but if one looks at completion rates of any online courses (Udemy/Coursera - under 4%), then one understands why physical cohort-led education system can work.
Happy to chat with anyone who'd like to delve deeper on this.
This would mean moving to 100% weighted exams, and there's good reasons why there has been a general trend away from that over recent decades. For one thing, some students simply perform better under pressure than others, independent of their preparedness and knowledge of the material.
Mind you, I don't really have any alternative suggestions.
> Mind you, I don't really have any alternative suggestions.
This is thing.
If this choice is between:
1. A gameable system that will be gamed by most students.
2. An ungameable system that will unfairly punish those bad under pressure and time constraints.
There isn't really a choice at all.
One option would be a school-provided proctoring system, allowing teachers to outsource the actual test-taking times. It could be done outside of class time, at the student's convenience, and they could have 3-4 hours if they chose.
Fair point, but the solution I propose would only apply to those parts of the assessment involving solo writing assignments -- so excluding class participation, group assignments, etc. (Which is not to say that students can't use AI to cheat on these, but they have other solutions.)
I mean, the real answer is that the other students were cheating on their assignments. It's that simple. We keep making up excuses for all of this shit. Some people don't "test well". Turns out those people don't know shit.
Let's get real here. I know why these nonsensical memes keep propagating but dear god. People will just believe anything these days, including that gas stoves cause asthma or whatever other bullshit is being peddled.
This isn't true. I'm one of those people who tested remarkably well, and back in college would do fine on exams despite frantically copying all of my own (non-comp Sci) assignments. Better than my peers who knew more and helped me cram. Test anxiety is real.
It is definitely not the case that if student A performs better on a timed high-stakes test than student B, that means A must have worked harder / prepared better / know the material better / etc. than B. Some people are very skilled at bullshitting their way through stupid school tests, and others are not. Very few school tests are well enough designed that they can effectively measure the intended target of how well someone understands the topic, content, and course-specific skills which are being intentionally trained in the course.
Bullshitting though tests is a learnable / trainable skill, but schools generally do not teach it very coherently or well and most students do not deliberately practice it. It generally doesn't have that much to do with the content or other skills intentionally taught by any particular course or by schools in general (there's decent overlap with the skills involved in competitive debate and extemporaneous speech, which some students participate in as an extracurricular activity). Rating students on how good they are at bullshitting their way through exams is sadly a significant part of the way our education system is focused and organized, but in my opinion it is not a valuable or particularly valid approach. There are certain professional contexts/tasks where this kind of skill is useful, but developing it per se shouldn't be the focus of the education system.
Sometimes this and related skills are summarized as "intelligence" ("oh she aced the test without studying, she must just be really smart", etc.), but in my opinion it's quite a misleading use of the word.
This exactly how one of my English professors structured his class. The students would have to do the research beforehand and come in on test day with their works cited page completed. The actual paper would be written by hand during class time. You were only allowed the blank green book and a couple of pages of notes with direct quotes to incorporate into your paper.
He wasn’t worried about llms, they were not around, but plagiarism. It worked well.
It's impossible to design a system which is perfect for everyone. People with attention disorders might feel the opposite and will do better with the pressure of a test.
I had to do these for a couple college classes (The original OpenAI GPTs were just released around when I graduated, I remember reading about them and then avoiding pytorch because the wheels were a pain to build.)
You have to get a special blue book with a couple blank pages and then write an essay with the prompt that's given at exam time. Then you turn in the book at the end of the exam. I think it's a great idea and was surprised more classes didn't work that way but I guess it's like you say: grading written assignments like this is a lot of work.
> There are valid reasons why college students in particular might prefer that AI do their writing for them: most students are overcommitted;
Tangentially: I've helped out some college students with mentoring and advice from time to time. One common theme I've noticed is that their class load virtually doesn't matter. They find ways to run out of time no matter how much free time they start with.
We all like to imagine the poor, overburdened college student working 2 jobs and attending classes to make ends meet when reading statements like that. But to be completely honest, the students like that usually have their time management on point. The hardest ones to coach were the students who had no real responsibilities outside of classes, yet who found their free time slipping through their fingers no matter what they did.
Among all of the other problems with easy AI cheating, I wonder how much the availability of these tools will encourage even more procrastination. Feeling like you always have the fallback option of having ChatGPT write the homework for you leaves the door open to procrastinating even longer
> I asked my students to complete a baseline survey registering their agreement with several statements, including “It is unethical to use a calculator in a math class”
Unless there was more to this survey, this wording seems misleading. In a college-level math class, using a calculator is a common expectation depending on the type of class and the problem. The students would probably think of their TI-89, not a magical AI calculator that could solve every freeform problem for them.
I've gone back to university as an adult to study econometrics, already with a software background.
It has been quite an adjustment, the hardest part isn't really the content, there have just been assessment tasks that are more demanding in time than cognition. I think a big part has been these are introductory subjects.
You have much less agency over your schedule outside university itself than you would have when you're working as well, which has been a massive PITA having previously worked as a software engineer for quite a while. As an institution it just doesn't respect peoples time, which is fine when you're much younger and you're just coming out of school (another institution that doesn't care for your time outside of it).
I think my problems have been compounded by the fact this is a undergraduate degree not a masters. There was only a masters for applied Econ in my area (which is much less data and math focused than econometrics), I've had this conversation 100 times, not looking to repeat that, trust me the moment I see an option to change to a masters in my area of interest in the city I live in, I would jump at it. But my choice to go back to uni to some extent is an as much an act of consumption as it is getting a piece of paper to saying "you can trust this man with econ".
Anyways in these first year subjects, you'll have assessment tasks like "make a recording of you demonstrating your understanding of <basic function in excel>, explain the value of <basic function in excel>". They are easy subjects but they are also really time consuming so to some extent it feels degrading. I would take steps to skip the subject based on my prior experience but the focus of these subjects isn't even Excel, its just thrown in there because they anticipate you have little experience with these things and it'll be useful in later subjects when it becomes assumed knowledge.
> They find ways to run out of time no matter how much free time they start with
Despite how much I wish it weren't, this is exactly me. I seem to be able to only work under pressure, so as I got later into my degree I would start my work later and later, until it was standard that I'd start at 4am on the day it was due. A terrible way to function or to learn
the fact is that this is a terrible feedback loop, because every time you started late, but manage to finish on time, it is reinforcing the idea that you're "good" and can manage it even if you start late! This self-reinforcing condition makes you later next time, but because of your capability, you still make it on time.
But at some point, you either stop procrastinating as you find the absolute limit, or fail outright (which is why survivorship bias exists if you look at the student body and see the amount of procrastination).
>They find ways to run out of time no matter how much free time they start with
IMO you’re observing Parkinson’s law, “work expands to fill the allotted time”. The students who take a million classes look like they have time management in check, and I’m sure some do, but they also are benefiting from the inverse of Parkinson’s law — if they can _only_ allot X hours a week to a task, they’re going to make the most of those X hours. This practically holds regardless of the student’s time management skills. I should know, I have horrible time management, and only succeed by overcommitting and rising to the challenge.
I think this is covered in Micheal Easter’s notion that as societies become more comfortable, our brains lower the threshold of what constitutes a “problem” in our lives. We’re wired not only to be great at problem solving but also discovering new ones. Think this is based on prevalence theory related research in psychology.
Different tools in different years as expectations have changed. The TI-89 is incredibly powerful, but has to give way to MATLAB and Wolfram Alpha. It used to be productive to “google” your problems. Going further, there’s now LLMs writing python code to do calculations. Hard to say what’s next, but I’m sure what is considered ethically questionable today will be acceptable and the new thing will be the new questionable tool.
To be clear; there are plenty of contexts where having a TI-89 is 100% unambiguously cheating. There are even more places where MATLAB and Wolfram Alpha are cheating.
A times table test like we all took in grade school. A test applying Newton's method. Any test where the calculator can do the work instead of you, and a TI-89 can do quite a bit of work.
And more trivially/procedurally, the numerous tests I had to take in high school and university which explicitly said "no graphing calculators" (and the occasional test that said "no calculators" outright).
I was thinking of exams rather than day to day learning of basics, seeing tests as teaching rather than assessing. Point being you can't cheat at learning.
The calculator can't fake that you understand e.g. how to differentiate. The working is the bulk of available exam marks but the calculator only gives answers.
All forms/structures in society that think that a TI-89 is cheating deserve to have as many people as possible use TI-89 and similar CAS's.
Ludditism is the human death drive externalized for the modern age. Reject it in all of its incarnations.
There's objectively no value with learning how to perform calculations by hand that CAS's can perform better automatically.
We have well over the number of human beings on earth amounts of calculator capable calculating devices. I will not "be in a situation where I don't have a calculator".
Time spent not teaching folks how to use computational aids of math and teaching them to do it the "hard way" is time robbed from them.
Those who invent systems which subvert or short circuit the attempts to enslave individuals into doing hard work for hard works sake are ontologically good. Those who try to defend work for works sake are ontologically evil and I hope that they reincarnate into a durian fruit in the next life.
> I will not "be in a situation where I don't have a calculator". [...] Those who invent systems which subvert or short circuit the attempts to enslave individuals into doing hard work for hard works sake are ontologically good. Those who try to defend work for works sake are ontologically evil and I hope that they reincarnate into a durian fruit in the next life.
By that logic, you must also permit students another extremely well-established labor-saving technique: Hiring someone else do the homework/exams on their behalf!
After all, for students from certain families, forcing them to use the calculator personally would likewise be "robbing time from them", denying them valuable experience honing their skills in managing and subcontracting. They'll never be in a situation where they don't have a calc^H^H^H^H lackey.
I trust we can agree that (A) permitting outsourcing is absurd, (B) that cheating is not "ontologically good" even when reduces the total human labor, and (C) prohibiting the practice should not put a teacher into a state of stinky-fruit damnation.
I submit that we do care about how students do it and which skills they use, even if we disagree on what those are. It is not as simple as saying that good education means spurring them to supply Valid Answers By Any Contemporary Efficient Means.
You take these classes in order to internalize the intuition for the subject. That's the point of the class. Getting the right answers is only important insofar as it serves that goal.
Your attitude makes sense in engineering but not in a math class for example.
We were not permitted to use a TI-89 in math courses since we were expected to learn the underlying concepts, which is much more than learning how to use a tool.
The reasoning was similar in physics. The instructors couldn't care less whether we used the CAS functionality of the TI-89 because that wasn't a part of their curriculum, but they were concerned about students downloading solver programs. (Most of the instructors would agree that creating our own programs to solve problems would be a valuable learning tool, but they had no way to determine whether we created those programs or someone else.)
Academic courses tend to be biased towards laying the foundations so that we can build upon our own knowledge. There are other types of courses that are purely concerned about applications and the use of tools.
Frankly, I'm quite glad to be able to do basic arithmetic in my head -- having to pick up a calculator every time I need to multiply two numbers would burn through far more time over the course of my life than what I spent in Kindergarten (or whenever) learning to do it myself.
All you want is for people to be better slaves for your capitalist machine. Never learning how to do things themselves: only consuming, buying what could otherwise have been achieved by the simple (but non-GDP-increasing) exercise of their own mind. Your rhetoric betrays your own class.
Try reading more than one leftist book: Fischer is great, but he doesn't try to talk praxis. So I'm talking about pragmatics here: just pretending that "we can abolish work by automating it" is pure ideology, totally ignorant of the centuries of historical examples to the contrary. Any time you get back will immediately be seized. What matters are your personal capabilities, what the working class is capable of doing with their own hands. You advocate waiting on the bourgeoisie to hand over control (what, out of good will?), but Fischer could tell you that only by seizing power for themselves can the working class ever become free.
> Any time you get back will immediately be seized.
I imagine some guy several thousand years ago: "Once we finish teaching this animal to pull the plough for us, descendants of The Tribe Between The Two Rivers shall have lives of pure leisure!"
I feel like you're glossing over the contexts of "acceptable" and "questionable".
The tool itself is not questionable or acceptable, it becomes questionable or acceptable depending on the usage. A pencil and paper can be questionable if the test is designed and expected to be completed without it.
You can design tests where an LLM spitting out python is an expected tool, but what are you testing for then? I doubt there are classes that teach whatever that test would be for yet.
Being one of those students and with a career under my belt of process analysis and coaching, I have an interesting observation: I harness free time as an explicit part of my writing process, rather than something that interferes with it.
I write at about 1200 words per day and I considered each fo the major multi-week assignments in my entry-level English courses to be worth no more than one day of my time apiece. For the finals, I gave them two days apiece, because I wanted an extra day to define the scaffolding for my argument.
