[jessym]

Thanks!

I'm definitely going to have a look at some of the resources from your blog post.

With regards to the standard documentation: I notice that I'm able to read the sentences, but I'm finding it a bit difficult to really turn it into a working project I actually understand.

(For instance: the docs give an example of how bind to a socket address, but I'm not sure how to morph this into a simple echo server, for example, which writes back whatever I sent to it. Feels a bit like I might be missing some general information about TCP and how to test it, how it relates to HTTP, etc.)

[arthurcolle]

Totally not an OpenAI shill or anything, but based on your question and this specific type of QA stuff, here's my advice: spend the 20 bucks on ChatGPT Plus to get GPT-4 access - it is an invaluable learning tool for questions such as these.

For example, with your specific question: https://chat.openai.com/share/0d922193-b9b6-4d51-85e8-d31c05...

Zero-shot correct solution to the example query. Based on your question, you could even just say "Can you explain X based on Y info that I found in the docs without writing a solution" if you truly just want to learn about the topic. Then with some iterative refinement of the results over a few back and forths, you can learn about the concept without necessarily having it write the code for you. Cheers!

[jessym]

This is pretty good!

I've been trying to have the free GPT-3.5 explain me things, but the results were a bit mixed (I caught it making a few things up earlier). But I might give GPT-4 a try, thanks!

[arthurcolle]

GPT-4 is significantly better than GPT-3.5-turbo. Do not use GPT-3.5-turbo outputs as your benchmark for the utility of the more advanced models! :) Especially when used for LLM Agents, or any kind of harder question.

[jessym]

Got it, thank you so much!