Live data from Hacker News

Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes

dailycal.org

491–500 of 828 posts

Re: Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes

#491

Earlier quoted context omitted.

In my own experience, the only path I truly gain intellectual benefits is the one where I work closely with the LLM, test very narrow hypotheses, and leverage it for learning over producing. Trying 5N paths is useful and sometimes yields interesting insights I’ll retain, but it’s not the rich, challenging, deeply engaging kind of process I find I need in order to develop useful knowledge and skills. So yes it’s an ac…

I don’t see these at odds. Sometimes through working closely with an LLM, N paths emerge. Having it go off and test each with defined metrics to determine which is better is the natural follow up. Even better if you dive into the why it ended up being better which the LLM seems to be able to expose well in a lot of cases. The part I find weird is all the claims that LLM usage leads to less thinking and exploring and…

You’re right, they don’t have to be at odds. Before LLMs, there were jobs where I had to power through and make a sort of ‘minimum effort’ approximation without applying much analytical or investigative energy or skill. This isn’t a lot different from churning something out with an LLM. There’s not much to learn, the end product is mediocre, it’s more of a rote path.

The only distinction I wanted to make is that the learning doesn’t come by default. Yet that was largely true when people copied mystery solutions from stack overflow and used black box libraries for 90% of the complex work their programs facilitated.

Perhaps not much has changed but we’re now operating at a much larger scale and the opportunity to not be curious is actually more present than ever.

People who are curious are massively benefited by this tooling, in my opinion. Like you’re saying, if you want to investigate and learn, there has never really been a better time. If you’re sincerely applying yourself and pulling all of those threads, there has never been a better teacher.

I’ve wondered about the matter of finding and cultivating curiosity too. I’ve come to believe most humans, let alone programmers specifically, are not all that curious. A lot of us are path-followers and we’d rather not get into the weeds most of the time. Then some of us see weeds and dive in, even when it’s not pragmatic to do so. I don’t know how much it can be cultivated or even removed from a person who has more than enough.

Re: Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes

#492

I have some sympathy for these kids. If LLMs were around when I was a student, I would've also used them to "speed up" my homework assignments then proceed to fail all my tests. Now I work mostly with PhDs who were at the top of every academic environment they've ever been in. And yet I can see their thinking skills rapidly declining as well; many of them can no longer brainstorm, code, think deeply, or write without…

> If LLMs were around when I was a student, I would've also used them to "speed up" my homework assignments then proceed to fail all my tests. You go to a university because you are deeply interested in understanding the subject that you study. Doing the homework and the tests are just the "goalposts" to check for yourself whether you made progress on this. So, as long as you are not under time pressure (which you in…

> You go to a university because you are deeply interested in understanding the subject that you study.

You must come from a wealthy background because what you described is far beyond the vast majority of people's means - at least here in the US.

Most of us go to college because it's the only reliable way to get a tollerable job that pays well. Only a few of my college courses aligned with my interests. The rest were just the price paid for the degree.

Re: Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes

#493

Even as a software engineer myself, I'm feeling a bit of cognitive decline having AI doing some/most of the thinking for me The solution? I'm not sure but possibly use AI as more of a collaborate partner to discuss with rather than letting it give you the answers

>The solution? I'm not sure

The solution is extremely obvious, just stop using it on 2 days out of the week or something like that.

You need to go to the gym, but for your brain.

If what you are building is too complex for you to meaningfully contribute to in the absence of LLM assistance then that should tell you something important.

Re: Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes

#494
post #259
post #38

In unrelated news "More than 600 University of California faculty members, led by mathematicians at UC Berkeley, are calling on the system to reinstate standardized testing requirements for science, technology, engineering and mathematics applicants, saying that six years of test-free admissions has not reliably assessed readiness and professors are often teaching middle school math to incoming students." https://arc…

Who even makes the decision to drop having a standard bar to verify students? And what possible benefit would that have?

I thought for most universities it was because some schools didn't organise the tests and others did during COVID.

Re: Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes

#495
post #457

Earlier quoted context omitted.

Counterpoint, I think this is true for some archetypes of people, but certainly not everyone. I personally use it like the socratic method. I am an intermediate user, I spend a ton of time with LLMs at work and personally, both prompting and letting some crappy agents try to automate boring work. I primarily use Gemini and ChatGPT models, along with some Chinese smaller weight models (eg qwen) locally. If you treat t…

Are you also one of those people who believes that advertising works on other people, but not you?

Advertising is fine (not great), especially when highly targeted and relevant. Spam, misleading, or predatory advertising is not.

