The role of the University is to resist AI | Hacker News Reader

The role of the University is to resist AI

danmcquillan.org

23 pointsconferza28 comments

This post is close to pure waffle. Yes there are parts of common sense but just to give some example “spice” thrown in amongst other lines:

> Deep learning has historical and epistemological connections to eugenics through its mathematics, its metrics and through concepts like AGI, and we shouldn't be surprised if and when it gets applied in education to weed out 'useless learners'.

It might be partially correct but this is similar to saying Germans should not be trusted because of WWII.

(sad face) this post subtracts from the valid arguments against the usage of AI tools in some valid scenarios, it is because some folks have a knee jerk reaction and label authors as Luddites

> This post is close to pure waffle

Frankly an LLM would have done a better job and been more succinct.

It's bad on purpose to make you click.

The article's main point seems to be that students cheat with AI; hence, universities must "resist" AI to preserve critical thought.

Although universities are certainly against cheating, the responsibility has always been on the student not to cheat. Universities do not oppose useful technologies simply because they may be misused for cheating.

Put another way, the role of a university is to "discover and invent the future." In this light, universities will be more interested in developing AI than so-called "resisting" it. This is especially since it has already yielded breakthroughs in science, e.g. a Nobel Prize in Chemistry for protein prediction.

The role of the university is to show that calling AI "AI" is just for idiots. Intelligence has nothing to do with AI.

Students are like the electrical current, given the path with least resistance they will follow it(as with most of people). Cutting off such a path is NP hard to do at scale IMO. But definitely agree on the idea of the article.

A damning statement of the masses.

How about LLM do away with the pointless painful and thankless brute forcing of huge databases of physical paper to find the key piece of information you want to check something against. Teaching that, that is what academic study is leads to the misconception that writing enough about something makes it valid. A serious problem in many of the areas of new academic prostration.

I'd say this opens research to a new audience who otherwise wouldn't go near it, but frankly we see what damage the internet did to society after people said this in the 90s so yes maybe put it back in the box...

> frankly we see what damage the internet did to society after people said this in the 90s

Isn't ignoring the incredible good the Internet has done exactly what you're advocating against with regards to LLMs?

Yes, new technology that impacts fundamental aspects of our society will have negative side effects that we will have to adapt to and solve - but on the whole, I think the Internet has produced far more good in the world than otherwise. We'll see how LLMs pan out, but I suspect it'll play out similarly.

It seem rather an unfortunate take. Universities have always been about both education and research to understand the world. Ignoring one of the main things happening in the world these days seems to go against that, like a few centuries ago saying the role of the university is to resist science or the printing press or something.

Schools in general are doomed to complete obsolescence now, as every kid can have their own personal AI tutor. Doesn't look much better for universities. Outside of lab work and testing, there is just no further need for them. There, I said it.

That's what people said about Google. Don't worry the worthwhile universities aren't going anywhere.

Of credentialism means nothing and the world turns to meritocracy most of us reading this now are doomed!

I tend to thing that universities will be OK

Even in subjects that just need a mind and maybe a computer like mathematics and computer science.

The main benefit of uni studies is nagging me to learn stuff. There, I said it.

It's been possible for a very long time to try to learn things yourself - and I don't just mean the courses and other resources available online these days; one could always have went to a library and learned that way.

The thing is, people still went to school. Why is that, do you think?

Because you can't just ask a book a question and instantly get presented with an answer at any required level of detail.

Apart from that, coercion and credentialism, where the latter can be provided with testing alone.

How do you know what questions to ask? How do you know the LLM is giving a good answer?

Asking questions and getting answers is only one part of learning; another, more important part, that's missing is putting it in context and laying it out in a manner that teaches the topic in a way that develops actual expertise. This is one part of the actual role of educators - which aren't going to go anywhere. To say nothing of the other benefits of an actual education: credentials, showing you can stick to something long-term, etc.

In any case, I don't disagree that LLMs can and should change education, and for the better - but I don't think universities are going anywhere, nor should they.

