4 ms·

The main question IMO is this: > when I first used these lookup books, I had a shaky understanding of logarithms. this implies that now you do have an underst

The main question IMO is this: > when I first used these lookup books, I had a shaky understanding of logarithms. this implies that now you do have an understanding of logarithms. how did you acquire this understanding? The fear that many here share is that LLMs will make too many people get by without learning anything substantial from the ground up. Take away their LLM access and they're completely useless - at least that's what people are afraid of. Myself I'm on the fence - it was always the case that a certain type of people seek to understand how thinks actually work, and others who just want to apply recipies. These trends will likely make the recipy-appliers at first more powerful. I'm afraid though that soon enough they will be driven out by further automation coming on top.

Sai_ 3y ago
> this implies that now you do have an understanding of logarithms. how did you acquire this understanding? Not knowing logarithms (a) didn't prevent me from successfully using the log tables books and (b) using the log tables books didn't prevent me from developing a better understanding of logarithms. I expect something similar with happen with GPT based education. I acquired a deeper understanding of logarithms through a bit of reflection and thought. The math lessons and teachers didn't help illuminate logarithms in any significant way. They simply rattled off the theory of logarithms and moved on to teaching us how to use them. I had to go out of my way to understand it myself. Is this any different from a GPT giving you a superficial, regurgitated level of knowledge which you then have to deepen through some effort on your part?
- m_mueller 3y ago
  Yes, agreed. again though, the question is what jobs will be left in a couple of years for people besides people like you who went and sought out the foundational knowledge to be able to actually check the work the LLMs are going to do. And from my experience it needs a lot of checking.
- saint_fiasco 3y ago
  It's not any different, but in my opinion both are not good. Someone should teach you how logarithms work before giving you a log table. Someone at some point should show you a trigonometric circle before telling you to press the cos button in the calculator to calculate a cosine. That said, some"one" could be a GPT chatbot. The problem is that you might not know the right question to ask.

Sai_ 3y ago

> this implies that now you do have an understanding of logarithms. how did you acquire this understanding? Not knowing logarithms (a) didn't prevent me from successfully using the log tables books and (b) using the log tables books didn't prevent me from developing a better understanding of logarithms. I expect something similar with happen with GPT based education. I acquired a deeper understanding of logarithms through a bit of reflection and thought. The math lessons and teachers didn't help illuminate logarithms in any significant way. They simply rattled off the theory of logarithms and moved on to teaching us how to use them. I had to go out of my way to understand it myself. Is this any different from a GPT giving you a superficial, regurgitated level of knowledge which you then have to deepen through some effort on your part?

m_mueller 3y ago
Yes, agreed. again though, the question is what jobs will be left in a couple of years for people besides people like you who went and sought out the foundational knowledge to be able to actually check the work the LLMs are going to do. And from my experience it needs a lot of checking.
saint_fiasco 3y ago
It's not any different, but in my opinion both are not good. Someone should teach you how logarithms work before giving you a log table. Someone at some point should show you a trigonometric circle before telling you to press the cos button in the calculator to calculate a cosine. That said, some"one" could be a GPT chatbot. The problem is that you might not know the right question to ask.