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

CheatGPT

blog.humphd.org

Post reply on HN
Print view View original

471–480 of 544 posts

Re: CheatGPT

#471

Post by whazor »

Al-Khwarizmi wrote: post #308

As a CS professor myself, I don't think ChatGPT really creates entirely new problems for assessment, but rather makes problems that we had been sweeping under the carpet more obvious and harder to ignore. Any assignment that can be cheated by using ChatGPT could already be cheated before by asking a friend, an expert or paying someone else to do it. But most teachers assumed this doesn't happen often, and thus acted…

For my university, not that long ago, it was possible to finish computer science, without touching a computer.

Still in my time most exams were done on paper. Think of algorithms in pseudocode. But also a lot of logic like reasoning, with process automata, or relational algebra (databases).

Re: CheatGPT

#472

Post by fidgewidge »

nemo44x wrote: post #338
jraph wrote: post #316

Earlier quoted context omitted.

> the continuous assessment fad that has always been rather disastrous for equity Can you expand on this? I lived the two models as a student and still can't make up my mind on which is best. I feel like it just takes catching some bad cold the week of the exams or a bad exam subject / topic to fail them which seems pretty unfair, but I also know my shit better and have more perspective the week of the exam and prefe…

They’re saying that people that were allowed into their programs/universities because of quotas and not merit struggle to prove competency through assessment.

Wouldn't that be an issue with point in time assessment too?

Re: CheatGPT

#473

Post by patrakov »

The real problem is that an _attempt_ to cheat is graded, on average, higher than no submission at all. We need to change that.

Re: CheatGPT

#474

Post by raincole »

yodsanklai wrote: post #424
stillsleepy wrote: post #91

Earlier quoted context omitted.

the other week i gave chatgpt a simple multiplication problem that it got wrong. very simple problem like 86 * 0.0007 or something. but ive been working with chatgpt for 4-5 weeks now and that wrong answer doesnt make up for all the "good answers" that are usually not perfect. like one day i needed to COALESCE in mysql. i didnt know that, but chatgpt did. theres a few times i would have written a function the complic…

I was dismissive at first, but I have to admit chatgpt brings a lot of value to programmers. It saves me a lot of time remembering some APIs or syntax for languages I use only occasionally. It's sometimes wrong, but for programming, it's easy to detect and fix. But it's more problematic for non-programming questions where it's hard to check the answer without googling it.

I don't know why people use ChatGPT for programming. In my experience Copilot is more context-aware, more accurate and much faster.

Re: CheatGPT

#475

Post by asdff »

jefftk wrote: post #435
asdff wrote: post #428

Earlier quoted context omitted.

And a counter for that could be a laptop device with the usb ports disabled. A lot of this stuff is solved with some basic IT imo.

You're now well past the original proposal of a Raspberry Pi. And people who need different input devices for ergonomic reasons (laptops keyboards are not a good fit for many people) are going to push back hard. The device can make network connections, right? Someone's going to come up with a very short program you can type by hand, compile, and then pull down arbitrary other code over the network.

Honestly I appreciate the exercise pentesting my hypothetical academic computer. Lets drop full network access entirely then. Allow it to only connect to a university intranet for assignment deposits, along with plenty of logging of what happened on the device. Lets rule out the "typing verbatim" method too by checking whether the user typed in all in one go or spent time debugging and revising using our own keylogger perhaps. If they need special tooling to interface with the computer, that will probably come with a recommendation from student disability services and be totally legit.

Its a classic arms race, but at the end of the day with enough effort on the issue, IT departments will either win entirely, or make it hard enough to cheat for the vast majority of users that only the extremely small minority who do manage to cheat probably deserve a cs degree. If its hard people won't do it, just like how they keep buying textbooks because finding a free pdf online is only slightly harder, but enough effort to put off most people from forcing campus bookstores out of business.

