GPT makes learning fun again • Hacker News Board

Re: GPT makes learning fun again

#201

Post by thealig » Tue, May 02, 2023, 3:22 PM UTC

alex-moon wrote: post #134

So many comments on this shovelling sh*t on OP...? I agree with the author entirely. In practically every field, the reason it's difficult to learn anything, starting with zero, is because you don't know what questions to ask. You need to find an in. ChatGPT is really good at this - the answers it gives provide an idea of what you don't know you don't know. Obviously, you then need to go verify the answers, but the d…

Finding out what you don't know to start learning is also a part of learning, and it is a bit difficult to trust LLMs like ChatGPT when it had a tendency to spout correct sounding answers that are incorrect, inaccurate etc.

Search engines are much better On uncovering guides on that, specially from experts and verified sources. It's a bit of work to verify that, but then it is part of learning itself, not sure trying to punt on that to fast track your learning your whatever is going to make a meaningful difference in terms of time

Re: GPT makes learning fun again

#202

Post by simonw » Tue, May 02, 2023, 3:22 PM UTC

notamy wrote: post #171

lcuff wrote: post #25

Earlier quoted context omitted.

I'm curious about what topics you have knowledge of where it's failed. Does it seem like there's a pattern to the failures? I've been using GPT for coding help, and it it is very helpful in ruby and bash, though it often delivers buggy software: Badly handled non-happy-path conditions, mostly, which when I tell it to handle the case, it may. It's a huge help for me finding gems and showing me standard ruby library sy…

> Does it seem like there's a pattern to the failures? Personally, trying to use it to write code (primarily Elixir backend and Rust systems/CLI), I tend to run into: - Hallucinating APIs that don't exist - Hallucinating entire libraries, despite being told repeatedly they don't exist - Saying it will make requested changes and not doing so - Not being anywhere close to idiomatic code - Not being able to explain code…

Yeah, GPT-4 is so, so much better at not doing the things you're describing there.

ChatGPT is also much better for Python and JavaScript than Rust and Elixir, presumably because it's seen an order of magnitude more example code for those languages.

Re: GPT makes learning fun again

#203

Post by sebastiennight » Tue, May 02, 2023, 3:29 PM UTC

mrrobot900 wrote: post #184

im reading the wealth of nations right now, using gpt on the side to help clarify long passages and decode archaic terms. it's been tremendously helpful. some actual use cases: - summarising entire chapters (summarise chapter 8 of book 1) - asking to verify my own summaries (in summary, smith writes that ... am i correct?) - explaining certain passages in ever more detail (explain ...; explain in simpler terms; im st…

Funny that the one comment in this thread that is talking about summaries is the longest on the entire page!

Re: GPT makes learning fun again

#204

Post by navane » Tue, May 02, 2023, 3:32 PM UTC

input_sh wrote: post #109

rolisz wrote: post #87

Earlier quoted context omitted.

If you're not DevOps, why bother learning whatever framework is used today, but which will be replaced by something else next year? I just want to get my stuff deployed and that's it.

Saying that DevOps tooling changes every year is absolute nonsense. The tooling is fairly standardised nowadays (Docker, Kubernetes, Ansible, Terraform, Prometheus, CI/CD tool of your choice) and every part of it is over half a decade old. ChatGPT will only help you with the simplest of tasks, and even then you're gonna fail if you don't know to correct it. If you go beyond the most basic of the tasks you can learn h…

over half a decade old! haha, so assuming you don't use it the day it comes out, but after a couple of years, most people have only used it a couple of years. Unless you've only worked a couple of years, you've gone trough a couple of these.

Re: GPT makes learning fun again

#205

Post by thethimble » Tue, May 02, 2023, 3:51 PM UTC

krychu wrote: post #159

I've been using GPT to have (insightful) educational conversations about Quake 1 source code: https://twitter.com/krychusamp/status/1649048047996014595 I always finish up by asking GPT to test my knowledge with a single-choice questionnaire. What I've observed is that the retention of the material is higher compared to "traditional" techniques. Perhaps the conversation style is more immersive, or perhaps focusing on…

Agreed - it seems that calling out LLM accuracy is a meme here - hyperbolically: “because LLMs can be inaccurate they are useless”.

There seems to be less next-level analysis: which topics are more prone to inaccuracy, does the critique loop actually help LLMs overcome those inaccuracies, and do the benefits of LLMs outweigh the consequences of these inaccuracies?

Re: GPT makes learning fun again

#206

Post by JohnFen » Tue, May 02, 2023, 3:51 PM UTC

Here's the aspect of this article that I found the most fascinating, and might explain why I don't get the same level of benefit from ChatGPT as many report.

If I were learning how LEDs work, I would not have wasted any time whatsoever on the search results that the author spent a lot of time on. They were obviously (to me) the wrong articles on the face of it, because they were covering aspects that weren't really what I was looking for (the wrong sort of detail and emphasis).

So I think I would have been off and running pretty much immediately with the web search results rather than spending time on the clear dead ends.

ChatGPT gets me there too, after enough back-and-forth, but it takes longer for me to zero in on what I'm looking for.

I say this not to say that ChatGPT is in any way bad for this. I'm just noticing a difference in how the two of us engage in learning new topics. Perhaps the reality is that for some people, ChatGPT is a godsend, and for others, it's fine... but hardly an improvement for this use use case.

It would explain a lot.

(Also, when did learning stop being fun??)

Re: GPT makes learning fun again

#207

Post by JohnFen » Tue, May 02, 2023, 3:57 PM UTC

throwaway2990 wrote: post #151

devjab wrote: post #7

I'm in two minds about it. On one hand the internet (in general) has become so hard to find information on, and I agree with the author that GPT is a breath of fresh air. On the other hand I've seen GPT fail so miserably at topics that I'm knowledge about that I have a very hard time trusting anything it tells me. I'm not sure what the answer really is, but I'm not sure it's GPT. I wish we could go back to having sea…

[flagged]

Maybe you're downvoted for calling a group of people liars rather than for liking AI.

Re: GPT makes learning fun again

#208

Post by namaria » Tue, May 02, 2023, 3:59 PM UTC

kerkeslager wrote: post #57

grrdotcloud wrote: post #4

Imagine being able to ask questions and get answers back based upon understanding, and not upon a curriculum or agenda. I have found this method to allow me to consume information much faster while skipping over the often tedious topics.

Do you think that "tediousness" is a good indicator of whether a topic is important?

Shhhh, the more Duning-Kruger, the more valuable true competence becomes.

Re: GPT makes learning fun again

#209

Post by cyrialize » Tue, May 02, 2023, 4:03 PM UTC

I've been using Phind [0] for specifically dev related stuff, and it's been great. I very much like that it gives me all the resources it uses.

On the other hand, I feel like my experience with Phind has been so great because with all dev stuff I ask it it's always been something I could try out right away.

With non-dev stuff that isn't always the case. I'd love an LLM that can do the same thing as Phind, as in providing the sources it's using, while also working for everything.

[0]: https://www.phind.com/

Re: GPT makes learning fun again

#210

Post by MattGrommes » Tue, May 02, 2023, 4:14 PM UTC

One solution to his problem with the Wikipedia being hard to learn from is to use the criminally under-known Simple Wikipedia (https://simple.wikipedia.org/wiki/Main_Page). Just replace 'en' in a wikipedia URL with 'simple' and hopefully there will be a nice and simple version of the page. I don't see it advertised very often but I think it should be.