My mother indicates that this is how she went through college too; very occasionally, a serious paper would require more effort than this, but for the most part it was “load assignment into brain, study assignment mentally until T-2d, write assignment, submit”. If several essays are due, then they have to be staged at various days numbered T-2d through T-5d for example — and it’s really important to not depend on T-1d existing at all due to courseware/internet/power outages.
I could technically write a worse essay the day it was assigned, but ultimately, I’m turning in A-tier work by this method. The hardest lesson was that I have to try not to wait until T-1d, because there’s a lot of risk encoded in that and it outweighs the value derived from having an extra day to think about it while I do other things.
But it wasn’t about “free time slipping away” — it’s just that I’m writing crap throwaway work that doesn’t matter after it’s done, and so I can barely motivate to care relative to literally anything else in my life that matters. Thus the T-2d compromise: I’m not about to give them precedence over literally anything, but I will concede that I do need to do so one day early, however boring it feels, because I’d rather have a crap day at T-2d than the same crap day at T-1d with the unproductive anxiety of risking a class-retake if my internet drops out.
Notably, when I actually genuinely care about what I’m writing, I’ll spend weeks researching sources and studying arguments and selecting quotes and then assembling it all over a couple days into a work of art — but assembly day is still always as late as possible in the time window assigned, because by then I’m most able to think and write about it efficiently and with a minimum of frustration. Not a zero of frustration, that is — I am a grouchy writer — but I’m healthy-grouchy on T-2d and bitter-grouchy on T-1d, so I do make the effort to put in my writing that day early now.
So: for your coaching efforts, try working with students to construct a working calendar that has non-writing activities in the leadup and then writing activities at the end. ie assuming a 7-day window,
T-7d: Assignment given: Read the assignment. (Seems obvious; is not obvious!)
T-6d: Think about your argument during your free time, while playing games or out at coffee or whatever.
T-5d: Try to construct a very halfass outline on a piece of paper. One sentence per argument you’d like to make, draw arrows to rearrange them. Not complete sentences, not punctuated, doesn’t have any structure at all. Point is that trying will help brain coalesce.
T-4d: Research references for fun. End up with far too many. Start highlighting quotes to yourself using highlighter or digital tools. If you’re going to experiment with a new tool, get it working and productive in 2 hours or discard it and do something shittier.
T-3d: Bind quotes to your argument phrases from that halfass outline. This may force reorg of outline; cool. Compile Works Cited from whatever you end up using so that you don’t have to fuck around with it tomorrow.
T-2d: Write paper, referring to outline / phrases handwritten note. Do one paragraph at a time. Plan to spend your entire day on this with 1 hour away from desk handling bio/sanity needs for every 2 hours at desk. Enforce that upon yourself.
T-1d: Finish whatever writing you didn’t feel like you were prepared to write on T-2d. Ideally try to do this earlier in the day than later, since that every hour you let this slip towards midnight l measurably increases your chances of a life outage causing you to fail the class.
The point of this schedule is to bake in the daydreaming / slow cooker aspect of the creative process but to keep it on the rails. I play video games extensively when I’m thinking about a paper because I can feed my literary brain the assignment to simmer and then go occupy my reflex brain with the game. I usually end up having to use some T-1d time but I’m getting better at managing my life’s dependencies ie. Food and Water and Sleep so that I’m more reliably at T-2d completion :)
> One common theme I've noticed is that their class load virtually doesn't matter. They find ways to run out of time no matter how much free time they start with.
12 credits is normally a minimum. That's roughly 12 hours a week in classroom, give or take. You need 3x that number of hours out of the classroom--that's 36 hours.
So, you're at 48 hours of academic work every single week. A 15 credit load means a 60 hour week.
Most people working jobs would start to complain about burnout at 50+ hours per week for 4 years running. They would almost certainly complain at 60 hours per week.
Most 3 credit hour classes are really 2.5 hours. And almost no one spends 36 hours hours outside of class on a 12 credit schedule unless they messed up and signed up for 4 difficult classes in the same semester. You definitely aren’t spending that much time outside of class all semester long.
You also have to consider that a semester is 16 weeks. The first week or 2 of each semester is very light.
So we’re talking 30/52 weeks a year for most people.
For most people, you’ll never have that much free time again in your life.
College feels like a lot of work because you aren’t good at time management yet. And you remember the last few weeks of each semester where you are actually extremely busy.
> The first week or 2 of each semester is very light.
Please tell that to the undergrad STEM professors, please. Almost all of mine had an assignment first class that was due by the third class.
Freshman engineering is generally Calc I, Physics I, Chem I, and English Composition/Writing and often some random engineering/computer thing. I assure you that schedule sucks even harder that it looks like it sucks, and it gets more time consuming as the years progress.
While there were lots of Party Hardy(tm) types in the College of Arts, the ones I knew who were taking their degrees seriously were working every bit as hard as the STEM folks. Possibly, they were working harder as they needed a lot more extracurricular work and achievement since what they were doing didn't have nice, clean objective measures like STEM does. They spent a lot of time being unpaid labor at functions and networking like crazy.
By contrast, no matter how many hours they worked me at my summer internship, it simply never compared to the grind at school.
Imo those numbers are pretty inflated unless you’re taking a full load of the hardest classes offered. Usually you pair some GE requirements or electives with heavier material. I really don’t want do some sort of humble brag here, so I’ll just say that if I followed your math it would come out to like 90+ hours. I promise, I was not that diligent.
Altman’s analogy didn’t hold up. Calculators were uncontroversial
Calculators are uncontroversial now. But when they first became cheap and widely available, they were not allowed in math classes. Then only four function calculators, then graphing calculators. But still today, programmable calculators are prohibited in many academic contexts.
The point that you're (and everyone is) glossing over here is relative positions on the skill gradient.
A first grader probably would be prohibited from using any kind of calculator on arithmetic tests, 4-fn or not. But 8th graders are usually permitted scientific (non-programmable) calculators.
As you go up in grade level, you "get access to" calculators capable of functionality at the level below you. Because the point is that when we're educating students we want them to actually learn the subject matter, but once we've deemed them to have understood it and we have them move onto the next goal, we give them the tools to make that prior goal easier. We lessen the burden of the little mechanical concepts they already know so that they have an easier time becoming familiar with the next more advanced concepts.
AI systems are so much more advanced than what's capable on a TI programmable calculator. It's hard to draw clean boundaries around the tiers and enforce them by telling the model "help the user with tasks of tier 1-4 but not 5+". That's the issue, that it's really infeasible to strictly use them strictly as learning tools. You can almost do it with a lot of self-discipline and self-reflection to analyze your own workflow, but it's not generalizable across domains.
I mentioned the continued ban on programmable calculators in many academic contexts. Those contexts still include some portions of undergraduate education. This is fifty years after the introduction of programmable calculators.
Realistically, the answers the students gave the teacher were probably motivated by the practical benefits that come with giving teachers the answers they want to hear…bullshit questions are likely to produce bullshit answers. It’s not like first year college students haven’t had twelve years of academic standards moralizing talked at them.
It’s easy to do at the syntactic level by controlling the sampling. For example, it’s easy and common to restrict output to be valid JSON but just not allowing any tokens that would make it not valid JSON.
But reliably restricting output at the semantic level is very much an open problem.
>>> imagine how radically math class must have changed when calculators became widely affordable
It didn't.
I was in math class when calculators were introduced. At least for high school level and beyond, the curricula were designed to make problems solvable without calculators, and they weren't of much use. This was still the case when I taught an undergrad college math class in 1997. Graphing calculators were allowed, and the kids who tried to use them just screwed themselves up.
I would have gladly changed the curriculum to use calculators and computers from the very beginning. As tools, and not just to administer the same old exercises and quizzes. Give them Jupyter Notebook. Math education has never been a success story.
Education faces a dilemma, which is that it has always used heuristics to guide study and assess performance. Exercises such as the "three paragraph essay" had no use in the real world, even long before AI could generate them on demand. When one of those heuristics is broken, another one has to be found. Even word processing forced teachers to grade papers on content, rather than mechanics.
> Exercises such as the "three paragraph essay" had no use in the real world
The rigid formatting imposed by graded assignments may not have any use. But on the other hand, having completed a liberal arts/"soft" science degree before my CS degree I've greatly benefited in the workplace from the writing experience it provided. I had to write so many papers that it became more or less effortless to produce long form, well structured writing and those written communication skills have helped me distinguish myself far more than my technical ability at work.
The generic ChatGPT overly formal corporate tone has no nuance or subtlety and is a poor substitute for well crafted, deliberate communication. I am always conscious of how my exact words and phrasing would be perceived by the intended audience, frequently requiring a balancing act between competing interests while maintaining clarity. Due to that I manage to avoid stepping on toes or sabotaging relationships due to inartful phrasing. It's frustrating to receive emails consisting of LLM boilerplate because it has such low information density and is so much more difficult to infer tone and emotion from the other side .
I'm very grateful I completed my education without the temptation to just churn out low effort writing or code and depriving myself of that experience. I'm not confident at all I would have been able to maintain that self-discipline.
I'm with you there. Though I have a technical degree, I went to a liberal arts college, and preferentially took courses that had a heavy writing component. I was a good typist thanks to coding, and one of the first students to use word processing. Today, I still use those skills to my advantage.
As a teen in the late 80's I had an HP calculator that I programmed to compute molecular weights given an input string like "H2SO4". It felt like having a secret superpower, especially when I participated in competitive exams. I was a very straightlaced kid and would not have used the program if it such things were explicitly forbidden, but as far as I could tell, they never were.
Reminds me of when I write a j2me app for matrix diagonalization because we could use the old feature phones as calculators. Nobody thought we'd be mad enough to use those to cheat...
Do you still have the source code for the j2me app?
I hadn’t yet learned to program back when I was still using a feature phone, but I have a lot of fond memories of J2ME applications that I installed on my phones. Mostly games, of course.
I encourage anyone that wrote J2ME games and utilities, no matter how small or big, to upload the source to GitHub :)
But normally it depends on the subject and if the automation/machine solves the primary skill being teached or if its just a "secondary/tertiary" skill. Are you in a Calculus 101 class? Calculators like TI-89 are likely to be prohibited when examining for deriving analytical solutions for derivatives and integrals.
Statistics, Physics or any other subjects that needs applied maths? Such a calculator is probably a minimum requirement to take the course.
My HP-49g+ was definitely load-bearing going through EE. I was never much good at memorizing big sheets of formulas but I was pretty good at memorizing a couple of simple differential equations (e.g. I(t) = C dv(t)/dt was easy, v(t) = v_s * e^(-t/RC) wouldn't stick). So I'd just... derive all of the "special case" formulas from scratch during the exam. Usually they were simple enough that I could just get them into the right form but I'd lean on my calculator doing the symbolic integration for me when they weren't.
The other thing it was awesome for was solving systems of linear equations. I could do the nodal or loop analysis just fine, I'd write down the matrix that represented the system of equations and then just punch that matrix in and invert it.
I was on pretty good terms with my chemistry teacher, so...maybe? It's been a while, but I don't remember either showing it off or taking pains to keep it secret. To adults, that is; my nerdy friends and I delighted in showing off the cool stuff we did with our calculators.
I vaguely remember thinking that one likely reason shortcuts like mine were not prohibited was because no one in charge suspected that such things were even possible with current technology, or if they were, that a child would be able to exploit it. But as long as I kept to the letter of the rules, I considered myself ethically in the clear.
> But as long as I kept to the letter of the rules, I considered myself ethically in the clear.
Yeah, totally, just to be clear I'm not judging.
In fact, if you programmed it to handle those operations, one could argue you had already learned a big chunk of what was going to be measured in the exams.
Kind of similar to the paradox of creating and using cheat sheets, is highly likely you're accidentally learning about the subject matter in the process of writing the sheet, sometimes up to a point where the cheat sheet is not necessary anymore.
> Kind of similar to the paradox of creating and using cheat sheets, is highly likely you're accidentally learning about the subject matter in the process of writing the sheet, sometimes up to a point where the cheat sheet is not necessary anymore.
The problem is (example from mathematics): even if you are capable of deriving some formula (you thus understood the topic well), it takes a lot of time in the exam. Looking at the cheat sheet is much faster - in particular when the time is somewhat precious in the exam.
A related personal story: During my statistics course in high school, we discovered that the TI-89 had some statistical functions that the TI-83 didn't have. So, the rule was that if we wanted to use the TI-89 ones, you had to write an application for the TI-83 one. It was a great way to really learn the algorithms.
I could do it correctly from the get-go. The program just saved me from drudgery many times over. Probably enough times to recoup my time investment to create the program, but in any case I enjoyed coding for its own sake.
