Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Trying to teach in the age of the AI homework machine

solarshades.club

Post reply on HN
Print view View original

611–620 of 672 posts

Re: Trying to teach in the age of the AI homework machine

#611

Post by Balgair »

blibble wrote: post #598
Balgair wrote: post #455

Earlier quoted context omitted.

I think you've identified the main issue here: LLMs aren't destroying the University or the essay. LLMs are destroying the cheap University or essay. Cheap can mean a lot of things, like money or time or distance. But, if Universities want to maintain a standard, then they are going to have to work for it again. No more 300+ person freshman lectures (where everyone cheated anyways). No more take-home zoom exams. No m…

Oxbridge supervisinons/tutorials are typically two students, and at a push three (rarely) certainly not anywhere close to ten!

Thanks for the clarification there!

Yeah, 1:3 teacher to student ratio would make university a lot more expensive

Re: Trying to teach in the age of the AI homework machine

#612

Post by NegativeLatency »

sshine wrote: post #69

I teach computer science / programming, and I don't know what a good AI policy is. On the one hand, I use AI extensively for my own learning, and it's helping me a lot. On the other hand, it gets work done quickly and poorly. Students mistake mandatory assignments for something they have to overcome as effortlessly as possible. Once they're past this hurdle, they can mind their own business again. To them, AI is not…

I had numerous in person paper exams in CS (2009 - 2013) where we had to not only pseudo code an algorithm from a description, but also do the reverse of saying/describing what a chunk of pseudo code would do.

Re: Trying to teach in the age of the AI homework machine

#613

Post by Balgair »

voilavilla wrote: post #604
Balgair wrote: post #455

Earlier quoted context omitted.

I think you've identified the main issue here: LLMs aren't destroying the University or the essay. LLMs are destroying the cheap University or essay. Cheap can mean a lot of things, like money or time or distance. But, if Universities want to maintain a standard, then they are going to have to work for it again. No more 300+ person freshman lectures (where everyone cheated anyways). No more take-home zoom exams. No m…

>> (where everyone cheated anyways) This is depressing. I'm late GenX, I didn't cheat in college (engineering, RPI), nor did my peers. Of course, there was very little writing of essays so that's probably why, not to mention all of our exams were in person paper-and-pencil (and this was 1986-1990, so no phones). Literally impossible to cheat. We did have study groups where people explained the homework to each other,…

Even before LLMs, if you walked into any frat and asked to see their test bank, you'd get thousands of files. Though not technically cheating, having every test a professor ever gave was a huge advantage. Especially since most profs would just reuse tests and HWs without any changes anyway.

To my generation, it wasn't that cheating was a 'thing' as much as it was impossible to avoid. Profs were so lazy that any semi-good test prep would have you discover that the profs were phoning it in and had been for a while. Things like not updating the course page with all the answers on them were unfortunately common. You could go and tell the prof, and most of us did, but then you'd be at a huge disadvantage relative to your peers who did download the answer key. Especially since the prof would still not update the questions! I want to make it clear: this is a common thing at R1 universities before LLMs.

The main issue is that at most R1s, the prof isn't really graded on their classes. That's maybe 5% of their tenure review. The thing they are most incentivized by is the amount of money they pull in from grants. I'm not all that familiar with R2 and below, but I'd imagine they have the same incentives (correct me if I'm wrong!). And with ~35% of students that go to R2 and below, the incentives for the profs for ~65% of students isn't well correlated with teaching said students.

Re: Trying to teach in the age of the AI homework machine

#614

Post by milesward »

Balgair wrote: post #564
rwyinuse wrote: post #560

Earlier quoted context omitted.

Over here in Finland, higher education is state funded, and the funding is allocated to universities mostly based on how many degrees they churn out yearly. Whether the grads actually find employment or know anything is irrelevant. So, it's pretty hard for universities over here to maintain standards in this GenAI world, when the paying customer only cares about quantity, and not quality. I'm feeling bad for the stud…

Gosh, I'm so myopic here. I'm mostly talking about US based systems. But, of course, LLMs are affecting the whole world. Yeah, I'd love to hear more about how other countries are affected by this tool. For Finland, I'd imagine that the feedback loop is the voters, but that's a bit too long and the incentives and desires of the voting public get a bit too condensed into a few choice to matter [0]. What are you seeing…

Huh, parameters as LLM bandwidth.

Re: Trying to teach in the age of the AI homework machine

#615

Post by anon84873628 »

voilavilla wrote: post #604
Balgair wrote: post #455

Earlier quoted context omitted.

I think you've identified the main issue here: LLMs aren't destroying the University or the essay. LLMs are destroying the cheap University or essay. Cheap can mean a lot of things, like money or time or distance. But, if Universities want to maintain a standard, then they are going to have to work for it again. No more 300+ person freshman lectures (where everyone cheated anyways). No more take-home zoom exams. No m…

>> (where everyone cheated anyways) This is depressing. I'm late GenX, I didn't cheat in college (engineering, RPI), nor did my peers. Of course, there was very little writing of essays so that's probably why, not to mention all of our exams were in person paper-and-pencil (and this was 1986-1990, so no phones). Literally impossible to cheat. We did have study groups where people explained the homework to each other,…

Here's how cheating advanced since then.

1. People in the Greek system would save all homework sets and exams in a "library" for future members taking a given course. While professors do change (and a single professor will try to mix up problems) with enough time you eventually have an inventory of all the possible problems, to either copy outright or study.

