Ask HN: How to learn Python in 2024? | Hacker News Reader

Ask HN: How to learn Python in 2024?

20 pointsnomilk17 comments

A young relative asked me and I did my best but my recommendations were probably out of date (codecademy, derek banas youtube channel, andrew ng ML course). She's interested in automations and tinkering with gpt and whisper APIs.

What's the best way for someone completely new to tech to learn some python in 2024?

Also interested in

- recommended installation method,

- IDE,

- REPL etc

(I'd recommend vscode but for lack of REPL, so will recommend pycharm unless there's a better option for a complete beginner)

View on HN

There are plenty of free resources:

* Python Programming by University of Helsinki (https://programming-24.mooc.fi/)

* Harvard CS50's Introduction to Programming with Python (https://cs50.harvard.edu/python/2022/)

* futurecoder (https://futurecoder.io/) — interactive, includes integrated debuggers, enhanced tracebacks, hints for exercises and more

* Think Python (https://allendowney.github.io/ThinkPython/) — gives you a solid foundation to programming, teaches debugging right the beginning, interesting exercises, etc

* The Python Coding Book (https://thepythoncodingbook.com/) — friendly, relaxed programming book for beginners

* PyFlo (https://pyflo.net/) — interactive beginners guide to becoming a Python programmer

See my resources list https://learnbyexample.github.io/py_resources/ for many more links

There are a lot of people coming to Python now, even among us who already code in other languages. The first thing I noticed was a confusing variety of ways to install Python, plus a lot of necessary-but-confusing tools for version management, package management, and environment management. I wrote a guide for my mac.install.guide website about installing Python [0] and then condensed it for freeCodeCamp [1].

In a nutshell, I recommend installing Python with Rye [2], an all-in-one project management tool for Python, written in Rust (for speed) and inspired by Cargo, borrowing a folder-based approach to development from other languages such as JavaScript and Ruby. Other people like Pyenv/Venv/Pip but I think it's easier to use a single all-in-one tool.

[0] "Install Python" (https://mac.install.guide/python/install)

[1] "How to Install Python on a Mac" (https://www.freecodecamp.org/news/how-to-install-python-on-a...)

[2] "Rye" (https://rye-up.com/)

Is rye stable enough to recommend to newcomers?

I'm a huge believer in running things on your own machine, but, for a true novice, starting in something like replit may be a better choice.

Rye has moved to https://rye.astral.sh/

rye's website is down.

They moved it to https://rye.astral.sh/

Python is just the tool, what you want to learn are the concepts and techniques for solving problems with computers.

Check out CS50x or CS50p

(python version)

https://cs50.harvard.edu/python/2022/

Corey Schafer has one of the best Python playlists out there. His channel at YouTube. I don't believe any more than his channel is needed, other than Python docs.

Agreed. This is the playlist I recommend for beginners:

https://www.youtube.com/watch?app=desktop&v=YYXdXT2l-Gg&list...

I would recommend [1].

It is challenging for a beginner, at the same time I remember this course (at least many years ago) having an outstanding community. No question without answer spirit.

They have detailed instructions on how to install Anaconda.

Once done with this training your young relative won't have any issue registering and creating an account with any given AI provider and making API calls.

deeplearning.ai happens to have course very beginner friendly. But in My opinion a person should focus on having the aforementioned course (or CS50 [Also on edx]) under its belt.

[1] https://www.edx.org/learn/computer-science/massachusetts-ins...

Strongly recommend NOT using Anaconda, ever, for anything.

Use Poetry, if you have to, use pip/venv.

When it comes time to managing your actual production code, if you learned in anaconda you will have no idea what you are doing. Pick your dependencies purposefully and carefully, not relying on whatever Conda tells you to use.

Managing deps in Python is awful and Poetry makes it suck less.

hmm, I did not have the same experience. Anaconda makes managing python versions and environments extremely easy, and you can still use pip inside it if you want. But I have installed countless ML packages from both conda, pip-only and local setup only repoes, haven't had any issues.

I've found this one excellent - it's by David Beazley:

https://dabeaz-course.github.io/practical-python/Notes/Conte...

May be a bit too quick jump for a complete beginner to tech, but it's definitely worth it.

https://inventwithpython.com/pygame/

I started with python’s own IDE and a few practical books. It was enough to get me started on programming. A beginner book and a python installation is all that’s required. An interesting (to her) project may be needed for motivation.

> I'd recommend vscode but for lack of REPL

vscode has a built-in terminal?

But I'd separate concerns here and have her focus on writing and running basic programs to start. Choose one of the tutorial/video recs from other comments and get her set up in Replit.

Let her work through the complexity of local setup _after_ she knows the basics and wants to do more.

Install VScode, optimize the environment (10^10 article, google, or best are on medium), pick a topic/app you are interested so you are more likely to keep investing, and get a sub to ChatGPT. Start with "I want to build a python3 application that does {insert high-level description}. Please wait for me to describe it in detail. Do you understand". It keeps you interested combine it with more medium articles that interest you and ask ChatGPT when you don't understand. Ultimately it's try through creativity. If do not know what to build, tap into python projects on GITHub. Step and repeat. Never look back. If you want to learn other things like html, css, do a web app with flask or stream-lit...you will learn that too...

Can recommend. This is how I did it over the last year. VSCode, Python, Flask, ChatGPT, and pick something challenging to build. Keep going for a year.

Reply on news.ycombinator.com