←back to thread

359 points FromTheArchives | 1 comments | | HN request time: 0s | source
Show context
cjs_ac ◴[] No.45293534[source]
I'm a former physics teacher, and while I'm impressed by the technology, I think this is a low efficacy innovation.

The real challenge in teaching Newton's laws of motion to teenagers is that they struggle to deal with the idea that friction isn't always there. When students enter the classroom, they arrive with an understanding of motion that they've intuited from watching things move all their lives, and that understanding is the theory of impetus: https://en.wikipedia.org/wiki/Theory_of_impetus

An AI system that can interrogate individual students' understanding of the ideas presented and pose questions that challenge the theory of impetus would be really useful, because 'unteaching' impetus theory to thirty students at once is extremely difficult. However, what Google has presented here, with slides and multiple guess quizzes, is just a variation on the 'chalk and talk' theme.

The final straw that made me leave teaching was the head of languages telling me that a good teacher can teach any subject. Discussions about 'the best pedagogy' never make any consideration of what is being taught; there's an implicit assumption that every idea and subject should be taught the same way. School systems have improved markedly since they were introduced in the nineteenth century, but I think we've got everything we can out of the subject-agnostic approach to improvement, and we need to start engaging with the detail of what's being taught to further improve.

replies(9): >>45294003 #>>45294211 #>>45294245 #>>45294249 #>>45294416 #>>45294671 #>>45294721 #>>45297478 #>>45298395 #
sky2224 ◴[] No.45294245[source]
Yeah, as a student, I have to agree.

The issue with learning things isn't that it hasn't been tailored to be interesting or relatable to me, it's just that it's a lot of content and it's hard. The solution is figuring out how to set up a type of spoon feed algorithm that checks that I'm understanding little bite size pieces along the way in addition to giving layman's terms for things that don't necessitate the formal description (e.g., deciphering math language).

ChatGPT Study mode has actually been quite good at this when you prompt it correctly and are studying a subject that it's well trained on.

replies(1): >>45295089 #
aDyslecticCrow ◴[] No.45295089[source]
Khan academy and brilliant are both excellent. They're hand crafted and limited in subjects and depth, but i think establish the current "roof" in how perfectly structured self-learning materials would look. I've heard from teachers using them in schools and found excellent results.

AI rephrasing words better to each individual isn't interesting to me. Automatic Interactive small quizzes, puzzles, and self adjusting difficulty level would be amazing, but i don't see AI really reaching that level.

When i see AI "quiz me on this" it gets stuck asking direct factual question about the text. But a good question challenges assumptions, and prod deeper understanding.

replies(1): >>45295706 #
sky2224 ◴[] No.45295706[source]
Here's a conversation using ChatGPT Study Mode I had a little bit ago covering Linear Algebra concepts that I wanted to learn. The concepts that were gone over in this chat aren't the most complicated, but I think you might find it interesting to look over since I think it actually does show we'll likely reach the level you're seeking. This conversation is with 4o shortly before the GPT-5.0 rollout, which is why it's a little less concise and more emotive.

https://chatgpt.com/share/68cc844a-14d4-8009-88e3-53f5d781b5...

replies(1): >>45300448 #
1. aDyslecticCrow ◴[] No.45300448[source]
That is indeed better than ive seen before. I do still find it does do what i dislike. "Here is the forumla for cosine similarty, try compute it"

Id rather it approach it from "we would like to calculate difference in vector direction. Dot product is almost what we want, but cosine similarty is even better"

The "Angle connection" of cosine similarty was instead of added as an extra note later. But i think its fundamental to its intuition that excluding it from the main explanation could lead to a misunderstanding.

(Heck the definition section of Wikipedia makes the formula very clear, and the introductionsection is also excellent to descibe its utility)

So we may reach is one day. But i still think its a ways off compared to a proper hand crafted learning materials. And this subject specifically is a best case scenario.

(3b1b is amazing if you still want to get a more inutitive grasp on matrix transformatios)