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

Harvard’s new computer science teacher is a chatbot

independent.co.uk

Post reply on HN
Print view View original

141–150 of 175 posts

Re: Harvard’s new computer science teacher is a chatbot

#141

Post by NotAFood »

thumbuddy wrote: post #66
RyanAdamas wrote: post #60

This is a terrible indication of what is to come. Cheap, accessible, and fully featured instructors that would otherwise be free and open source, contained and used for profit by in-name-only institutions, organizations and corporations for the purpose of increased profit from the reduction in human labor. Every ounce of AI should be open source and free to the public. This is absolutely technological terrorism on th…

Imagine going to an ivy League school and paying 10,000$ for a chat bot to teach you. What a racket

That's why I went to Berkeley /s

Jokes aside I'm sure our administration would salivate at deploying this for CS61A.

Re: Harvard’s new computer science teacher is a chatbot

#142

Post by some_random »

tbrownaw wrote: post #136
some_random wrote: post #126

Earlier quoted context omitted.

I don't believe that at all, the idea of challenging what you know and how you know it is extremely useful across all subjects. The problem is that the efforts to teach critical thinking has been colossal failures.

Thing is, doing a consistency check on your collection of purported facts is the easy part. What's hard is the soundness check where you need a lot of domain knowledge to see if things actually match objective reality.

I don't see any domain (except politics of course) where asking yourself "is there another explanation for what I'm seeing" isn't useful.

Re: Harvard’s new computer science teacher is a chatbot

#143

Post by chongli »

some_random wrote: post #139
SamoyedFurFluff wrote: post #135

Earlier quoted context omitted.

That’s tautological about whether or not the claim itself has merit. It only says if the claim doesn’t have merit, here is a set of motivations that might have biased the claim. But there’s no way to determine just from that whether the claim has merit. It goes double for the vast majority of the way we get information: ad hoc through conversations with other humans. If a neighbor tells me it’s thunder storming tomor…

You are massively overthinking this. Critical thinking isn't some giant flowchart of questions and rules that you have to run on every single thing you ever hear or think in your entire life, that's obviously insane. Frankly I don't understand how you could even think that. Critical thinking is just saying "Hey this neighbor has a history of lying to me and my other neighbor said that they tried to get a court order…

Critical thinking is just saying "Hey this neighbor has a history of lying to me and my other neighbor said that they tried to get a court order to stop them from putting up a fence. Maybe I should check the forecast before canceling my fence building plans".

How is that not an application of domain knowledge (know your neighbours) with a bit of intelligence? How do you teach people to do that in general? How do you teach people to deal with questions when they lack domain knowledge?

Re: Harvard’s new computer science teacher is a chatbot

#144

Post by some_random »

chongli wrote: post #143
some_random wrote: post #139

Earlier quoted context omitted.

You are massively overthinking this. Critical thinking isn't some giant flowchart of questions and rules that you have to run on every single thing you ever hear or think in your entire life, that's obviously insane. Frankly I don't understand how you could even think that. Critical thinking is just saying "Hey this neighbor has a history of lying to me and my other neighbor said that they tried to get a court order…

Critical thinking is just saying "Hey this neighbor has a history of lying to me and my other neighbor said that they tried to get a court order to stop them from putting up a fence. Maybe I should check the forecast before canceling my fence building plans". How is that not an application of domain knowledge (know your neighbours) with a bit of intelligence? How do you teach people to do that in general? How do you…

Again, you're massively overthinking this. Then entire point is to not blindly accept things and instead actually evaluate claims. You can do that with domain knowledge you have, which appears to what what you're calling all knowledge. If there exists some kind of knowledge that isn't domain knowledge, I'm sure you could find a use for it too.

Re: Harvard’s new computer science teacher is a chatbot

#145

Post by sharma-arjun »

chongli wrote: post #138
anticensor wrote: post #137

Earlier quoted context omitted.

There is a solution to that too: Have the placement exam counted as a proficiency course in the student grades, like foreign language requirements or scientific prep courses in graduate-level schools for those majored in a different field.

My school is a public university. They’re not allowed to introduce mandatory placement exams like that. They have to work within the government-set high school curriculum. They also don’t want to allow the advanced students to “test out” of the intro course, as some schools do, because the advanced students may be highly skilled at programming but totally lacking in introductory CS knowledge.