Re: CheatGPT

#476

Post by psyklic »

_gabe_ wrote: post #452
QuadrupleA wrote: post #405

With all these ChatGPT academic apocalypse stories, I keep thinking - what the hell is wrong with students, that they don't want to actually learn the subjects they've enrolled in? If your goal is to fake-learn and have ChatGPT do your homework, just drop out already. You're paying big tuition for no educational benefit. Maybe your degree gets your foot in the door at some company, but it won't be too long before the…

I'm thinking the exact same thing. It also makes me wonder why professors still have mandatory homework when they know students will cheat. Why not give optional homework, and allow students to have their work reviewed by the professor or TA if the student wants. But if the student doesn't want to do it then, they can skip it and waste their money. Then, have mandatory in class tests that cover the material. I feel l…

Students have lots of competing deadlines. If one class has no deadlines, most students will deprioritize it until the very end when it's too late to realize their mistake. So, having homework worth something (no matter how minor) would be beneficial to most.

It's unfortunate, but a key component of education is motivation -- and deadlines are one key way of providing that.

Re: CheatGPT

#477

Post by singularity2001 »

alexpetralia wrote: post #156
singularity2001 wrote: post #107

Maybe take the approach many physics classes embrace (in germany) : All tools are allowed in tests: calculators, books, as long as the correct answer is derived with "correct derivation path". If chatgpt makes some problems too easy maybe it's a good thing because we can raise the bar. Find problems that require true understanding beyond auto complete / copy pasting.

Almost certainly the bar for problems that ChatGPT cannot solve is far, far higher than a student who is just learning computer science.

Haha try it for some basic functions and algorithms and you're up for a surprise

Re: CheatGPT

#478

Post by MengerSponge »

arvidkahl wrote: post #320
jraph wrote: post #316

Earlier quoted context omitted.

> the continuous assessment fad that has always been rather disastrous for equity Can you expand on this? I lived the two models as a student and still can't make up my mind on which is best. I feel like it just takes catching some bad cold the week of the exams or a bad exam subject / topic to fail them which seems pretty unfair, but I also know my shit better and have more perspective the week of the exam and prefe…

There is also an option outside "one exam" and "many exams" — and that's "no exams." Obviously, this is spectacularly hard to fit into the current system, but you see this a lot in vocational schools where project-based work makes up the majority of the grade and exams are held to comply with regulatory requirements. ChatGPT might just fit into an educational world that values outcomes over fact retention. One may dr…

Well-structured exams are the single most effective teaching tool available. Assessment is a fringe benefit. I'm a strong advocate for "even more exams", where you take the highest score on particular topics.

There are variations, but if you're curious check out "Standards-based grading" or Mastery-based grading. It's in a weird and shitty limbo where a lot of primary and secondary schools have adopted it partially with inadequate training/support, and it's also getting fucked by our shitty publishing industry, same as the botched common-core rollout.

Re: CheatGPT

#479

Post by ryanjshaw »

josephg wrote: post #261
epicureanideal wrote: post #162

Earlier quoted context omitted.

> "What's the weirdest bug you debugged? What made it weird?" If you can remember, 5-10 years after you solved it.

Its an interview question. If you draw a blank on that, rephrase the question for yourself to something like this: "When you think of bugs you've recently fixed in your code, what comes to mind? Which bug had the most interesting story?". If that feels illegal to you, you can check with your interviewer: "I can't remember many weird old bugs, but can I tell you about a bug I fixed the other week?". Interviews aren't…

I fix plenty of weird bugs. But I don't build a reverse index in my mind of "weird bugs".

Re: CheatGPT

#480

Post by ryanjshaw »

taneq wrote: post #376
epicureanideal wrote: post #162

Earlier quoted context omitted.

> "What's the weirdest bug you debugged? What made it weird?" If you can remember, 5-10 years after you solved it.

If you don’t have at least a couple of weird enough debugging war stories to still remember years later, you probably don’t have very broad debugging experience. (Not to say you aren’t a good developer, but you haven’t seen a lot of weird edge cases.)

You're generalising how your brain works to how everybody's brain work. My brain doesn't respond well to questions out of nowhere. I don't maintain a "weird bugs" index. If we are just talking and something comes up there is a good chance I will have a related story, though.
Post reply on HN