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

161–170 of 672 posts

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

#161

Post by Gigachad »

wisty wrote: post #133
fn-mote wrote: post #83

Earlier quoted context omitted.

Variations in this system are in active use in the US as well. Do you feel it is effective? It seems to me that there is a massive asymmetry in the war here: proctoring services have tiny incentives to catch cheaters. Cheaters have massive incentives to cheat. I expect the system will only catch a small fraction of the cheating that occurs.

You can't stop people hiring someone who looks similar from sitting the exam, or messages in morse code via Bluetooth. It's hard to stop a palm card. But it stops a casual cheater from having ChatGTP on a second device.

Both of those are so hard and so expensive that usually just learning the material is more practical.

LLMs and remote exams changed the equation so now cheating is incredibly easy and super effective compared to trying to morse code someone with a button in your shoe.

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

#162

Post by frelupin_ »

ramraj07 wrote: post #152
frelupin_ wrote: post #147

Earlier quoted context omitted.

> it gets work done quickly and poorly This is only temporary. It will be able to code like anyone in time. The only way around this will be coding in-person, but only in elementary courses. Everyone in business will be using AI to code, so that will be the way in most university courses as well.

It already can. Im flabbergasted how people haven't still figured out how good gemini 2.5 is.

Claude 3.7 and 4 are better for me than Gemini 2.5 for vibing with legacy code. Gemini 2.5 has some great solutions if you handhold it, but tends to make too many assumptions about what would be better which can tear things up as an agent, imo. In other words, Gemini is smarter, but less practical when working with existing code, from what I’ve experienced. To each their own, though.

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

#163

Post by const_cast »

gonzobonzo wrote: post #153
const_cast wrote: post #81

Earlier quoted context omitted.

Typical classroom experience works and has worked for thousands of years. Edutech is pretty new and virtually all of it has been a disaster. Sitting in a lecture and taking notes on paper is tried, tested, and research backed. It works. Not for everyone, but for a lot of people.

Talk is cheap; if you want to see what people really believe, ignore what they claim and look at what they actually do. And when you do that, you see that people generally don't find typical college classes to be worth it outside of the credentials they give. Almost no one with a CS degree goes back to college to take a college algorithm course when they want to get better at algorithms; they study on their own. You…

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 all of them fail.

At the end of the day, we go home and we don't crack open a textbook. We sit and watch TV. Maybe we go for a walk or go to the gym. The vast majority of people do not have the mindset required to be self-educated.

We used to do the "everyone self-educate" thing. Most people couldn't read or write. Humans are unintuitive. You can't just give them access to things and expect results. They require accountability, they require structure. We're not machines, we're faulty fleshy creatures. Our reward feedback loops were never built for self-determination at this high of a level.

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

#164

Post by AStonesThrow »

grogenaut wrote: post #155
wisty wrote: post #133

Earlier quoted context omitted.

You can't stop people hiring someone who looks similar from sitting the exam, or messages in morse code via Bluetooth. It's hard to stop a palm card. But it stops a casual cheater from having ChatGTP on a second device.

You can. I did a remote proctored exam for the NREMT last year. They had me walk the camera around the room, under the desk, etc. All devices had to be in my backpack. No earbuds. They made me unplug the conference tv on the wall, lift picture frames etc. I had to keep my hands above the table the whole time, I couldn't look down if I was scratching an itch. They installed rootkit software and closed down all of the…

I took 3 CompTIA certification tests at a community college testing center. This was the procedure, more or less.

> When you take the test in a proctored testing center location they lock all of your stuff in a locker, check your hands, pockets, etc. They give you earplugs. You use their computer. They record you the whole time. They check your drivers license and take a fingerprint.

While attending there, I also took a virtual Calculus class. The instructor was based in the satellite campus, several miles away. The virtual class required a TI graphing calculator, used Pearson textbook & video lectures, and all the tests and quizzes were in Canvas. I worked from home or the main campus, where there was a tutoring center, full of students and tutors making the rounds to explain everything. I received tutoring every other week.

But then our instructor posted the details on our final exams. We were expected to arrive in-person, for the first time of the semester, on that satellite campus at specified times.

I protested, because everything I'd ever done was on the main campus, and I rode public transit, and the distance and unfamiliarity would be a hardship. So the disability services center accommodated me.

They shut me into in a dimly lit one-person room with a desk, paper, and pencil, and I believe there was a camera, and no calculator required. The instructor had granted an extended period to complete the exam, and I finished at the last possible moment. I was so thankful to be done and have good results, because I had really struggled to understand Calculus.