Re: Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes

#496
post #171
post #93

Earlier quoted context omitted.

I feel the opposite. My brainstorming has increased rapidly. I can now just throw ideas at an LLM to rapidly validate.

[flagged]

Not judge no. Implement and create a working MVP, yes of course.

And yes, I fire off ideas all over. Many require predicting the future to decide what to focus my individual effort on. This is a terrible way to do things because humans (and LLMs) are notoriously terrible at predicting the future. The gold standard is to try everything and eliminate what doesn't work. This is impossible using human labor. With LLM labor, it's simply a matter of relatively cheap money.

It's amazing. Technical problems are now no longer having to predict what the best implementation is. You can just try each one.

Again, no need to have an LLM judge, because the metrics that define 'better' are well-defined, and this is the interesting part of computer science, not the implementation.

Re: Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes

#497
post #266

The obvious cognitive deterimental effects of using map apps, when we all realized we lose directional sense and our previous ability to navigate without the smartdevices, was society's canary in a coal mine and a headsup of what was coming. WAL-E and Idiocracy. The future.

This goes with calculator to do basic math, contacts apps to store phone numbers instead of remembering them, and watches to tell time imo. Sure we could use our brain power with old techniques to do these, but why? I don't want to do any of these. I'd rather use that brain power for other problems. Same with maps. I don't want to have to store a bunch of location or routing data in my head. I think what you're point…

> I don't want to have to store a bunch of location or routing data in my head.

hmmm, given how closely memory is linked to spatial navigation sense, and not just in humans, but in evolutionary terms-- think squirrels remembering where they buried nuts, birds and fish remembering migration routes, ...

suggests the ability to store location/routing is foundational to much of intelligence.

Even simple tasks, typing, for example, depends on my knowing where the keys are. Imagine if your keyboard re-organized its keymap randomly every third keystroke.

Re: Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes

#498
post #457

Earlier quoted context omitted.

Counterpoint, I think this is true for some archetypes of people, but certainly not everyone. I personally use it like the socratic method. I am an intermediate user, I spend a ton of time with LLMs at work and personally, both prompting and letting some crappy agents try to automate boring work. I primarily use Gemini and ChatGPT models, along with some Chinese smaller weight models (eg qwen) locally. If you treat t…

> If you treat the model like an excellent bluffer, it has never been more fun to challenge a model. To me, there is something deeply intellectually satisfying about "proving" it incorrect, and I like being deeply critical of what the model spits back out. I find that refinement process (with the constant sycophancy turned down in the system prompt) creates a really good loop of critical evaluation that would be hard…

new session. It's easy to lead a model into getting the response you want, deliberately or accidently.

The point is not to literally win an argument (it doesn't matter), it is to use the model like a partner to poke holes in your own understanding. Once it's poked a hole, it has served its purpose. Plus, you eventually run out of context or the model trails off into babbble.

Re: Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes

#499

I have some sympathy for these kids. If LLMs were around when I was a student, I would've also used them to "speed up" my homework assignments then proceed to fail all my tests. Now I work mostly with PhDs who were at the top of every academic environment they've ever been in. And yet I can see their thinking skills rapidly declining as well; many of them can no longer brainstorm, code, think deeply, or write without…

When LLMs and ChatGPT first came out, it struck me as obvious and dangerous to a deep thinker or a knowledge worker the answering capacity. So, from my initial use I did not ask them questions, I have always "done my own work" and then asked the LLMs to criticize that work. This has been an exponential ladder of learning, and my cognitive growth is personally noticeable. I'm not hesitating to scribble out calculus an…

That's nice except when you work somewhere where more and more developers are pushed to pump out slop generated by AI as fast as possible. So far I am not there yet but I have plenty of friends in the industry who are basically 'not allowed' to code manually anymore.

Re: Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes

#500
post #259
post #38

In unrelated news "More than 600 University of California faculty members, led by mathematicians at UC Berkeley, are calling on the system to reinstate standardized testing requirements for science, technology, engineering and mathematics applicants, saying that six years of test-free admissions has not reliably assessed readiness and professors are often teaching middle school math to incoming students." https://arc…

Who even makes the decision to drop having a standard bar to verify students? And what possible benefit would that have?

The decision wasn't specifically to drop a standard bar. It was to drop the existing bars because they have become heavily gamed and are far more reliable indicators of your family's resources than your ability or likelihood of success. That was the equity argument.

Unfortunately, the lost signal wasn't replaced with anything. (I don't know what could replace it. It's an incredibly hard problem. )

Post reply on HN