2. Eventually a similar thing moved online, both with "black market" hired help, then the likes of Chegg Inc.

3. All the students in a course join a WhatsApp or Discord group and text each other the answers. (HN had a good blog about this from a data science professor, but I can't find it now. College cheating has been mentioned many times on HN).

Re: Trying to teach in the age of the AI homework machine

#616

Post by const_cast »

HPsquared wrote: post #414
intended wrote: post #405

Earlier quoted context omitted.

Mooc completion rates hovered at single degree percents. The vast majority of people do not complete. The people who do complete are outliers. I suppose we can build for outliers, but then most people are just going to be ignored in this system, and if they have way to respond (vote), they won’t be happy about it.

Most academic education is already built for outliers. Courses are designed around building up to the next generation of professors. Most knowledge that's taught in university is unused, basically wasted education for 99% of the "person-facts" that are picked up by the class.

This isn't true, education as a whole is built for the majority of people. If we look at K-12 it's explicitly built to work best for the neuro-typical child with a normal household. Outliers struggle, and they need other systems to catch them, like special education.

> basically wasted education

People say this but they don't understand how learning things works. Learning is inherently cummulative, everything builds off of everything else. We can't skip steps because then there will be holes. It's a big Jenga tower, and you're essentially advocating taking pieces out willy nilly because you don't think they're important.

Even coding, if you look at it, relies on English. English language arts and computer science are, in a lot of people's minds, completely opposite each other in the sphere of education. But they're actually not - because coding is a subset of the English language. Even naming variables, the people best at it are also the people who are good at conveying meaning in an essay. Because it's the same problem: conveying to an audience your intention as concisely, yet clearly, as possible.

Some relationships are obvious, like you can't learn calculus without algebra without arithmetic. But most are non-obvious.

Re: Trying to teach in the age of the AI homework machine

#617

Post by hirvi74 »

vonneumannstan wrote: post #450
aerhardt wrote: post #437

Earlier quoted context omitted.

Current AI can ace math and programming psets at elite institutions, and yet prior to GPT not only did I learn loads from the homework, I often thoroughly enjoyed it too. I don’t see how you can make that logical leap.

Its a problem of incentives. For many courses the psets make up a large chunk of your grade. Grades determine your suitability for graduate school, internships, jobs, etc. So if your final goal is one of those then you are highly incentivized to get high grades, not necessarily to learn the material.

I think you somewhat touched upon what I believe is the root of the problem:

> highly incentivized to get high grades, not necessarily to learn the material

Based on my own experiences and observations, I think grading is a far larger issue than cheating. I am not convinced that good grades are necessarily reflective of enrichment nor how much material has been learned. If a person makes the a high grade in a particular class, what does that actually mean?

I made high grades in plenty of classes that I couldn't tell you anything about what I actually learned.

Re: Trying to teach in the age of the AI homework machine

#618

Post by tangjurine »

math_dandy wrote: post #28

I teach math at a large university (30,000 students) and have also gone “back to the earth”, to pen-and-paper, proctored and exams. Students don’t seem to mind this reversion. The administration, however, doesn’t like this trend. They want all evaluation to be remote-friendly, so that the same course with the same evaluations can be given to students learning in person or enrolled online. Online enrollment is a huge…

why not pay for students to take the pen and paper exams at some proctored location, perhaps independent of the university.

Re: Trying to teach in the age of the AI homework machine

#619

Post by const_cast »

boredhedgehog wrote: post #273
const_cast wrote: post #163

Earlier quoted context omitted.

Conversely, if we're noticing what people actually do, you'll realize close to zero people who want to pursue computer science are doing it on their own. And not due to a lack of information. The draw of education hasn't been access, not since the internet anyway. Structure, pacing, curriculum, schedule, and measurement cannot be recreated. I've had many people tell me they're going to learn to program online. Almost…

> We used to do the "everyone self-educate" thing. Most people couldn't read or write. Do you regard that as intrinsically problematic? The people themselves weren't unhappy about their state, and society, too, could function well without mass literacy. There was a certain period where we thought training wage workers for their duties required them to be literate, but that might turn out to be unnecessary, if supplyi…

Yes, because it's hostile to democracy and freedom as whole. Even if we could let everyone not read and write and use LLMs, this is just asking for trouble. Reading and writing is precursor for other skills, in fact, just about every skill.

This includes analysis, critical thinking, skepticism, morality, you name it. Without literacy, people are easy to manipulate. It won't happen immediately, but it won't be long until we revert to a world state where organizations like the Catholic Church control everything and order around millions of people to kill each other.

Re: Trying to teach in the age of the AI homework machine

#620

Post by wwweston »

Jensson wrote: post #401
wwweston wrote: post #208

Earlier quoted context omitted.

That’s a magnifier but it shouldn’t be the cause; for that you need a shift in management culture from optimizing for academic missions to optimizing for careers/influence of management and trustees.

Large endowments causes that unless you have very strict rules around it like the Nobel prize endowment. You can see how every large charity starts to focus on growing larger rather than its mission, Mozilla is a good example of that.

I see a correlation in some orgs, but can you explain the mechanics? I think I’ve also seen cases where the original mission continues to be the primary compass, so maybe the mechanics of the failure mode are what I don’t have a clear picture of yet.
Post reply on HN