I always felt (and my maths teachers agreed) that if I understood something sufficiently to automate it, I’d proved my point and didn’t need to do the rest of the exercises.
Edit: Automate in the sense of coding it myself, not in the sense of downloading some software.
The concepts of adding machines and calculators were also slowly phased in over the span of a century. The first commercially successfully adding machines hit the market in the 1890's, and pocket calculators took off in the 1980's. AI went from theory to answering hand written math homework questions from a photograph in a few years.
I only had a calculator (at a technical university) starting in the mid-1970s. Prices were dropping like a stone in about that period. In high school it was pretty much slide rules.
Totally correct. In the 90's as a kid in school using a calculator was highly debated amongst teachers and the ability to bring one out on your desk depended on the teacher.
In grade 2 i had a teacher who would say "I don't believe in erasers", you know, the things that "undo" pencil. As a ~6 yr old i actually didn't understand this phrase: "Well I have one, they're real!"
It's also an extremely misleading comparison. Basic calculator functions do not in the slightest replace anything taught in a maths class. Using ChatGPT not just to write entire paragraphs (replacing composition), or even providing the writer with ideas (replacing the creative aspects of writing) isn't comparable to adding two large numbers together.
The equivalent in maths would be if you handed students a theorem prover or have Wolfram Alpha give you step-by-step solutions and obviously nobody to this day allows this, because like ChatGPT for writing it'd defeat the point, that students think.
When I was in uni we were allowed basic but not programmable calculators during exams and a lot of CS classes even were pen&paper, if the prof was a bit hardcore
Turns out education done right is vaguely a speed-run of how the knowledge was developed. Adding calculating tools makes sense as you advance the the corresponding point in the process. Honestly, I think there should be a chunk of precal and calc where they use slide rules only, then calculators of increasing complexity (or just increasingly complex features of one calculator).
"When will I use this in real life" is a declaration that you have no expectations of learning the next lesson that builds upon this one.
>> Altman’s analogy didn’t hold up. Calculators were uncontroversial
> Calculators are uncontroversial now. But when they first became cheap and widely available, they were not allowed in math classes.
The author of TFA means specifically for his cohort of students, not in general. He polled his students, and the result was that they thought calculators weren't seen as unethical but they were more skeptical/uncertain about AI. By his current students, now, not in general.
LLM’s have been widely available for approximately five years.
Five years into the availability of a calculator with an arbitrary advanced feature, it was controversial in academic contexts. Some of the author’s students could be grand-children of students from the early days of consumer calculators.
The author is comparing a new technology with an old one. And ignoring programable calculators which are still sometimes banned after fifty years…and many of the author’s students probably have used LLM’s for homework despite their statements that please the author.
Among other things, he's analyzing his students attitudes' re: AI and cheating with AI, and also comparing what they claim to feel vs what they actually do! It's mostly a reflection of what his students feel about the use of AI in English writing, not about calculators vs AI.
It seems as if you're responding to the line you quoted (out of context) through your preconceptions of what the article is about instead of actually reading the article!
> and many of the author’s students probably have used LLM’s for homework despite their statements that please the author.
"Probably have used LLM's"!? Don't take this the wrong way, but it seems you didn't read past the line you quoted, am I right? Because this is explicitly addressed multiple times, and the answer may surprise you.
Yeah I don't believe you, especially because you have an axe to grind about AI singularity bullshit. No one in their right mind should allow a graphing calculator to be used on an algebra exam, might as well let them bring a laptop and open Wolfram alpha.
Maybe my teachers didn’t know you could program equation solvers into it. Either way it helped me understand that those in authority didn’t understand the nature of the world changing beneath them, which prepared me for how many people would end up misunderstanding AI and under-appreciating it despite how much it is evidently changing the world.
> imagine how radically math class must have changed when calculators became widely affordable
I was there. We had been given slide rules, and a decent chunk of my 11-yr-old maths classes were devoted to teaching us how to use them. Calculators were banned because they "didn't teach us anything" (but somehow slide rules did? It didn't make any sense at the time either).
Over the course of the next few years, calculators became more acceptable, and by my 18-yr-old Maths A-level class, we were being advised on which scientific calculators to buy.
Really enjoyed that. Shows the "messy middle" that people are caught in during this current wave of AI tech. I think one undeniable positive outcome has been the like collective introspection AI has sparked among so many people. With a tool challenging what it means to be "human" or "creative" (more so than prior technological advances), I've been seeing a lot of wonderful discussions, articles, videos, etc. with people wrestling with those questions and also just affirming their own singular voice and unique creative essence. It's been cool to see.
I think one undeniable positive outcome has been the like collective introspection AI has sparked among so many people.
Has it sparked that? I'm doubtful but I'd be very interested if anyone has a references about such introspection.
I grew up reading Douglas Hofstadter back in the 70s and what I appreciated with his ideas was using AI to illuminate what is human. His wave of AI failed, of course. But still, it's disheartening how little of that kind of inquiry seems missing from the current wave of AI.
This essay resonated with me because it highlighted the similarity between AI-written texts, describing the result as a word salad. And this also reminded me about some words from my teacher of Russian Literature: that the "bright future" themed novels of the pre-WW2 Soviet writers — works produced under strict political control — read like one big novel without a beginning and an end, and not as separate works.
And this grayness and sameness is what happens when people are forced to "think" as a chorus, either by the authorities or their censorship, or voluntarily by using the same AI's help.
> most students are overcommitted; college is expensive, so they need good grades for a good return on their investment; and AI is everywhere, including the post-college workforce.
Yeah. Overcommitted to partying and skipping class.
Has this author ever been to an average American university?
Let's rephrase it to "most of even the best students" then.
I went to about the cheapest US school that had a decent math program. It costs $17,500/yr between tuition, rent, books, and rice and beans.
That's a lot of money. It's over $10/hr in pre-tax income, even if you work full-time all year, which isn't an easy bar to clear in the sort of towns with cheaper universities. Wages don't scale well enough with more expensive tuition for there to be substantially better options.
Classes are another 22+ hours each week (you could complete school in 3-4 yrs instead, but that makes it even harder to afford and doesn't really reduce the workload enough to make a difference, however I'll also factor in a 15-hour workload later).
The rule of thumb is that you should study 3 hours for every hour of class. I found that approximately correct. Some classes took a lot less. Some took a little more. Combined with the self study you need in adjacent topics, 3hrs is a fair bit low.
During the school year then, you have something that looks like a 128hr/week schedule, or 100hr if you're finishing in 4yrs, and still 60hr/week even if you're finishing in 4yrs and racking up $70k in debt.
Don't get me wrong; I had free time (I worked more during the summer, less during the school year, allowing loans to cover the slack, which bought extra time here and there), but it wasn't exactly a party either. When I skipped class it was because I had to work, had to study for some other more pressing class, or found it more efficient to study the book than to try to understand that particular lecturer.
This matched my experience 30 years ago. Work 20 hours a week, but tuition and living expenses were a lot cheaper back then ($215-$330/month for a room! $900/quarter tuition). The 3 hour for every 1 hour of class is especially true for computer science, and skipping class in favor of self study worked well if the lecturer was really bad. Lectures were pretty much bonus reinforcement if useful at all, a lot of what you learned relied on self study.
A lot of students didn't do what I did, and they washed out pretty quickly (I had a lot of classmates from HS that didn't last the first quarter). My first quarter was pretty harsh (only got one 4.0, and a 2.7 in a chemistry class I had no reason being in), but I wised up quickly. It was hard going from High school where I could do all my homework in the time between classes, to college where I had to do real actual studying.
My experience was only 7 years ago. I don't think it's a generational difference so much as the fact that a university is a big place, enough so that even if you engage actively with 100+ people you won't see the whole picture.
It has gotten a lot more expensive and competitive. I’m almost embarrassed to mention how crazy cheap my school was in 1995 compared to 2025. Also, I doubt I would have gotten in with my high school achievements even though I graduated in a fairly decent cum laude position. Life for kids these days is a lot tougher.
The numbers I quoted were current prices for the school/city I attended (UND, if anyone wants to cross check). When I attended prices were lower, but so were wages and a number of other things, roughly proportionally AFAICT.
I could maybe see your point about admittance (I had something like a D average, maybe C- or C+ or something, in high school), but I think my financial estimates were about correct. Is something majorly incorrect?
Even if so, there are a dozen cheaper states in the country with halfway decent programs (if any future internet denizen is reading this, Fayetteville Arkansas is actually great for math right now, both in quality and in cost/jobs). I doubt my observations are too far off-base for a typical student trying to go to school economically.
> The rule of thumb is that you should study 3 hours for every hour of class. I found that approximately correct. Some classes took a lot less. Some took a little more.
lol stop the cap.
Its more like 0.5 hours of grand total work for every hour of scheduled class, since most kids are skipping 50% of the classes and using "chat" (a really fetch name for a new digital drug btw) to make up the rest.
What people are telling you in the comments is that your perspective is not universal. I've personally only ever skipped one class in my time in undergrad (as an American at an American university), and not for a party. I'm not a special case or anything, those classes are very expensive!
Isn't it basically a remedial class, or rather a class for people who did the minimum of HS math? I would expect it to be filled with more partiers than studiers, and also that plenty of the people in it would already know a lot of the material that's covered (so why not ditch?)
In 2012 I took my first course in undergrad. The teacher asked everyone to introduce themselves and say something they liked to do. Every single student, except myself, said they like to party. That was shocking to me. Most students weren’t skipping classes though.
Their writing skills were also abysmal. Frankly if they were that bad by college they didn’t seem likely to improve in my opinion.
I've been taking one class at a time for the last six years at what I think is an average American university. The twenty somethings that your comment is aimed at have been my lab partners and such. You're describing maybe 10% of them. As a group they're all over the map.
As learning goes, I'd say anybody taking more than two classes at a time is overcommitted. They might manage to get A's but I speak with these students about the courses we took together a year or two after the fact and it's clear to me that taking four or five classes at once is an awful strategy for retention.
> several statements, including “It is unethical to use a calculator in a math class”
That's too broad to be of any use, if the math class is teaching you to calculate in your head, then using any calculator is cheating. If the math class is teaching some algebra equation solving skills, then using a programmable calculator that auto solves them is cheating.
That's the similar issue with such experiments - they unfortunately aren't rigorous to provide any insight into education
I understand why much of the discussion about AI and university education has focused on first-year writing classes in the U.S. Some of my own first experiments with ChatGPT in December 2022 were having it write school-like essays [1].
Over the past few months, I cotaught a university class in which we also had first-year students use and reflect on their use of AI in their classwork. But the context was different: the class was a seminar on science communication (how to make science engaging to children and the general public), and most of the time was spent with the students doing group projects. Also, the class was at a university in Tokyo, and we taught it in Japanese.
We have just started analyzing the feedback from the students, but my impression is that they were less conflicted about the use of AI for their group work than they might have been if they had been doing their projects as individuals.
Meanwhile, as the semester progressed, agentic frameworks started to mature. I spent a lot of time on my own experimenting with Claude Code and Gemini CLI. While none of the students in that class seemed to use them, it became clear to me that such higher-level cognitive tools will pose an even greater challenge to higher education than essay-writing chatbots do now.
AIs grading AI-generated essays looks like a recipe for model collapse. That's why we certainly need people who go after the "diminishing returns" of improving their writing skills beyond the "good enough" that AI delivers.
Should education systems aim for that for all of their students? Certainly, because AI alone is not sufficient to raise the bar. As impressive an AI is when it seems to invent a new molecule, it is still only possible because of the original works of many people.
What a lovely essay. Reminds me of the way I loved the liberal arts growing up. I missed having classes like that in college (AP'd and ACT'd my way out of most requirements).
English teachers seem especially prone to that friendly and sporting demeanor the author has. Professors from the engineering schools are far more prescriptive, probably due to the nature of the material.
Since discovering Analytical Philosophy, I think it is irresponsible to combine Nonfiction and fiction under the term 'English'.
As an engineer, I write emails, they need to be clear, factual, etc... This is in huge contrast with fiction, where writers get merit for being intentionally ambiguous with things like metaphors and symbolism.
What an incredible disservice to students and society to consider English(nonfiction) an 'art'. It should be treated like math and science.
I had to become a middle aged adult and learn this for myself.
All the best writers that I know in the sense that you mean (communicating information precisely), including non-native speakers, are also avid fiction readers. Many also write fiction or prose for fun. Familiarity and fluency with the details of usage and vocabulary are what let one employ these things precisely for whatever purpose, fictional or not.