Yeah, it's extremely common for people to be extremely prolific programmers in high school, without having much knowledge of foundational computer science or algorithm concepts.

Source: Am a high school programmer

Re: Harvard’s new computer science teacher is a chatbot

#147

Post by sharma-arjun »

Lio wrote: post #26
userbinator wrote: post #3

Professor Malan said students would be warned of the pitfalls of the AI, saying they should “always think critically” when presented with information. I want to believe that this will actually happen, but my experience with people blindly trusting what a tool (not even an AI-based one) says, including but not limited to students, suggests that's not going to happen.

My experience with LLMs so far is that they very confidently give you an answer and very convincingly explain that answer but for certain portion of the time that answer is very wrong. You can even get it to break its reasoning down into little convincing but, none the less, incorrect steps. If I was running this course I would make that "artefact" part of the process. i.e. I would tell students that the chatbot will…

One useful trick to avoid some of the pitfalls of ChatGPT for non-creative tasks is to respond to its output with something like:

_"Point out incorrect assumptions or statements in the above answer."_

I got the idea from Khan Academy's [implementation](https://youtu.be/A7REVn9gzgs?t=1208) of GPT-4, where they allow it to 'think' by generating an internal response first, before generating a final response. This apparently improved the veracity of its output by a significant margin.

If ChatGPT doesn't double down on its original answer, then I know definitely not to trust it.

This doesn't solve the underlying problem, because it doesn't ultimately correct the error in a reliable way, but it does help me avoid some of the more obviously garbage output.

Re: Harvard’s new computer science teacher is a chatbot

#148

Post by kevin_thibedeau »

hgsgm wrote: post #88
ekianjo wrote: post #64

Earlier quoted context omitted.

> My experience with LLMs so far is that they very confidently give you an answer and very convincingly explain that answer but for certain portion of the time that answer is very wrong. This is true of humans just as well

Humans are not programmed to avoid saying "I don't know."

There is a prominent culture where this is such a common occurrence that you have to be careful about accepting their confident assertions.

Re: Harvard’s new computer science teacher is a chatbot

#149

Post by eesmith »

Kon-Peki wrote: post #106
eesmith wrote: post #69

Earlier quoted context omitted.

I must be missing something. Harvard has 7,153 undergrads says Wikipedia. At most that's 1,790 students per year taking the class each year. But surely not everyone is taking the course? So why does it need to scale? Harvard tuition is $54,269, says https://registrar.fas.harvard.edu/tuition-and-fees . Assuming it's a 3 hour class with a 120 credit hours to graduate, that's 30 hours per year, so each student pays abou…

> At most that's 1,790 students per year taking the class each year Anyone in the world can take CS50 for credit via Harvard Extension [1], and if you happen to be in the Boston area you are welcome to come to campus and attend lectures, sections, office hours, etc (but they also have everything you need fully online). Like other large classes at Harvard (Stat 110, etc), these operate just like the weed-out courses a…

The pricing doesn't make sense. Harvard Extension charges $2,040/4 credit course[1]. I estimated this course cost Harvard students $5,000, for low-quality access to educational staff.

Does that mean Harvard Extension students get even worse access?

Surely paying twice as much in tuition should mean something more than getting to say "I graduated from Harvard."

[1] https://extension.harvard.edu/paying-for-school/

Re: Harvard’s new computer science teacher is a chatbot

#150

Post by lelandbatey »

eesmith wrote: post #130
wil421 wrote: post #76

Earlier quoted context omitted.

My coworker’s kid just graduated from Harvard. He said it cost about $315,000k. Does your number include all the fees and such most colleges tack on to tuition?

I linked to https://registrar.fas.harvard.edu/tuition-and-fees which shows I quoted only tuition. That's the amount which most directly goes toward paying for educational staff salaries. The total tuition+fees for Academic Year 2023-24 from the same page, is $79,450/year, giving $317,800 for 4 years, and matching your number.

I think they're pointing out that $315k is different from $315,000k as both include the k (for thousand) but one includes the thousands as 0's and the other does not.
Post reply on HN