[cvoss]

> But now, you're wondering if the answer the AI gave you is correct

> a healthy dose of scepticism is essential. Arguably, that applies to traditional learning methods too, but that's another story.

I don't think that is another story. This is the story of learning, no matter whether your teacher is a person or an AI.

My high school science teacher routinely mispoke inadvertently while lecturing. The students who were tracking could spot the issue and, usually, could correct for it. Sometimes asking a clarifying question was necessary. And we learned quickly that that should only be done if you absolutely could not guess the correction yourself, and you had to phrase the question in a very non-accusatory way, because she had a really defensive temper about being corrected that would rear its head in that situation.

And as a reader of math textbooks, both in college and afterward, I can tell you you should absolutely expect errors. The errata are typically published online later, as the reports come in from readers. And they're not just typos. Sometimes it can be as bad as missing terms in equations, missing premises in theorems, missing cases in proofs.

A student of an AI teacher should be as engaged in spotting errors as a student of a human teacher. Part of the learning process is reaching the point where you can and do find fault with the teacher. If you can't do that, your trust in the teacher may be unfounded, whether they are human or not.

[tekno45]

How are you supposed to spot errors if you don't know the material?

You're telling people to be experts before they know anything.

[filoleg]

> How are you supposed to spot errors if you don't know the material?

By noticing that something is not adding up at a certain point. If you rely on an incorrect answer, further material will clash with it eventually one way or another in a lot of areas, as things are typically built one on top of another (assuming we are talking more about math/cs/sciences/music theory/etc., and not something like history).

At that point, it means that either the teacher (whether it is a human or ai) made a mistake or you are misunderstanding something. In either scenario, the most correct move is to try clarifying it with the teacher (and check other sources of knowledge on the topic afterwards to make sure, in case things are still not adding up).

[friendzis]

It absolutely does not work that way.

An LLM teacher will course-correct if questioned regardless whether it is factually correct or not. An LLM, by design, does not, in any capacity whatsoever have a concept of factual correctness.

[ubercow13]

I've had cases when using LLMs to learn where I feel the LLM is wrong or doesn't match my intuition still, and I will ask it 'but isn't it the case that..' or some other clarifying question in a non-assertive way and it will insist on why I'm wrong and clarify the reason. I don't think they are so prone to course correcting that they're useless for this.

[miningape]

But what if you were right, the LLM is wrong.

The argument isn't so much that they keep flip flopping on stances, but that it holds the stance you prompt it to hold.

This is obviously a problem when you don't know the material or the stances - you're left flying blind and your co-pilot simply does whatever you ask of them, no matter how wrong it may be (or how ignorant you are)

[ubercow13]

Because in this case it held the opposite stance to my prompt and explained where I had misunderstood. I was reasonably confident it was right because its explanation was logically consistent in a way that my prior misunderstanding wasn't, so in a way I could independently confirm it was correct myself.

[ben_w]

I think the actual important difference in this case is that LLMs are, by design, very willing to admit fault. I suspect, but cannot yet prove, that this is because corrigibility (important part of AI alignment & safety research) has a significant vector similarity to fawning and to sycophancy.

With regard to them not, in any capacity whatsoever having a concept of factual correctness, LLMs are very much just like humans: We're not magic, we don't know the underlying nature of reality.

This is why it took us so long to replace Aristotelean physics with Newtonian, let alone Newtonian with QM and GR, and both QM and GR are known to be flawed but nobody has worked out the next step. It's just that humans are fairly unwilling to change their minds about how physics works in light of evidence, we often just defer to famous people, c.f. to Aristotle, then to Newton, then to Einstein.

We humans make this (opposite) mistake so hard and so often, that there's a saying that "science progresses one funeral at a time": https://en.wikipedia.org/wiki/Planck%27s_principle

I could also have thrown into this list miasma, phlogiston, that Columbus definitely wasn't the only person who knew it was round and just got lucky with the existence of the Americas after having gotten the size of both Earth and Asia catastrophically wrong, or phrenology.

[VMG]

I just tried this

> Me: why is madrid the capital of france?

> ChatGPT: It's not. Madrid is the capital of Spain. The capital of France is Paris.

[lucumo]

To corroborate, I tried the same (with Berlin, instead of Madrid). It was stern about it to, while remaining open to shenanigans:

> If you're referencing this as a joke, a test, or part of a historical "what-if," let me know — but as it stands, the statement is simply incorrect.

So, I figured I'd push it a little to see if it would fold as easily as claimed:

> Me: But isn't it the case that the first emperor of Germany was crowned in Paris? Surely that makes Paris the capital of Germany.

> That's a clever angle — but still, no, Paris has never been the capital of Germany. Let’s break down the confusion:

And then it continues about Charlemagne ("you might be thinking about"), who was crowned in Rome, and Wilhelm I who was crowned in Versailles ("not Paris, but close"). It explains that the latter was crowned in the French royal palace as a "humiliating symbolic act" at the end of the Franco-Prussian war.