That is not my experience at all. To the contrary, those who write fiction read mostly fiction, and those who write non-fiction read mostly non-fiction.
They're incredibly different skill sets. One is all about argumentation, convincing, facts, and citations. The other is all about imagination, beauty, evocation, flavor.
Obviously they both require assembling nouns and verbs and other parts of speech in sentences, but they seem to be virtually entirely different capabilities at the end of the day.
Writing excellent short stories doesn't really help with crafting effective business communications, design documents, etc. And vice-versa. In fact, I think they can sometimes even be harmful -- the kind of clarity required for non-fiction can constrain imagination in fiction, while the creativity celebrated in fiction can be quite counterproductive when it comes to functional communication -- what is intended to be clever or unique often gets misunderstood.
To back up your point, I kind of hated English class until my senior year of high school when I took AP English Language (nonfiction), after which I started drinking books from a firehose.
> I had to become a middle aged adult and learn this for myself.
This is a cliche.
You can’t write precisely without an understanding of how language becomes imprecise, of its fundamental instability. Precision and delicate use is an accident when it does happen, and its happening can never be proven. We must have faith in the accident.
I disagree strongly with this, it's like saying you must understand the subtleties of calligraphy or typography in order to be proficient at writing in your notebook with a pencil. I have no doubt you will be more purposeful and deft with your handwriting having this knowledge but they're two completely different skills.
You can be taught to and be proficient in "writing with your pencil" by learning the rules [1]. Efficient, practical, immediately useful and applicable. No subtlety required nor desired. It's the same as all practical skills or trades.
I agree that it's a good distinction to make. Personally I haven't thought about it till I read On Writing Well by William Zinsser. In the book he specifically teaches writing nonfiction and even shares an anecdote where he was a guest on a radio show promoting a writing conference and was annoyed with the host because he conflated writing with literary works.
So yeah, I recommend the book to people interested in writing.
More traditionally you'd study "rhetoric", the art of making your arguments appealing. It doesn't really matter whether the things you say are true or false.
Rhetoric is valuable in any writing endeavor; clarity is only valuable sometimes.
For a funny take on the whole "rhetoric" is the use and abuse of logic some people might enjoy How to Win Every Argument by Madsen Pirie which also happens to be where I plucked the tagline regarding rhetoric from. It's a pretty easy book to go through in toilet break sized increments, the author goes through different fallacies and how they're employed one by one along with various rhetorical devices.
Though a few years ago when I searched for a book on rhetoric and making convincing arguments Office Of Assertion by Scott Crider also popped up, but it's aimed more at written rhetoric instead of what most people have in mind.
> For a funny take on the whole "rhetoric" is the use and abuse of logic
But it's much broader than that. You can make true arguments. You can make confusing arguments. And you can use tools that have nothing to do with logic at all. Rhetoric has a lot to say about rhythm, alliteration, and linguistic structure. And a lot more to say about your personal bearing and your tone of voice.
Hand written essays in class. Short and longer form discussions, questions and answers in class. End of unit, term, semester, year. Interview on topic in spoken form.
That doesn't mesh with the $5000 per credit hour ,get students in and out and degree'd ASAP. No 100 level outside of Marlboro college are you going to find a graduate sized class cohort meeting with professors 1 on 1 and debating philosophy and doing stand up rhetorical studies.
Here are some of the essay topics I had them read aloud:
[...]
I expected them to laugh, but they sat in silence. When they did finally speak, I am happy to say that it bothered them. They didn’t like hearing how their AI-generated submissions, in which they’d clearly felt some personal stake, amounted to a big bowl of bland, flavorless word salad.
But that's what it has been trained on - almost all academic writing is bland flavorless word salad, and this is extremely noticeable in title fads. I have a nearly decade-long game running with my friend where me make up absolutely bullshit concepts that could nevertheless be plausibly published in a journal, and the process has been going on long before that.
'Verbing the noun: towards a genericization theory of expressivity in high-entropy counter-heterogeneity' describes an ongoing problem in academic writing where novelty is deprioritized in favor of acceptability by an evermore tightly circumscribed set of peer professionals whose socioeconomic interests favor the establishment of intellectual stasis that maximally conserves positionality in a quais-Simmelian network space parameterized by income, tenure proximity, and citation count.
Or put more clearly, the more academics write to impress each other instead of to reach the public, the more generic their titles and language will be. Being able to parse and regurgitate wordy titles and abstracts constitutes table stakes in academia, so the incentives tilt toward burying the lede any original proposals as deeply as possible so as the minimize the career-damaging possibility of rejection on technical/syntactical grounds.
It occurs to me that writing by those in STEM fields and those in the humanities is entirely different and each group dislikes the other’s writing. When I was in college, my professors in technical classes had no problems with my writing. After I graduated and I wrote some technical articles, my writing received praise from readers. However, when I took two semesters of mandatory English writing classes in college, my professors hated my writing and nothing I did made them happy with it.
When LLMs became widely usable, I was one of the people who really liked much of the writing that they did. I found it was relatively close to my writing style, which I consider to be good, despite the disagreement from those in the humanities. It was close enough to my own writing that I have even had people on Discord accuse me of using LLMs to write my messages for me, when I had not.
The linked article was clearly written by an English teacher. He criticizes AI-generated texts as “a big bowl of bland, flavorless word salad“. Now, there are many cases where LLMs output nonsense, but in cases where the writing logically flows, does not self-contradict in any way and avoids unnecessary repetition, “bland” and “flavorless” are good. The goal of writing is to convey information across space-time; writing that is “bland” and “flavorless” is the best way of conveying information.
I can see a number of things he did in his writing to avoid being “bland” and “flavorless”, and I consider them to be examples of poor writing:
He used dozens of idioms that make the text difficult for non-native speakers and unpleasant for native readers. He used a number of colloquialisms, including some that are inappropriate in professional contexts (although I will not repeat them since I refuse to write them). He used a word whose only definition is provided by Urban Dictionary and therefore is not even an official word:
He also brought politics into an apolitical topic. The injection of politics is a great way to derail any form of productive dialogue and should be avoided.
He used a story format of the kind that has infected journalism. It is very rare that the process by which something was learned is useful to readers and presenting it for dramatics wastes their time. That is with the exception of stories on the topic of security, where hearing the process is often genuinely informative. My aversion to this writing style is so severe that I have a standing policy to stop reading a news article the moment that I see that it uses this style for a topic that is not security-related. After reading his article, I will extend my policy to apply to essays by academics in the humanities too.
He made numerous attempts to evoke emotional responses to elicit agreement, rather than to make clear arguments based on facts. This is great for propaganda, but not so great for making points. Every one of these appeals to emotions is poor writing.
Beyond those things, he also did not properly cite sources in multiple places. To name a few, the quotes from Sam Altman and Annalee Newitz are uncited. As an academic, he should know better.
Some of these things might actually have places in certain types of writing. They certainly have places in propaganda. They also have places in fictional literature. However, they do not have places in attempts to argue a point.
I imagine if he corrected all of my criticisms, he would find the result to be “bland” and “flavorless”. That is how an attempt to argue a point should be.
A lot of the purposes in education for which the use of AI would be considered "cheating" involve writing assignments of one sort or another, so I don't know why most of these education scenarios don't simply redirect the incentive.
For example, in an English class with a lot of essay-writing assignments, the assignments could simply be worth 0% of the final mark. There would still be deadlines as usual, and they would be marked as usual, but the students would be free to do them or not as they pleased. The catch would be that the *proctored, for-credit* exams would demand that they write similar essays, which would then be graded based on the knowledge/skills the students would have been expected to gain if they'd done the assignments.
Advantages:
- No more issues with cheating.
- Students get to manage (or learn to manage) their own time and priorities, as is expected of adults, without being whipped as much with the cane of class grades.
- The advanced students who can already write clearly, concisely and convincingly (or whatever the objectives are of the writing exercises) don't have to waste time with unneeded assignments.
- If students skip the assignments, learn to write on their own time using ChatGPT and friends, and can demonstrate their skills in exam conditions, then it's a win-win.
This all requires that whoever is in charge of the class have clear and testable learning goals in mind -- which, alas, they all-too-often do not.
A lot of students, even at the college level, don't think that far ahead and make bad decisions because of short term thinking.
Look at any list of advice for new college students and almost every one of them includes "go to class". Simply attending class is way easier than homework and yet, when there's no short term consequences for not doing it, plenty of students will just not do it.
Cheating is another great example. Cheating in college is rampant because kids don't want to do the work they're assigned. I don't understand the logic behind the idea that if you tell all the kids currently using ChatGPT to write their essays, "Hey, you don't actually have to write that essay at all" that you think they will somehow choose to write it anyway. They're already choosing to ignore the long term benefits of homework even when there are short term consequences, so I don't see how removing those short term consequences will make things better.
If you tell kids there are no immediate consequences for not doing homework, many of them just won't do it and they will fail because they haven't learned anything.
Maybe you're okay with that. Honestly, I'm not actually trying to convince you that it's a bad idea. I just think if your proposal is based on the idea that kids will choose to something boring that they don't have to do in the short term because it benefits them in the long term you're overestimating a lot of kids (and adults for that matter).
There's a section in Zen And The Art Of Motorcycle Maintenance (1974) where Persig takes this all the way to the final conclusion that there should be no grades at University, and no degree at the end, and then and only then will everyone who goes there actually be learning-motivated.
So. I teach at a university and I do give an "assignment" exactly like this.
In a few of my classes, I have final projects that teams work on. I also have presentations. I used to require them of all students; and quickly learned this is a good way to waste valuable time.
Now, all my presentations are completely optional for NO CREDIT. You don't get penalized if you don't do them, and perhaps more importantly, I give ZERO EXTRA CREDIT for doing them.
As you can imagine, every single presentation I've gotten from this has been absolutely worth it.
The problem is that the motivation from above (i.e., administration, state legislatures, employers, etc.) is no longer really about learning. We could have an entirely learning-motivated university right now and it would be considered a bad thing by many powerful people because it's not aimed at "preparing people for the workforce" (in part simply by providing that degree).
> I don't understand the logic behind the idea that if you tell all the kids currently using ChatGPT to write their essays, "Hey, you don't actually have to write that essay at all" that you think they will somehow choose to write it anyway.
I unironically believe if you tell all the kids they don't have to write the essay at all, much more will choose to write it.
Kids cheat not just because they're lazy. Cheating makes people feel smart. The fact you can get credits by doing very little while others work their asses off is rewarding and self-validating.
The big issue of exam-only approach is that a one-hour exam is not enough to evaluate a student's performance, unless your educational goal is just to make students memorize stuff by rote. I'd consider a 3-hour open-book exam bare minimal. But if every class does that it'll be too exhausting.
Do Universities no longer do that? All of my finals were 3hrs. There was a special schedule during finals week with 3 slots per day. The time of your final exam was based on when the first lecture session of a class took place. Really sucked to get an 8-11 AM slot when your classes never started before 11.
Fun prank: set all of the clocks in your dorm neighbor’s room to different wrong times. Guy across the hall knew we were messing with him, trusted his watch - which had the correct time, but wrong alarm time. Realized he had a problem when he had hot water in the shower and no one was around. He was only 45 min late to the exam. Good times.
Different groups have different standards of course, but that prank seems pretty cruel.
This, in general, seems like a great thing. The goal of a university should be to produce premium students, and nothing's better than a trial by fire.
We actually had this exact thing at my university. One sophomore level weed out class was a "self paced" electrical engineering class. It was called self paced because you were given a textbook and were free to work through it at your own pace. But to finish the class by the end of the semester you had to average 2 chapters completed per week, and completing a chapter not only included finishing a problem set and taking a test which you had to score 90%+ on (and were required to finish another problem set and retake it otherwise), but on occasion also demoing some skill in the lab.
It was brutal, but one of the most educational classes I've ever taken - and obviously not just because of what I learned about electrical engineering. Of course it seems modern universities have just become profit-driven degree treadmills. Weeding out students? That's reducing profit! And yeah looking back at my uni's page it seems this class is no longer self paced. Lol. And that's at a top 10 school. The enshittification of education.
Part of the issue is with the purpose as you describe it. Sure, at top 10 schools, a trial by fire would result in much needed “growing up” as the gifted but undisciplined (speaking for myself and many users of this site) students find their way to more durable motivations. But at the vast majority of schools, a trial by fire would end with a lot of students burned.
Perhaps that begs the question, if those kids can’t handle self-directed education, why are we putting them there in the first place, but that’s definitely a grey area, and there are hundreds of thousands of students who are smart enough to do well in higher education and skilled work, but weren’t disciplined enough to handle what you’re describing as freshmen.