Anyway, if learning becomes so easily accessible, wouldn't it be even more important for employers to be able to vet potential hires? Surely companies can't be expected to interview the thousand new applicants that have put applicable skills on their resume simply because they talked to an LLM for a few hours and think that's sufficient to learn it?

Yes, you design a curriculum and feed it to the LLM. One person con do that for a course once, and then you can broadly leave it like that, with occasional updates - while each student gets their own personal teacher.

Educational cost could plummet like a rock while significantly improving results, if we just dare to allow it!

> wouldn't it be even more important for employers to be able to vet potential hires

Yes, I mentioned "testing" twice already. And most of that could be done via LLM as well, at least a coarse pre-screening for the obvious frauds. Also, where are all those people who claim to be Physicists "simply because they talked to an Physicist for a few hours"? You send the children through the curriculum, then you test them, as I said. Doesn't even have to be the same institution that does it.

School as we know it is a dead institution walking, and good riddance, and I very much hope the universities as we know them will cease to exists right along with it.

"Testing" is not really a good indicator of whether you were properly educated - that's why students get an overall grade not solely based on the year-end test.

I'm also not sure why you think testing would not go obsolete with traditional schooling?

> School as we know it is a dead institution walking, and good riddance, and I very much hope the universities as we know them will cease to exists right along with it.

Also, when can we expect this to happen? I mean, I've been waiting for this since computers, then the Internet, and now LLMs have been around for several years, so just curious.

If this is true then why have education at all?

If AI is the equivalent of a college instructor, just deploy AI everywhere instead of wasting time and energy teaching humans, a fairly limited resource?

We exist in an era in which coursework as a medium of assessment has suddenly become nearly worthless. It does not surprise me that since this is the way things have been done for centuries they haven’t quickly rustled up some easy solutions.

No. And there are a LOT of assumptions baked into this about people's passive engagement with models, early model flaws being projected indefinitely into the future, and the effectiveness of AI.

A characteristic of LLMs bad articles is that they try to stack up small multidimensional criticisms in hopes that the at least one would stick.

LLMs curb independent thinking (links some n=5 article published recently)

LLMs reduce wages (wrong!!)

LLMs cause environmental damage (wrong!)

Add in some vague leftist jargon about decolonisation and you have your standard llms bad article. There is I admit, a nugget of truth in each criticism and I hope we can explore it from an unbiased angle.

After thinking about it I have a theory on the real reason some people have a bad opinion on LLMs.

Saying that LLM prevent independent thinking is like saying books prevent independent thinking. It depend how you use them.

Not everyone has a deep understanding of LLMs or the self-control to avoid the temptation of off-loading thinking to one. Many people are desperate enough to look to LLMs for medical advice or even friendship[0]. As workplaces and schools partner with Big Tech[1], LLM interfaces have already become embedded in everyday life as something you can't ignore.

There are no guardrails to LLMs. They remove friction from tasks that used to require critical thinking. We're constantly pressured into using them. I think only blaming the end-user is naive.

[0]: https://aeon.co/essays/our-crisis-is-not-loneliness-but-huma...

[1]: https://www.forbes.com/sites/danfitzpatrick/2025/02/26/chatg...

Oh hell no.

This is like saying mathematics should avoid the calculator.

Don't be so naive, the only AI models have are able to perform more powerful contextual lookups and reference and on the case of chat bots hallucinate when this fails.

These tools have no agency, people who blindly trust it beyond this are the highest of fools and don't understand garbage in garbage out.

The role of a university is to train in the use of tools and that includes the squishy one used for critical thinking. If the university wasn't doing that before now it wasn't doing its job.

None of this pseudo intellectualism and politico opinion posting.

The role of the University is to curate datasets useful for building new models, and perhaps training those models.

If the cost of doing so ever truely drops then yes, verified pure signal to train models on vs everything ever written would make for better models :)

Reply on news.ycombinator.com