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

#165

Post by TychoCelchuuu »

foxglacier wrote: post #138

Schools need to re-think what the purpose of essays was in the first place and re-invent homework to suit the existance of LLMs. If it's to understand the material, then skip the essay writing part and have them do a traditional test. If it's to be able to write, they probably don't need that skill anymore so skip the essay writing. If it's to get used to researching on their own, find a way to have them do that whic…

[dead]

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

#166

Post by nyarlathotep_ »

QuadmasterXLII wrote: post #159
jumploops wrote: post #55

A bit off-topic, but I think AI has the potential to supercharge learning for the students of the future. Similar to Montessori, LLMs can help students who wander off in various directions. I remember often being “stuck” on some concept (usually in biology and chemistry), where the teacher would hand-wave something as truth, this dismissing my request for further depth. Of course, LLMs in the current educational land…

The longer I go without seeing cases of ai supercharging learning, the more suspicious I get that it just won’t. And no, self reports that it makes internet denizens feel super educated, don’t count.

Wasn't this the promise of MOOCs in the 2010s?

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

#167

Post by eru »

timr wrote: post #140
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'm not that old, and yet my university CS courses evaluated people with group projects, and in-person paper exams. We weren't allowed to bring computers or calculators into the exam room (or at least, not any calculators with programming or memory). It was fine. I don't see why this is so hard, other than the usual intergenerational whining / a heaping pile of student entitlement. If anything, the classes that requi…

> I don't see why this is so hard, other than the usual intergenerational whining / a heaping pile of student entitlement.

You know that grading paper exams is a lot more hassle _for the teachers_?

Your overall point might or might not still stand. I'm just responding to your 'I don't see why this is so hard'. Show some imagination for why other people hold their positions.

(I'm sure there's lots of other factors that come into play that I am not thinking of here.)

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

#168

Post by eru »

frelupin_ wrote: post #162
ramraj07 wrote: post #152

Earlier quoted context omitted.

It already can. Im flabbergasted how people haven't still figured out how good gemini 2.5 is.

Claude 3.7 and 4 are better for me than Gemini 2.5 for vibing with legacy code. Gemini 2.5 has some great solutions if you handhold it, but tends to make too many assumptions about what would be better which can tear things up as an agent, imo. In other words, Gemini is smarter, but less practical when working with existing code, from what I’ve experienced. To each their own, though.

The Claudes are a lot worse at even mildly challenging algorithmic problems than Gemini 2.5 Pro.

However, most legacy code is fairly primitive on that level, so my observation is in no way contradicting yours.

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

#169

Post by eru »

presentation wrote: post #160
zer00eyz wrote: post #141

Earlier quoted context omitted.

> 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. > small handful of dum-dums made it all the way to exam without having connected two dots, and I have to fail them ... won't lead them to a meaningful existence I don't see a problem, the system is working. The same group of people that are going to loose their job to an LLM arent getting sm…

Ideally the system would encourage those dum-dums to realize they need to change their ways before they're screwed. Unless the system working is that people get screwed and cause problems for the rest of society.

I want to agree with your point, but also: someone who's middle-class enough to make it to uni in the first place won't cause much trouble for society.

Paternalism in the sense of 'we know what's better for you than you do' is perhaps justified for those people who really don't know better. But I don't think we should overextend that notion.

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

#170

Post by TychoCelchuuu »

datahack wrote: post #146

There is a tremendous lack of understandings between the genx and millennial teachers and the way they see and use AI, and how younger people are using it. Kids use AI like an operating system, seamlessly integrated into their workflows, their thinking, their lives. It’s not a tool they pick up and put down; it’s the environment they navigate, as natural as air. To them, AI isn’t cheating—it’s just how you get things…

Decades of research into learning shows that "desirable difficulty" is not, as you put it, "just a fetishized relic of an agrarian education system designed to churn out obedient workers, not creative thinkers." Rather, difficulty means you are encountering things you do not already understand. If you are not facing difficulties then your time is being wasted. The issue is that AI allows people to avoid facing difficulties and thus allows them to waste their time.

You think we will make progress by learning to use AI in certain ways, and that assignments can be crafted to inculcate this. But a moment's acquaintance with people who use AI will show you that there is a huge divide between some uses of AI and others, and that some people use AI in ways which is not creative and so on. Ideally this would prompt you to reflect on what characteristics of people incline them towards using AI in certain ways, and what we can do to promote the characteristics that incline people to use AI in productive and interesting ways, etc. The end result of such an inquiry will be something like what the author of this piece has arrived at, unfortunately. Any assignment you think is immune to lazy AI use is probably not. The only real solution is the adversarial approach the author adopts.

Post reply on HN