The idea of "weeding out" students implies that many students are "weeds" who need to be uprooted and thrown away rather than grown.
A teacher who thinks this way is probably in the wrong profession. A university that operates this way is failing to educate the students it admitted.
Weeding out as I've seen it is a class that requires a certain level of commitment and ability to either plan your work or tough it out that a high school just can't really prepare anyone for. So in a way the student isn't a "weed" but their motivation or maturity might be and they're free to retake the class once they know that university will require them to put in more work than high school. If they can't put in the work then completing a thesis and graduating is going to be very hard and that happens the last year of uni so better to set the expectations early with a "weed out" class.
On the class topic, I suspect that attendance was more impactful for students pre-internet as the alternative was to wade through the library piecing together material.
With lecture notes/slides available online, well prepared books and study forums readily available - in-person attendance can feel archaic.
We may be experiencing a similar dynamic in education with AI. In a world where we can create individualized curriculum’s for each student encompassing the entire tree of knowledge - Perhaps it’s time to rethink how we educate students rather than push them into lecture halls designed for the Middle Ages.
Here's an alternative hypothesis...
People thrive under regularity, and young people (especially) tend not to understand that. Similarly, being able to focus on a single thing is a kind of super-power, while multi-tasking generally hurts performance on tasks.
Going to class (and paying attention) means that you've got a regular period of focus on the class topic. That combination of regularity and focus translates into long-term learning and better performance.
> the assignments could simply be worth 0% [..] that the proctored, for-credit exams would demand that they write similar essays.
We run university programs at my company, and arrived at this bit of insight as well. That said, some of your points are incorrect or incomplete:
- You can't build systems assuming responsible individuals. These systems are guaranteed to fail. Instead, assume individuals are mould-able, and build a system which nurtures discipline towards goals. This works. - There are still issues with cheating, but it's more of an older way of thinking, that we developed methods to reset. - Advanced students need to be given more challenging assignments - quantum of assignments should be the same no matter the capability of students. This solution was unworkable until GenAI came about.
Looked from a pure individual skill-building perspective your ideas are alluring, but if one looks at completion rates of any online courses (Udemy/Coursera - under 4%), then one understands why physical cohort-led education system can work.
Happy to chat with anyone who'd like to delve deeper on this.
This would mean moving to 100% weighted exams, and there's good reasons why there has been a general trend away from that over recent decades. For one thing, some students simply perform better under pressure than others, independent of their preparedness and knowledge of the material.
Mind you, I don't really have any alternative suggestions.
> Mind you, I don't really have any alternative suggestions.
This is thing.
If this choice is between:
1. A gameable system that will be gamed by most students.
2. An ungameable system that will unfairly punish those bad under pressure and time constraints.
There isn't really a choice at all.
One option would be a school-provided proctoring system, allowing teachers to outsource the actual test-taking times. It could be done outside of class time, at the student's convenience, and they could have 3-4 hours if they chose.
> An ungameable system that will unfairly punish those bad under pressure and time constraints.
Given modern communication technology it’s still gameable
Fair point, but the solution I propose would only apply to those parts of the assessment involving solo writing assignments -- so excluding class participation, group assignments, etc. (Which is not to say that students can't use AI to cheat on these, but they have other solutions.)
> For one thing, some students simply perform better under pressure than others
Learning to perform under pressure is the main purpose of attending college.
I mean, the real answer is that the other students were cheating on their assignments. It's that simple. We keep making up excuses for all of this shit. Some people don't "test well". Turns out those people don't know shit.
Let's get real here. I know why these nonsensical memes keep propagating but dear god. People will just believe anything these days, including that gas stoves cause asthma or whatever other bullshit is being peddled.
This isn't true. I'm one of those people who tested remarkably well, and back in college would do fine on exams despite frantically copying all of my own (non-comp Sci) assignments. Better than my peers who knew more and helped me cram. Test anxiety is real.
It is definitely not the case that if student A performs better on a timed high-stakes test than student B, that means A must have worked harder / prepared better / know the material better / etc. than B. Some people are very skilled at bullshitting their way through stupid school tests, and others are not. Very few school tests are well enough designed that they can effectively measure the intended target of how well someone understands the topic, content, and course-specific skills which are being intentionally trained in the course.
Bullshitting though tests is a learnable / trainable skill, but schools generally do not teach it very coherently or well and most students do not deliberately practice it. It generally doesn't have that much to do with the content or other skills intentionally taught by any particular course or by schools in general (there's decent overlap with the skills involved in competitive debate and extemporaneous speech, which some students participate in as an extracurricular activity). Rating students on how good they are at bullshitting their way through exams is sadly a significant part of the way our education system is focused and organized, but in my opinion it is not a valuable or particularly valid approach. There are certain professional contexts/tasks where this kind of skill is useful, but developing it per se shouldn't be the focus of the education system.
Sometimes this and related skills are summarized as "intelligence" ("oh she aced the test without studying, she must just be really smart", etc.), but in my opinion it's quite a misleading use of the word.
This exactly how one of my English professors structured his class. The students would have to do the research beforehand and come in on test day with their works cited page completed. The actual paper would be written by hand during class time. You were only allowed the blank green book and a couple of pages of notes with direct quotes to incorporate into your paper.
He wasn’t worried about llms, they were not around, but plagiarism. It worked well.
So one bad day can ruin your marks.
It’s also a disadvantage for people with test anxiety.
It's impossible to design a system which is perfect for everyone. People with attention disorders might feel the opposite and will do better with the pressure of a test.
I had to do these for a couple college classes (The original OpenAI GPTs were just released around when I graduated, I remember reading about them and then avoiding pytorch because the wheels were a pain to build.)
You have to get a special blue book with a couple blank pages and then write an essay with the prompt that's given at exam time. Then you turn in the book at the end of the exam. I think it's a great idea and was surprised more classes didn't work that way but I guess it's like you say: grading written assignments like this is a lot of work.
> There are valid reasons why college students in particular might prefer that AI do their writing for them: most students are overcommitted;
Tangentially: I've helped out some college students with mentoring and advice from time to time. One common theme I've noticed is that their class load virtually doesn't matter. They find ways to run out of time no matter how much free time they start with.
We all like to imagine the poor, overburdened college student working 2 jobs and attending classes to make ends meet when reading statements like that. But to be completely honest, the students like that usually have their time management on point. The hardest ones to coach were the students who had no real responsibilities outside of classes, yet who found their free time slipping through their fingers no matter what they did.
Among all of the other problems with easy AI cheating, I wonder how much the availability of these tools will encourage even more procrastination. Feeling like you always have the fallback option of having ChatGPT write the homework for you leaves the door open to procrastinating even longer
> I asked my students to complete a baseline survey registering their agreement with several statements, including “It is unethical to use a calculator in a math class”
Unless there was more to this survey, this wording seems misleading. In a college-level math class, using a calculator is a common expectation depending on the type of class and the problem. The students would probably think of their TI-89, not a magical AI calculator that could solve every freeform problem for them.
I've gone back to university as an adult to study econometrics, already with a software background.
It has been quite an adjustment, the hardest part isn't really the content, there have just been assessment tasks that are more demanding in time than cognition. I think a big part has been these are introductory subjects.
You have much less agency over your schedule outside university itself than you would have when you're working as well, which has been a massive PITA having previously worked as a software engineer for quite a while. As an institution it just doesn't respect peoples time, which is fine when you're much younger and you're just coming out of school (another institution that doesn't care for your time outside of it).
I think my problems have been compounded by the fact this is a undergraduate degree not a masters. There was only a masters for applied Econ in my area (which is much less data and math focused than econometrics), I've had this conversation 100 times, not looking to repeat that, trust me the moment I see an option to change to a masters in my area of interest in the city I live in, I would jump at it. But my choice to go back to uni to some extent is an as much an act of consumption as it is getting a piece of paper to saying "you can trust this man with econ".
Anyways in these first year subjects, you'll have assessment tasks like "make a recording of you demonstrating your understanding of <basic function in excel>, explain the value of <basic function in excel>". They are easy subjects but they are also really time consuming so to some extent it feels degrading. I would take steps to skip the subject based on my prior experience but the focus of these subjects isn't even Excel, its just thrown in there because they anticipate you have little experience with these things and it'll be useful in later subjects when it becomes assumed knowledge.
Edit: phrasing
> They find ways to run out of time no matter how much free time they start with
Despite how much I wish it weren't, this is exactly me. I seem to be able to only work under pressure, so as I got later into my degree I would start my work later and later, until it was standard that I'd start at 4am on the day it was due. A terrible way to function or to learn
Only being able to work under pressure is textbook ADHD.
the fact is that this is a terrible feedback loop, because every time you started late, but manage to finish on time, it is reinforcing the idea that you're "good" and can manage it even if you start late! This self-reinforcing condition makes you later next time, but because of your capability, you still make it on time.
But at some point, you either stop procrastinating as you find the absolute limit, or fail outright (which is why survivorship bias exists if you look at the student body and see the amount of procrastination).
>They find ways to run out of time no matter how much free time they start with
IMO you’re observing Parkinson’s law, “work expands to fill the allotted time”. The students who take a million classes look like they have time management in check, and I’m sure some do, but they also are benefiting from the inverse of Parkinson’s law — if they can _only_ allot X hours a week to a task, they’re going to make the most of those X hours. This practically holds regardless of the student’s time management skills. I should know, I have horrible time management, and only succeed by overcommitting and rising to the challenge.
[1] https://en.wikipedia.org/wiki/Parkinson's_law?wprov=sfti1
I think this is covered in Micheal Easter’s notion that as societies become more comfortable, our brains lower the threshold of what constitutes a “problem” in our lives. We’re wired not only to be great at problem solving but also discovering new ones. Think this is based on prevalence theory related research in psychology.
Different tools in different years as expectations have changed. The TI-89 is incredibly powerful, but has to give way to MATLAB and Wolfram Alpha. It used to be productive to “google” your problems. Going further, there’s now LLMs writing python code to do calculations. Hard to say what’s next, but I’m sure what is considered ethically questionable today will be acceptable and the new thing will be the new questionable tool.
To be clear; there are plenty of contexts where having a TI-89 is 100% unambiguously cheating. There are even more places where MATLAB and Wolfram Alpha are cheating.
Please share examples of contexts where TI-89 is cheating
A times table test like we all took in grade school. A test applying Newton's method. Any test where the calculator can do the work instead of you, and a TI-89 can do quite a bit of work.
And more trivially/procedurally, the numerous tests I had to take in high school and university which explicitly said "no graphing calculators" (and the occasional test that said "no calculators" outright).
Thanks
I was thinking of exams rather than day to day learning of basics, seeing tests as teaching rather than assessing. Point being you can't cheat at learning.
The calculator can't fake that you understand e.g. how to differentiate. The working is the bulk of available exam marks but the calculator only gives answers.
Is it unambiguous cheating if it doesn't help?
My Calc 1 class forbade all calculators. Grades were also completely based on quizes and exams. Homework assignments were all optional and un-graded.
All forms/structures in society that think that a TI-89 is cheating deserve to have as many people as possible use TI-89 and similar CAS's.
Ludditism is the human death drive externalized for the modern age. Reject it in all of its incarnations.
There's objectively no value with learning how to perform calculations by hand that CAS's can perform better automatically.
We have well over the number of human beings on earth amounts of calculator capable calculating devices. I will not "be in a situation where I don't have a calculator".
Time spent not teaching folks how to use computational aids of math and teaching them to do it the "hard way" is time robbed from them.
Those who invent systems which subvert or short circuit the attempts to enslave individuals into doing hard work for hard works sake are ontologically good. Those who try to defend work for works sake are ontologically evil and I hope that they reincarnate into a durian fruit in the next life.
> I will not "be in a situation where I don't have a calculator". [...] Those who invent systems which subvert or short circuit the attempts to enslave individuals into doing hard work for hard works sake are ontologically good. Those who try to defend work for works sake are ontologically evil and I hope that they reincarnate into a durian fruit in the next life.
By that logic, you must also permit students another extremely well-established labor-saving technique: Hiring someone else do the homework/exams on their behalf!
After all, for students from certain families, forcing them to use the calculator personally would likewise be "robbing time from them", denying them valuable experience honing their skills in managing and subcontracting. They'll never be in a situation where they don't have a calc^H^H^H^H lackey.
I trust we can agree that (A) permitting outsourcing is absurd, (B) that cheating is not "ontologically good" even when reduces the total human labor, and (C) prohibiting the practice should not put a teacher into a state of stinky-fruit damnation.
I submit that we do care about how students do it and which skills they use, even if we disagree on what those are. It is not as simple as saying that good education means spurring them to supply Valid Answers By Any Contemporary Efficient Means.
