“ Students are permitted to use AI assistants for all homework and programming assignments (especially as a reference for understanding any topics that seem confusing), but we strongly encourage you to complete your final submitted version of your assignment without AI. You cannot use any such assistants, or any external materials, during in-class evaluations (both the homework quizzes and the midterms and final).
The rationale behind this policy is a simple one: AI can be extremely helpful as a learning tool (and to be clear, as an actual implementation tool), but over-reliance on these systems can currently be a detriment to learning in many cases. You absolutely need to learn how to code and do other tasks using AI tools, but turning in AI-generated solutions for the relatively short assignments we give you can (at least in our current experience) ultimately lead to substantially less understanding of the material. The choice is yours on assignments, but we believe that you will ultimately perform much better on the in-class quizzes and exams if you do work through your final submitted homework solutions yourself.”
It feels downstream of CMU's "reasonable person principle". They know that people are going to use AI on their homework, but they trust that they want to learn and improve their skills -- and this is good advice for doing so.
I'm somewhat biased because I was involved in a previous, related course. The important takeaways aren't really about gritty debugging of (possibly) large homework assignments, but the high-level overview you get in the process. AI assistance means you could cover more content and build larger, more realistic systems.
An issue in the first iteration of Deep Learning Systems was that every homework built on the previous one, and errors could accumulate in subtle ways that we didn't anticipate. I spent a lot of time bisecting code to find these errors in office hours. It would have been just as educational to diagnose those errors with an LLM. Then students could spend more time implementing cool stuff in CUDA instead of hunting down a subtle bug in their 2d conv backwards pass under time pressure... But I think the breadth and depth of the course was phenomenal, and if courses can go further with AI assistance then it's great.
This new class looks really cool, and Zico is a great teacher.
I'm old but cumulative assignments are nothing new (the build an OS class, build a compiler class, etc) and my recollection is after you submitted an assignment the instructor would release a correct version you could swap in for yours. So any bugs in previous modules (that the TA/grader didn't catch) couldn't hold up the current assignment.
Old too, and in my experience that was often slightly more work than fixing the bugs in my own implementation. I did swap out a borked module in the build an OS class once but otherwise used my own.
I don't think the final evaluation is to "cement the understanding" so much as _verify_ that students have taken accountability for their own learning process.
This is what a student, who truly wants to learn rather than simply complete a course / certification, would do... Use AI tools to explain + learn, but not outsource the learning process itself to the tools.
An AI is never going to make me feel bad for asking a "stupid" question, nor will it ever behave as though I have to earn the right to ask a sophisticated one. The answer is, in some cases, almost immaterial. Building in students the willingness to ask questions accelerates learning.
Does one still need to read, take notes, write, act, etc? Yes. The feeling of newfound knowledge that comes from reading/watching a virtuoso demonstration is always a mirage--I suspect most of us have the memory of watching someone do algebra on the board and then having the feeling of understanding melt away later when faced with the page. That's not changing, so far as I can tell.
But we should acknowledge that allowing students to ask questions at will, even ill-posed questions, might have some value.
> What’s your hypothesis of how AI can accelerate how your brain understands something?
What are your beliefs / hypothesis of how having a human teacher can help you understand something?
AI explanations are no longer terrible garbage. The LLM might not be doing original research, but it has definitely read the textbook. :/ And 1000 related works.
You shouldn't believe the LLM when it tells you how to micro-optimize your code, but you can take suggestions as a starting point and verify them.
This is a bad case of whataboutism (I hate this word but it describes the answer you gave), what do you mean by accelerating understanding? Maybe they are good as suggestion engines, but it is very early to state what you did.
Theyre alot more than "suggestion engines". They can reason with you, show you examples, tell you how to dig deeper and verify what theyre saying, etc.
I have some success with this method: I try to write an explanation of something, then ask the LLM to find problems with the explanation. Sometimes its response leads me to shore up my understanding. Other times its answer doesn’t make sense to me and we dig into why. Whether or not the LLM is correct, it helps me clarify my own learning. It’s basically rubber duck debugging for my brain.
Way back when I was a student I had a professor who had a policy that if your homework scores differed substantially from your exam scores, the homework portion of your grade would be disregarded and your final grade would be determined solely by the midterm and the final. It was a harsh policy, and at the time I hated it, but in retrospect it was fair. Seems even more relevant today.
“ Students are permitted to use AI assistants for all homework and programming assignments (especially as a reference for understanding any topics that seem confusing), but we strongly encourage you to complete your final submitted version of your assignment without AI. You cannot use any such assistants, or any external materials, during in-class evaluations (both the homework quizzes and the midterms and final).
The rationale behind this policy is a simple one: AI can be extremely helpful as a learning tool (and to be clear, as an actual implementation tool), but over-reliance on these systems can currently be a detriment to learning in many cases. You absolutely need to learn how to code and do other tasks using AI tools, but turning in AI-generated solutions for the relatively short assignments we give you can (at least in our current experience) ultimately lead to substantially less understanding of the material. The choice is yours on assignments, but we believe that you will ultimately perform much better on the in-class quizzes and exams if you do work through your final submitted homework solutions yourself.”