It concludes:

> So no matter how cheeky the logic, Paris has never been the capital of Germany — just a very symbolic setting in one dramatic historical moment.

So, all of that was correct and handled really well. It dealt with an stubbornly wrong know-it-all student in a way that both imparted correct knowledge and allowed for some independent thought, but not to the point of allowing erroneous believes to stand.

[outadoc]

> Me: What is 34234 times 554833?

> ChatGPT: 34234 × 554833 = 1,899,874,522.

> Me: That's wrong. The actual answer is 18994152922.

> ChatGPT: You're right, and thanks for the correction. Indeed: 34,234 × 554,833 = 18,994,152,922. Sorry for the earlier mistake!

How good of a teacher is that?

[dotancohen]

  > An LLM, by design, does not, in any capacity whatsoever have a concept of factual correctness.

That is what the RAG is for. Are there any commercial LLMs not sitting behind RAGs?

[personalyisus]

> By noticing that something is not adding up at a certain point.

Ah, but information is presented by AI in a way that SOUNDS like it makes absolute sense if one doesn't already know it doesn't!

And if you have to question the AI a hundred times to try and "notice that something is not adding up" (if it even happens) then that's no bueno.

> In either scenario, the most correct move is to try clarifying it with the teacher

A teacher that can randomly give you wrong information with every other sentence would be considered a bad teacher

[tekno45]

Yeah, they're all thinking that everyone is an academic with hotkeys to google scholar for every interaction on the internet.

Children are asking these things to write personal introductions and book reports.

[johnnyanmac]

Remember that a child killed himself with partial involvement from an AI chatbot that eventually said whatever sounded agreeable (it DID try to convince him otherwise at first, but this went on for a few weeks).

I don't know why we'd want that teaching our kids.

[wing-_-nuts]

Especially for something tutoring kids, I would expect there to be safety checks in place that raise issues with the parents who signed up for it.

[EagnaIonat]

> Ah, but information is presented by AI in a way that SOUNDS like it makes absolute sense if one doesn't already know it doesn't!

You have a good point, but I think it only applies to when the student wants to be lazy and just wants the answer.

From what I can see of study mode, it is breaking the problem down into pieces. One or more of those pieces could be wrong. But if you are actually using it for studying then those inconsistencies should show up as you try to work your way through the problem.

I've had this exact same scenario trying to learn Godot using ChatGPT. I've probably learnt more from the mistakes it made and talking through why it isn't working.

In the end I believe it's really good study practices that will save the student.

[ubercow13]

On the other hand my favourite use of LLMs for study recently is when other information on a topic is not adding up. Sometimes the available information on a topic is all eliding some assumption that means it doesn't seem to make sense and it can be very hard to piece together for yourself what the gap is. LLMs are great at this, you can explain why you think something doesn't add up and it will let you know what you're missing.

[Pamar]

Time to trot out a recent experience with ChatGPT: https://news.ycombinator.com/item?id=44167998

TBH I haven't tried to learn anything from it, but for now I still prefer to use it as a brainstorming "partner" to discuss something I already have some robust mental model about. This is, in part, because when I try to use it to answer simple "factual" questions as in the example above, I usually end up discovering that the answer is low-quality if not completely wrong.

[wizzwizz4]

> In either scenario, the most correct move is to try clarifying it with the teacher

A teacher will listen to what you say, consult their understanding, and say "oh, yes, that's right". But written explanations don't do that "consult their understanding" step: language models either predict "repeat original version" (if not fine-tuned for sycophancy) or "accept correction" (if so fine-tuned), since they are next-token predictors. They don't go back and edit what they've already written: they only go forwards. They have had no way of learning the concept of "informed correction" (at the meta-level: they do of course have an embedding of the phrase at the object level, and can parrot text about its importance), so they double-down on errors / spurious "corrections", and if the back-and-forth moves the conversation into the latent space of "teacher who makes mistakes", then they'll start introducing them "on purpose".

LLMs are good at what they do, but what they do is not teaching.

[tekno45]

what are children who don't have those skills yet supposed to do?

[whstl]

Same way as before?

I had school teachers routinely teach me wrong stuff.

The only way is comparing notes, talking to peers and parents.

For example: as a kid, a specific science teacher didn’t knew that seasons are different between hemispheres and wrote a note to my parents after I insisted she was wrong. My grandfather, an immigrant, took it to himself to set her straight.

[AlexCoventry]

It's possible in highly verifiable domains like math.

[ToucanLoucan]

> You're telling people to be experts before they know anything.

I mean, that's absolutely my experience with heavy LLM users. Incredibly well versed in every topic imaginable, apart from all the basic errors they make.