You take these classes in order to internalize the intuition for the subject. That's the point of the class. Getting the right answers is only important insofar as it serves that goal.
Your attitude makes sense in engineering but not in a math class for example.
From my experience:
We were not permitted to use a TI-89 in math courses since we were expected to learn the underlying concepts, which is much more than learning how to use a tool.
The reasoning was similar in physics. The instructors couldn't care less whether we used the CAS functionality of the TI-89 because that wasn't a part of their curriculum, but they were concerned about students downloading solver programs. (Most of the instructors would agree that creating our own programs to solve problems would be a valuable learning tool, but they had no way to determine whether we created those programs or someone else.)
Academic courses tend to be biased towards laying the foundations so that we can build upon our own knowledge. There are other types of courses that are purely concerned about applications and the use of tools.
What about work for learnings sake?
Frankly, I'm quite glad to be able to do basic arithmetic in my head -- having to pick up a calculator every time I need to multiply two numbers would burn through far more time over the course of my life than what I spent in Kindergarten (or whenever) learning to do it myself.
All you want is for people to be better slaves for your capitalist machine. Never learning how to do things themselves: only consuming, buying what could otherwise have been achieved by the simple (but non-GDP-increasing) exercise of their own mind. Your rhetoric betrays your own class.
The idea of work abolitionism / being anti-work is somehow turned into an accusation of wanting to enslave folks for capitalist ends.
You can't even imagine a reality any different than "Capitalist Realism"
Go read Mark Fischer.
Try reading more than one leftist book: Fischer is great, but he doesn't try to talk praxis. So I'm talking about pragmatics here: just pretending that "we can abolish work by automating it" is pure ideology, totally ignorant of the centuries of historical examples to the contrary. Any time you get back will immediately be seized. What matters are your personal capabilities, what the working class is capable of doing with their own hands. You advocate waiting on the bourgeoisie to hand over control (what, out of good will?), but Fischer could tell you that only by seizing power for themselves can the working class ever become free.
> Any time you get back will immediately be seized.
I imagine some guy several thousand years ago: "Once we finish teaching this animal to pull the plough for us, descendants of The Tribe Between The Two Rivers shall have lives of pure leisure!"
I feel like you're glossing over the contexts of "acceptable" and "questionable".
The tool itself is not questionable or acceptable, it becomes questionable or acceptable depending on the usage. A pencil and paper can be questionable if the test is designed and expected to be completed without it.
You can design tests where an LLM spitting out python is an expected tool, but what are you testing for then? I doubt there are classes that teach whatever that test would be for yet.
Being one of those students and with a career under my belt of process analysis and coaching, I have an interesting observation: I harness free time as an explicit part of my writing process, rather than something that interferes with it.
I write at about 1200 words per day and I considered each fo the major multi-week assignments in my entry-level English courses to be worth no more than one day of my time apiece. For the finals, I gave them two days apiece, because I wanted an extra day to define the scaffolding for my argument.
My mother indicates that this is how she went through college too; very occasionally, a serious paper would require more effort than this, but for the most part it was “load assignment into brain, study assignment mentally until T-2d, write assignment, submit”. If several essays are due, then they have to be staged at various days numbered T-2d through T-5d for example — and it’s really important to not depend on T-1d existing at all due to courseware/internet/power outages.
I could technically write a worse essay the day it was assigned, but ultimately, I’m turning in A-tier work by this method. The hardest lesson was that I have to try not to wait until T-1d, because there’s a lot of risk encoded in that and it outweighs the value derived from having an extra day to think about it while I do other things.
But it wasn’t about “free time slipping away” — it’s just that I’m writing crap throwaway work that doesn’t matter after it’s done, and so I can barely motivate to care relative to literally anything else in my life that matters. Thus the T-2d compromise: I’m not about to give them precedence over literally anything, but I will concede that I do need to do so one day early, however boring it feels, because I’d rather have a crap day at T-2d than the same crap day at T-1d with the unproductive anxiety of risking a class-retake if my internet drops out.
Notably, when I actually genuinely care about what I’m writing, I’ll spend weeks researching sources and studying arguments and selecting quotes and then assembling it all over a couple days into a work of art — but assembly day is still always as late as possible in the time window assigned, because by then I’m most able to think and write about it efficiently and with a minimum of frustration. Not a zero of frustration, that is — I am a grouchy writer — but I’m healthy-grouchy on T-2d and bitter-grouchy on T-1d, so I do make the effort to put in my writing that day early now.
So: for your coaching efforts, try working with students to construct a working calendar that has non-writing activities in the leadup and then writing activities at the end. ie assuming a 7-day window,
T-7d: Assignment given: Read the assignment. (Seems obvious; is not obvious!)
T-6d: Think about your argument during your free time, while playing games or out at coffee or whatever.
T-5d: Try to construct a very halfass outline on a piece of paper. One sentence per argument you’d like to make, draw arrows to rearrange them. Not complete sentences, not punctuated, doesn’t have any structure at all. Point is that trying will help brain coalesce.
T-4d: Research references for fun. End up with far too many. Start highlighting quotes to yourself using highlighter or digital tools. If you’re going to experiment with a new tool, get it working and productive in 2 hours or discard it and do something shittier.
T-3d: Bind quotes to your argument phrases from that halfass outline. This may force reorg of outline; cool. Compile Works Cited from whatever you end up using so that you don’t have to fuck around with it tomorrow.
T-2d: Write paper, referring to outline / phrases handwritten note. Do one paragraph at a time. Plan to spend your entire day on this with 1 hour away from desk handling bio/sanity needs for every 2 hours at desk. Enforce that upon yourself.
T-1d: Finish whatever writing you didn’t feel like you were prepared to write on T-2d. Ideally try to do this earlier in the day than later, since that every hour you let this slip towards midnight l measurably increases your chances of a life outage causing you to fail the class.
The point of this schedule is to bake in the daydreaming / slow cooker aspect of the creative process but to keep it on the rails. I play video games extensively when I’m thinking about a paper because I can feed my literary brain the assignment to simmer and then go occupy my reflex brain with the game. I usually end up having to use some T-1d time but I’m getting better at managing my life’s dependencies ie. Food and Water and Sleep so that I’m more reliably at T-2d completion :)
> One common theme I've noticed is that their class load virtually doesn't matter. They find ways to run out of time no matter how much free time they start with.
12 credits is normally a minimum. That's roughly 12 hours a week in classroom, give or take. You need 3x that number of hours out of the classroom--that's 36 hours.
So, you're at 48 hours of academic work every single week. A 15 credit load means a 60 hour week.
Most people working jobs would start to complain about burnout at 50+ hours per week for 4 years running. They would almost certainly complain at 60 hours per week.
Most 3 credit hour classes are really 2.5 hours. And almost no one spends 36 hours hours outside of class on a 12 credit schedule unless they messed up and signed up for 4 difficult classes in the same semester. You definitely aren’t spending that much time outside of class all semester long.
You also have to consider that a semester is 16 weeks. The first week or 2 of each semester is very light.
So we’re talking 30/52 weeks a year for most people.
For most people, you’ll never have that much free time again in your life.
College feels like a lot of work because you aren’t good at time management yet. And you remember the last few weeks of each semester where you are actually extremely busy.
> The first week or 2 of each semester is very light.
Please tell that to the undergrad STEM professors, please. Almost all of mine had an assignment first class that was due by the third class.
Freshman engineering is generally Calc I, Physics I, Chem I, and English Composition/Writing and often some random engineering/computer thing. I assure you that schedule sucks even harder that it looks like it sucks, and it gets more time consuming as the years progress.
While there were lots of Party Hardy(tm) types in the College of Arts, the ones I knew who were taking their degrees seriously were working every bit as hard as the STEM folks. Possibly, they were working harder as they needed a lot more extracurricular work and achievement since what they were doing didn't have nice, clean objective measures like STEM does. They spent a lot of time being unpaid labor at functions and networking like crazy.
By contrast, no matter how many hours they worked me at my summer internship, it simply never compared to the grind at school.
Imo those numbers are pretty inflated unless you’re taking a full load of the hardest classes offered. Usually you pair some GE requirements or electives with heavier material. I really don’t want do some sort of humble brag here, so I’ll just say that if I followed your math it would come out to like 90+ hours. I promise, I was not that diligent.
Altman’s analogy didn’t hold up. Calculators were uncontroversial
Calculators are uncontroversial now. But when they first became cheap and widely available, they were not allowed in math classes. Then only four function calculators, then graphing calculators. But still today, programmable calculators are prohibited in many academic contexts.
The point that you're (and everyone is) glossing over here is relative positions on the skill gradient.
A first grader probably would be prohibited from using any kind of calculator on arithmetic tests, 4-fn or not. But 8th graders are usually permitted scientific (non-programmable) calculators.
As you go up in grade level, you "get access to" calculators capable of functionality at the level below you. Because the point is that when we're educating students we want them to actually learn the subject matter, but once we've deemed them to have understood it and we have them move onto the next goal, we give them the tools to make that prior goal easier. We lessen the burden of the little mechanical concepts they already know so that they have an easier time becoming familiar with the next more advanced concepts.
AI systems are so much more advanced than what's capable on a TI programmable calculator. It's hard to draw clean boundaries around the tiers and enforce them by telling the model "help the user with tasks of tier 1-4 but not 5+". That's the issue, that it's really infeasible to strictly use them strictly as learning tools. You can almost do it with a lot of self-discipline and self-reflection to analyze your own workflow, but it's not generalizable across domains.
I mentioned the continued ban on programmable calculators in many academic contexts. Those contexts still include some portions of undergraduate education. This is fifty years after the introduction of programmable calculators.
Realistically, the answers the students gave the teacher were probably motivated by the practical benefits that come with giving teachers the answers they want to hear…bullshit questions are likely to produce bullshit answers. It’s not like first year college students haven’t had twelve years of academic standards moralizing talked at them.
Why can't they be restricted to produce only the concepts of a grade below you? It sounds doable and is actually a great idea.
Can LLMs be reliably restricted to produce any specific subset of content? AFAICT they're still consistently jailbroken.
It’s easy to do at the syntactic level by controlling the sampling. For example, it’s easy and common to restrict output to be valid JSON but just not allowing any tokens that would make it not valid JSON.
But reliably restricting output at the semantic level is very much an open problem.
How are you going to ensure that it is impossible for the student to work around whatever measures you take?
>>> imagine how radically math class must have changed when calculators became widely affordable
It didn't.
I was in math class when calculators were introduced. At least for high school level and beyond, the curricula were designed to make problems solvable without calculators, and they weren't of much use. This was still the case when I taught an undergrad college math class in 1997. Graphing calculators were allowed, and the kids who tried to use them just screwed themselves up.
I would have gladly changed the curriculum to use calculators and computers from the very beginning. As tools, and not just to administer the same old exercises and quizzes. Give them Jupyter Notebook. Math education has never been a success story.
Education faces a dilemma, which is that it has always used heuristics to guide study and assess performance. Exercises such as the "three paragraph essay" had no use in the real world, even long before AI could generate them on demand. When one of those heuristics is broken, another one has to be found. Even word processing forced teachers to grade papers on content, rather than mechanics.
> Exercises such as the "three paragraph essay" had no use in the real world
The rigid formatting imposed by graded assignments may not have any use. But on the other hand, having completed a liberal arts/"soft" science degree before my CS degree I've greatly benefited in the workplace from the writing experience it provided. I had to write so many papers that it became more or less effortless to produce long form, well structured writing and those written communication skills have helped me distinguish myself far more than my technical ability at work.
The generic ChatGPT overly formal corporate tone has no nuance or subtlety and is a poor substitute for well crafted, deliberate communication. I am always conscious of how my exact words and phrasing would be perceived by the intended audience, frequently requiring a balancing act between competing interests while maintaining clarity. Due to that I manage to avoid stepping on toes or sabotaging relationships due to inartful phrasing. It's frustrating to receive emails consisting of LLM boilerplate because it has such low information density and is so much more difficult to infer tone and emotion from the other side .
I'm very grateful I completed my education without the temptation to just churn out low effort writing or code and depriving myself of that experience. I'm not confident at all I would have been able to maintain that self-discipline.
I'm with you there. Though I have a technical degree, I went to a liberal arts college, and preferentially took courses that had a heavy writing component. I was a good typist thanks to coding, and one of the first students to use word processing. Today, I still use those skills to my advantage.
As a teen in the late 80's I had an HP calculator that I programmed to compute molecular weights given an input string like "H2SO4". It felt like having a secret superpower, especially when I participated in competitive exams. I was a very straightlaced kid and would not have used the program if it such things were explicitly forbidden, but as far as I could tell, they never were.
Reminds me of when I write a j2me app for matrix diagonalization because we could use the old feature phones as calculators. Nobody thought we'd be mad enough to use those to cheat...
Do you still have the source code for the j2me app?
I hadn’t yet learned to program back when I was still using a feature phone, but I have a lot of fond memories of J2ME applications that I installed on my phones. Mostly games, of course.
I encourage anyone that wrote J2ME games and utilities, no matter how small or big, to upload the source to GitHub :)
That one would be definitely lost to time...
Did you tell your teachers about your superpower?
But normally it depends on the subject and if the automation/machine solves the primary skill being teached or if its just a "secondary/tertiary" skill. Are you in a Calculus 101 class? Calculators like TI-89 are likely to be prohibited when examining for deriving analytical solutions for derivatives and integrals.
Statistics, Physics or any other subjects that needs applied maths? Such a calculator is probably a minimum requirement to take the course.
My HP-49g+ was definitely load-bearing going through EE. I was never much good at memorizing big sheets of formulas but I was pretty good at memorizing a couple of simple differential equations (e.g. I(t) = C dv(t)/dt was easy, v(t) = v_s * e^(-t/RC) wouldn't stick). So I'd just... derive all of the "special case" formulas from scratch during the exam. Usually they were simple enough that I could just get them into the right form but I'd lean on my calculator doing the symbolic integration for me when they weren't.
The other thing it was awesome for was solving systems of linear equations. I could do the nodal or loop analysis just fine, I'd write down the matrix that represented the system of equations and then just punch that matrix in and invert it.
You did the capacitor DE in your head? The analytical solution requires a “trick” (integrating factor) usually taught in an intro DE class.
I was on pretty good terms with my chemistry teacher, so...maybe? It's been a while, but I don't remember either showing it off or taking pains to keep it secret. To adults, that is; my nerdy friends and I delighted in showing off the cool stuff we did with our calculators.
I vaguely remember thinking that one likely reason shortcuts like mine were not prohibited was because no one in charge suspected that such things were even possible with current technology, or if they were, that a child would be able to exploit it. But as long as I kept to the letter of the rules, I considered myself ethically in the clear.
> But as long as I kept to the letter of the rules, I considered myself ethically in the clear.
Yeah, totally, just to be clear I'm not judging.
In fact, if you programmed it to handle those operations, one could argue you had already learned a big chunk of what was going to be measured in the exams.
Kind of similar to the paradox of creating and using cheat sheets, is highly likely you're accidentally learning about the subject matter in the process of writing the sheet, sometimes up to a point where the cheat sheet is not necessary anymore.
> Kind of similar to the paradox of creating and using cheat sheets, is highly likely you're accidentally learning about the subject matter in the process of writing the sheet, sometimes up to a point where the cheat sheet is not necessary anymore.
The problem is (example from mathematics): even if you are capable of deriving some formula (you thus understood the topic well), it takes a lot of time in the exam. Looking at the cheat sheet is much faster - in particular when the time is somewhat precious in the exam.
A related personal story: During my statistics course in high school, we discovered that the TI-89 had some statistical functions that the TI-83 didn't have. So, the rule was that if we wanted to use the TI-89 ones, you had to write an application for the TI-83 one. It was a great way to really learn the algorithms.
Stuff like that seems harder to do than learning the damn thing correctly
I could do it correctly from the get-go. The program just saved me from drudgery many times over. Probably enough times to recoup my time investment to create the program, but in any case I enjoyed coding for its own sake.
I you like Chemistry, then yes. If you like programming but dislike chemistry, then no.
I always felt (and my maths teachers agreed) that if I understood something sufficiently to automate it, I’d proved my point and didn’t need to do the rest of the exercises.
Edit: Automate in the sense of coding it myself, not in the sense of downloading some software.
The calculator tricked them into studying. Same trick as the "one note card, front back" but in this case accidental.
The concepts of adding machines and calculators were also slowly phased in over the span of a century. The first commercially successfully adding machines hit the market in the 1890's, and pocket calculators took off in the 1980's. AI went from theory to answering hand written math homework questions from a photograph in a few years.
I only had a calculator (at a technical university) starting in the mid-1970s. Prices were dropping like a stone in about that period. In high school it was pretty much slide rules.
Totally correct. In the 90's as a kid in school using a calculator was highly debated amongst teachers and the ability to bring one out on your desk depended on the teacher.
In grade 2 i had a teacher who would say "I don't believe in erasers", you know, the things that "undo" pencil. As a ~6 yr old i actually didn't understand this phrase: "Well I have one, they're real!"
It's also an extremely misleading comparison. Basic calculator functions do not in the slightest replace anything taught in a maths class. Using ChatGPT not just to write entire paragraphs (replacing composition), or even providing the writer with ideas (replacing the creative aspects of writing) isn't comparable to adding two large numbers together.
The equivalent in maths would be if you handed students a theorem prover or have Wolfram Alpha give you step-by-step solutions and obviously nobody to this day allows this, because like ChatGPT for writing it'd defeat the point, that students think.
When I was in uni we were allowed basic but not programmable calculators during exams and a lot of CS classes even were pen&paper, if the prof was a bit hardcore
Turns out education done right is vaguely a speed-run of how the knowledge was developed. Adding calculating tools makes sense as you advance the the corresponding point in the process. Honestly, I think there should be a chunk of precal and calc where they use slide rules only, then calculators of increasing complexity (or just increasingly complex features of one calculator).
"When will I use this in real life" is a declaration that you have no expectations of learning the next lesson that builds upon this one.
>> Altman’s analogy didn’t hold up. Calculators were uncontroversial
> Calculators are uncontroversial now. But when they first became cheap and widely available, they were not allowed in math classes.
The author of TFA means specifically for his cohort of students, not in general. He polled his students, and the result was that they thought calculators weren't seen as unethical but they were more skeptical/uncertain about AI. By his current students, now, not in general.
LLM’s have been widely available for approximately five years.
Five years into the availability of a calculator with an arbitrary advanced feature, it was controversial in academic contexts. Some of the author’s students could be grand-children of students from the early days of consumer calculators.
The author is comparing a new technology with an old one. And ignoring programable calculators which are still sometimes banned after fifty years…and many of the author’s students probably have used LLM’s for homework despite their statements that please the author.
The author is doing nothing of the sort.
Among other things, he's analyzing his students attitudes' re: AI and cheating with AI, and also comparing what they claim to feel vs what they actually do! It's mostly a reflection of what his students feel about the use of AI in English writing, not about calculators vs AI.
It seems as if you're responding to the line you quoted (out of context) through your preconceptions of what the article is about instead of actually reading the article!
> and many of the author’s students probably have used LLM’s for homework despite their statements that please the author.
"Probably have used LLM's"!? Don't take this the wrong way, but it seems you didn't read past the line you quoted, am I right? Because this is explicitly addressed multiple times, and the answer may surprise you.
> Calculators are uncontroversial now.
Yes, they are uncontroversially bad. Schools that don't use them have higher scores.
Unfortunately, even SAT/ACT have calculator slop now.
No middle schooler is using a graphing calculator on their algebra exam
I literally used a graphing calculator on my algebra exams in middle school.
Yeah I don't believe you, especially because you have an axe to grind about AI singularity bullshit. No one in their right mind should allow a graphing calculator to be used on an algebra exam, might as well let them bring a laptop and open Wolfram alpha.
Maybe my teachers didn’t know you could program equation solvers into it. Either way it helped me understand that those in authority didn’t understand the nature of the world changing beneath them, which prepared me for how many people would end up misunderstanding AI and under-appreciating it despite how much it is evidently changing the world.
> imagine how radically math class must have changed when calculators became widely affordable
I was there. We had been given slide rules, and a decent chunk of my 11-yr-old maths classes were devoted to teaching us how to use them. Calculators were banned because they "didn't teach us anything" (but somehow slide rules did? It didn't make any sense at the time either).
Over the course of the next few years, calculators became more acceptable, and by my 18-yr-old Maths A-level class, we were being advised on which scientific calculators to buy.
It's an interesting analogy, as TFA says.
Really enjoyed that. Shows the "messy middle" that people are caught in during this current wave of AI tech. I think one undeniable positive outcome has been the like collective introspection AI has sparked among so many people. With a tool challenging what it means to be "human" or "creative" (more so than prior technological advances), I've been seeing a lot of wonderful discussions, articles, videos, etc. with people wrestling with those questions and also just affirming their own singular voice and unique creative essence. It's been cool to see.
I think one undeniable positive outcome has been the like collective introspection AI has sparked among so many people.
Has it sparked that? I'm doubtful but I'd be very interested if anyone has a references about such introspection.
I grew up reading Douglas Hofstadter back in the 70s and what I appreciated with his ideas was using AI to illuminate what is human. His wave of AI failed, of course. But still, it's disheartening how little of that kind of inquiry seems missing from the current wave of AI.
This essay resonated with me because it highlighted the similarity between AI-written texts, describing the result as a word salad. And this also reminded me about some words from my teacher of Russian Literature: that the "bright future" themed novels of the pre-WW2 Soviet writers — works produced under strict political control — read like one big novel without a beginning and an end, and not as separate works.
And this grayness and sameness is what happens when people are forced to "think" as a chorus, either by the authorities or their censorship, or voluntarily by using the same AI's help.
Or by market consolidation, as we're experiencing now.
> most students are overcommitted; college is expensive, so they need good grades for a good return on their investment; and AI is everywhere, including the post-college workforce.
Yeah. Overcommitted to partying and skipping class.
Has this author ever been to an average American university?
Let's rephrase it to "most of even the best students" then.
I went to about the cheapest US school that had a decent math program. It costs $17,500/yr between tuition, rent, books, and rice and beans.
That's a lot of money. It's over $10/hr in pre-tax income, even if you work full-time all year, which isn't an easy bar to clear in the sort of towns with cheaper universities. Wages don't scale well enough with more expensive tuition for there to be substantially better options.
Classes are another 22+ hours each week (you could complete school in 3-4 yrs instead, but that makes it even harder to afford and doesn't really reduce the workload enough to make a difference, however I'll also factor in a 15-hour workload later).
The rule of thumb is that you should study 3 hours for every hour of class. I found that approximately correct. Some classes took a lot less. Some took a little more. Combined with the self study you need in adjacent topics, 3hrs is a fair bit low.
During the school year then, you have something that looks like a 128hr/week schedule, or 100hr if you're finishing in 4yrs, and still 60hr/week even if you're finishing in 4yrs and racking up $70k in debt.
Don't get me wrong; I had free time (I worked more during the summer, less during the school year, allowing loans to cover the slack, which bought extra time here and there), but it wasn't exactly a party either. When I skipped class it was because I had to work, had to study for some other more pressing class, or found it more efficient to study the book than to try to understand that particular lecturer.
This matched my experience 30 years ago. Work 20 hours a week, but tuition and living expenses were a lot cheaper back then ($215-$330/month for a room! $900/quarter tuition). The 3 hour for every 1 hour of class is especially true for computer science, and skipping class in favor of self study worked well if the lecturer was really bad. Lectures were pretty much bonus reinforcement if useful at all, a lot of what you learned relied on self study.
A lot of students didn't do what I did, and they washed out pretty quickly (I had a lot of classmates from HS that didn't last the first quarter). My first quarter was pretty harsh (only got one 4.0, and a 2.7 in a chemistry class I had no reason being in), but I wised up quickly. It was hard going from High school where I could do all my homework in the time between classes, to college where I had to do real actual studying.
> This matched my experience 30 years ago.
You know that's fair, I hadn't considered the generational differences to be this vast.
My experience was only 7 years ago. I don't think it's a generational difference so much as the fact that a university is a big place, enough so that even if you engage actively with 100+ people you won't see the whole picture.
It has gotten a lot more expensive and competitive. I’m almost embarrassed to mention how crazy cheap my school was in 1995 compared to 2025. Also, I doubt I would have gotten in with my high school achievements even though I graduated in a fairly decent cum laude position. Life for kids these days is a lot tougher.
The numbers I quoted were current prices for the school/city I attended (UND, if anyone wants to cross check). When I attended prices were lower, but so were wages and a number of other things, roughly proportionally AFAICT.
I could maybe see your point about admittance (I had something like a D average, maybe C- or C+ or something, in high school), but I think my financial estimates were about correct. Is something majorly incorrect?
Even if so, there are a dozen cheaper states in the country with halfway decent programs (if any future internet denizen is reading this, Fayetteville Arkansas is actually great for math right now, both in quality and in cost/jobs). I doubt my observations are too far off-base for a typical student trying to go to school economically.