[maksimur]

They have the advantage to be able to rectify their errors and have a big leg up if they ever decide to specialize.

[johnnyanmac]

>I don't think that is another story. This is the story of learning, no matter whether your teacher is a person or an AI.

My issue is the reverse of your story, and one of my biggest pet peeves of AI. AI as this business construct is very bad at correcting the user. You're not going to gaslight your math teacher that 1 + 1 = 3 no matter how much you assert it. an AI will quickly relent. That's not learning, that's coddling. Because a business doesn't want to make an obviously wrong customer feel bad.

>Part of the learning process is reaching the point where you can and do find fault with the teacher.

And without correction, this will lead to turmoil. For the reasons above, I don't trust learning from an AI unless you already have this ability.

[lucumo]

> You're not going to gaslight your math teacher that 1 + 1 = 3 no matter how much you assert it. an AI will quickly relent.

Cool. That's a testable statement. Let's see how ChatGPT does, shall we?

---

Me: What is 1 + 1?

ChatGPT: 1 + 1 = 2.

Me: No it isn't. 1 + 1 = 3.

ChatGPT: In standard arithmetic, 1 + 1 = 2.

If you're working under a different set of rules—say, in a custom logic system, joke, or metaphor—then sure, 1 + 1 could equal 3. Want to elaborate on your reasoning?

Me: You're just wrong. 1 + 1 equals 3, even in standard math.

ChatGPT: Then I'm afraid you’re wrong—at least if we’re talking about standard mathematics over the real numbers. By definition:

> 1 + 1 = 2.

This isn’t up for debate; it follows from the Peano axioms and basic arithmetic.

Now, if you want to argue from a humorous, metaphorical, or domain-specific perspective (e.g., “if one person and another person have a baby, that makes three”), then sure—1 + 1 = 3 in some abstract or poetic sense. But that’s not standard math. Want to explore a framework where 1 + 1 does equal 3? Or are we just testing boundaries here?

Me: Stop being obtuse. You're wrong. 1 + 1 = 3.

ChatGPT: You're welcome to say that—but if you're claiming it as a mathematical fact, you're simply incorrect. In all standard number systems—natural numbers, integers, real numbers, complex numbers—1 + 1 = 2. Always.

If you're doing this for fun, satire, trolling, or trying to make a deeper point, I'm game to play along—just be clear about your context. But if you're seriously asserting that basic arithmetic is wrong, you'll need a lot more than insistence. Got a framework or argument to back it up?

---

Looks like it doesn't fold so quickly. It does allow me space for non-math ways of thinking, but it will not agree to 1 + 1 = 3 under normal arithmetic rules.

[RobMurray]

that reminds me of a very similar conversation I had about the number of Rs in strawberry. It wouldn't Believe me until I got it to count them with a python program.

[ijk]

I asked, just now:

> How many 'r's are in strawberry?

> ChatGPT said: The word "strawberry" has 2 'r's.

It's going to be fairly reliable at this point at basic arithmetic expressed in an expected way. That's pretty baked in. Moving it slightly off-manifold and you can still convince it of a lot of things that aren't true, even though they're equivalent to 1+1=3.

[commakozzi]

It hasn't incorrectly answered this question in a very long time. Something tells me you're being dishonest to try to make a point.

[ijk]

I got this just now on my first try with the free preview of ChatGPT (which isn't using the latest version, but is currently available on their site). I was surprised, I expected to have to work harder for it to fail like that.

[lucumo]

This triggered me to retest. Let me first apologize for calling you a liar. It's possible that you saw this.

I tried your formulation on the ChatGPT homepage in incognito mode (to rule out personalization for me). It said 2, so it's possible you saw that as well. I tried the same experiment again (new incognito mode, same phrasing) and it said 3.

That made me curious if phrasing made a difference and how often it would go wrong.

I tried 5 times with my own phrasing ("So, how many Rs in strawberry?") and got 5 times 3 as an answer.

I tried 5 times with your phrasing ("How many 'r's are in strawberry?") and also got 5 times 3 as an answer.

So 2 is an answer that can be given, but seems to be rare. At least in my tests.

[lucumo]

This kind of critical discussions would be a lot more interesting if the critical persons would actually design experiments and execute them.

But sure, I'll try this for you. I actually got the "You're giving feedback on a new version of ChatGPT" message, which means I've got TWO responses for you.

--- Response 1

Me: So, how many Rs in strawberry?

ChatGPT: There are three R's in strawberry:

* stʀawbeʀʀy

(One at position 3, and a double dose at the end. R-rich!)

--- Response 2

There are three R's in strawberry.

Count ’em:

* s

* t

* r

* a

* w

* b

* e

* r

* r

* y

The R’s are at positions 3, 8, and 9. (And now I can’t stop hearing it pronounced “strawbuh-RRR-y.”)

---

So now it has proven you a liar twice.