> The rule of thumb is that you should study 3 hours for every hour of class. I found that approximately correct. Some classes took a lot less. Some took a little more.
lol stop the cap.
Its more like 0.5 hours of grand total work for every hour of scheduled class, since most kids are skipping 50% of the classes and using "chat" (a really fetch name for a new digital drug btw) to make up the rest.
College is just a hoop, remember?
What people are telling you in the comments is that your perspective is not universal. I've personally only ever skipped one class in my time in undergrad (as an American at an American university), and not for a party. I'm not a special case or anything, those classes are very expensive!
You are much closer to a special case than you think. Average attendance in my and my colleagues Math 101 classes is around 30% by mid semester.
i would guess math 101 to be the least attended class across the board though
Isn't it basically a remedial class, or rather a class for people who did the minimum of HS math? I would expect it to be filled with more partiers than studiers, and also that plenty of the people in it would already know a lot of the material that's covered (so why not ditch?)
>plenty of the people in it would already know a lot of the material that's covered (so why not ditch?)
Their test scores show that they don't already know a lot of the material that's covered.
>Isn't it basically a remedial class, or rather a class for people who did the minimum of HS math?
It's College Algebra (or the non-trig part of Precalculus).
In 2012 I took my first course in undergrad. The teacher asked everyone to introduce themselves and say something they liked to do. Every single student, except myself, said they like to party. That was shocking to me. Most students weren’t skipping classes though.
Their writing skills were also abysmal. Frankly if they were that bad by college they didn’t seem likely to improve in my opinion.
Overcommitted in this context probably means "has a schedule that is packed too full"
Between work and school and other responsibilities they have no time to decompress so they burn out
The only time I’ve had as much free time as in university was in my first job.
Since you said "in university" I assume that wasn't in the U.S.
I've been taking one class at a time for the last six years at what I think is an average American university. The twenty somethings that your comment is aimed at have been my lab partners and such. You're describing maybe 10% of them. As a group they're all over the map.
As learning goes, I'd say anybody taking more than two classes at a time is overcommitted. They might manage to get A's but I speak with these students about the courses we took together a year or two after the fact and it's clear to me that taking four or five classes at once is an awful strategy for retention.
Speak for yourself maybe
Speaking for myself, has the author ever been to an average American university?
Heh, especially the non-major freshman classes. A few weeks in and half the seats are empty compared to the first day of class.
I have, I'd recommended speaking for yourself
Cool, but has the author?
It me.
But I read an article recently about the death of partying in the USA: https://news.ycombinator.com/item?id=44514550
> several statements, including “It is unethical to use a calculator in a math class”
That's too broad to be of any use, if the math class is teaching you to calculate in your head, then using any calculator is cheating. If the math class is teaching some algebra equation solving skills, then using a programmable calculator that auto solves them is cheating.
That's the similar issue with such experiments - they unfortunately aren't rigorous to provide any insight into education
Interesting report.
I understand why much of the discussion about AI and university education has focused on first-year writing classes in the U.S. Some of my own first experiments with ChatGPT in December 2022 were having it write school-like essays [1].
Over the past few months, I cotaught a university class in which we also had first-year students use and reflect on their use of AI in their classwork. But the context was different: the class was a seminar on science communication (how to make science engaging to children and the general public), and most of the time was spent with the students doing group projects. Also, the class was at a university in Tokyo, and we taught it in Japanese.
We have just started analyzing the feedback from the students, but my impression is that they were less conflicted about the use of AI for their group work than they might have been if they had been doing their projects as individuals.
Meanwhile, as the semester progressed, agentic frameworks started to mature. I spent a lot of time on my own experimenting with Claude Code and Gemini CLI. While none of the students in that class seemed to use them, it became clear to me that such higher-level cognitive tools will pose an even greater challenge to higher education than essay-writing chatbots do now.
[1] https://www.gally.net/temp/202212chatgpt/defaultessay.html
AIs grading AI-generated essays looks like a recipe for model collapse. That's why we certainly need people who go after the "diminishing returns" of improving their writing skills beyond the "good enough" that AI delivers.
Should education systems aim for that for all of their students? Certainly, because AI alone is not sufficient to raise the bar. As impressive an AI is when it seems to invent a new molecule, it is still only possible because of the original works of many people.
What a lovely essay. Reminds me of the way I loved the liberal arts growing up. I missed having classes like that in college (AP'd and ACT'd my way out of most requirements).
English teachers seem especially prone to that friendly and sporting demeanor the author has. Professors from the engineering schools are far more prescriptive, probably due to the nature of the material.
>English teachers
Since discovering Analytical Philosophy, I think it is irresponsible to combine Nonfiction and fiction under the term 'English'.
As an engineer, I write emails, they need to be clear, factual, etc... This is in huge contrast with fiction, where writers get merit for being intentionally ambiguous with things like metaphors and symbolism.
What an incredible disservice to students and society to consider English(nonfiction) an 'art'. It should be treated like math and science.
I had to become a middle aged adult and learn this for myself.
All the best writers that I know in the sense that you mean (communicating information precisely), including non-native speakers, are also avid fiction readers. Many also write fiction or prose for fun. Familiarity and fluency with the details of usage and vocabulary are what let one employ these things precisely for whatever purpose, fictional or not.
That is not my experience at all. To the contrary, those who write fiction read mostly fiction, and those who write non-fiction read mostly non-fiction.
They're incredibly different skill sets. One is all about argumentation, convincing, facts, and citations. The other is all about imagination, beauty, evocation, flavor.
Obviously they both require assembling nouns and verbs and other parts of speech in sentences, but they seem to be virtually entirely different capabilities at the end of the day.
Writing excellent short stories doesn't really help with crafting effective business communications, design documents, etc. And vice-versa. In fact, I think they can sometimes even be harmful -- the kind of clarity required for non-fiction can constrain imagination in fiction, while the creativity celebrated in fiction can be quite counterproductive when it comes to functional communication -- what is intended to be clever or unique often gets misunderstood.
To back up your point, I kind of hated English class until my senior year of high school when I took AP English Language (nonfiction), after which I started drinking books from a firehose.
> I had to become a middle aged adult and learn this for myself.
This is a cliche.
You can’t write precisely without an understanding of how language becomes imprecise, of its fundamental instability. Precision and delicate use is an accident when it does happen, and its happening can never be proven. We must have faith in the accident.
I disagree strongly with this, it's like saying you must understand the subtleties of calligraphy or typography in order to be proficient at writing in your notebook with a pencil. I have no doubt you will be more purposeful and deft with your handwriting having this knowledge but they're two completely different skills.
You can be taught to and be proficient in "writing with your pencil" by learning the rules [1]. Efficient, practical, immediately useful and applicable. No subtlety required nor desired. It's the same as all practical skills or trades.
[1] https://stylepedia.net/style/
I agree that it's a good distinction to make. Personally I haven't thought about it till I read On Writing Well by William Zinsser. In the book he specifically teaches writing nonfiction and even shares an anecdote where he was a guest on a radio show promoting a writing conference and was annoyed with the host because he conflated writing with literary works.
So yeah, I recommend the book to people interested in writing.
More traditionally you'd study "rhetoric", the art of making your arguments appealing. It doesn't really matter whether the things you say are true or false.
Rhetoric is valuable in any writing endeavor; clarity is only valuable sometimes.
For a funny take on the whole "rhetoric" is the use and abuse of logic some people might enjoy How to Win Every Argument by Madsen Pirie which also happens to be where I plucked the tagline regarding rhetoric from. It's a pretty easy book to go through in toilet break sized increments, the author goes through different fallacies and how they're employed one by one along with various rhetorical devices.
Though a few years ago when I searched for a book on rhetoric and making convincing arguments Office Of Assertion by Scott Crider also popped up, but it's aimed more at written rhetoric instead of what most people have in mind.
> For a funny take on the whole "rhetoric" is the use and abuse of logic
But it's much broader than that. You can make true arguments. You can make confusing arguments. And you can use tools that have nothing to do with logic at all. Rhetoric has a lot to say about rhythm, alliteration, and linguistic structure. And a lot more to say about your personal bearing and your tone of voice.
Hand written essays in class. Short and longer form discussions, questions and answers in class. End of unit, term, semester, year. Interview on topic in spoken form.
That doesn't mesh with the $5000 per credit hour ,get students in and out and degree'd ASAP. No 100 level outside of Marlboro college are you going to find a graduate sized class cohort meeting with professors 1 on 1 and debating philosophy and doing stand up rhetorical studies.
Oh yeah, Marlboro is out bidness.
Works for Oxford and Cambridge.
Here are some of the essay topics I had them read aloud:
[...] I expected them to laugh, but they sat in silence. When they did finally speak, I am happy to say that it bothered them. They didn’t like hearing how their AI-generated submissions, in which they’d clearly felt some personal stake, amounted to a big bowl of bland, flavorless word salad.
But that's what it has been trained on - almost all academic writing is bland flavorless word salad, and this is extremely noticeable in title fads. I have a nearly decade-long game running with my friend where me make up absolutely bullshit concepts that could nevertheless be plausibly published in a journal, and the process has been going on long before that.
'Verbing the noun: towards a genericization theory of expressivity in high-entropy counter-heterogeneity' describes an ongoing problem in academic writing where novelty is deprioritized in favor of acceptability by an evermore tightly circumscribed set of peer professionals whose socioeconomic interests favor the establishment of intellectual stasis that maximally conserves positionality in a quais-Simmelian network space parameterized by income, tenure proximity, and citation count.
Or put more clearly, the more academics write to impress each other instead of to reach the public, the more generic their titles and language will be. Being able to parse and regurgitate wordy titles and abstracts constitutes table stakes in academia, so the incentives tilt toward burying the lede any original proposals as deeply as possible so as the minimize the career-damaging possibility of rejection on technical/syntactical grounds.
So many video ads!
It occurs to me that writing by those in STEM fields and those in the humanities is entirely different and each group dislikes the other’s writing. When I was in college, my professors in technical classes had no problems with my writing. After I graduated and I wrote some technical articles, my writing received praise from readers. However, when I took two semesters of mandatory English writing classes in college, my professors hated my writing and nothing I did made them happy with it.
When LLMs became widely usable, I was one of the people who really liked much of the writing that they did. I found it was relatively close to my writing style, which I consider to be good, despite the disagreement from those in the humanities. It was close enough to my own writing that I have even had people on Discord accuse me of using LLMs to write my messages for me, when I had not.
The linked article was clearly written by an English teacher. He criticizes AI-generated texts as “a big bowl of bland, flavorless word salad“. Now, there are many cases where LLMs output nonsense, but in cases where the writing logically flows, does not self-contradict in any way and avoids unnecessary repetition, “bland” and “flavorless” are good. The goal of writing is to convey information across space-time; writing that is “bland” and “flavorless” is the best way of conveying information.
I can see a number of things he did in his writing to avoid being “bland” and “flavorless”, and I consider them to be examples of poor writing:
He used dozens of idioms that make the text difficult for non-native speakers and unpleasant for native readers. He used a number of colloquialisms, including some that are inappropriate in professional contexts (although I will not repeat them since I refuse to write them). He used a word whose only definition is provided by Urban Dictionary and therefore is not even an official word:
https://www.urbandictionary.com/define.php?term=cyborging
He also brought politics into an apolitical topic. The injection of politics is a great way to derail any form of productive dialogue and should be avoided.
He used a story format of the kind that has infected journalism. It is very rare that the process by which something was learned is useful to readers and presenting it for dramatics wastes their time. That is with the exception of stories on the topic of security, where hearing the process is often genuinely informative. My aversion to this writing style is so severe that I have a standing policy to stop reading a news article the moment that I see that it uses this style for a topic that is not security-related. After reading his article, I will extend my policy to apply to essays by academics in the humanities too.
He made numerous attempts to evoke emotional responses to elicit agreement, rather than to make clear arguments based on facts. This is great for propaganda, but not so great for making points. Every one of these appeals to emotions is poor writing.
Beyond those things, he also did not properly cite sources in multiple places. To name a few, the quotes from Sam Altman and Annalee Newitz are uncited. As an academic, he should know better.
Some of these things might actually have places in certain types of writing. They certainly have places in propaganda. They also have places in fictional literature. However, they do not have places in attempts to argue a point.
I imagine if he corrected all of my criticisms, he would find the result to be “bland” and “flavorless”. That is how an attempt to argue a point should be.
Ai being a calculator is a bad analogy. A typical calculator doesn’t do the full assignment for you.
AI can do the full assignment and do it faster and better