Python Monthly Newsletter 💻🐍

Welcome to the 55th issue of Python Monthly!

If it’s your first time here, welcome, I like you already. If you want the full back story on this monthly newsletter, head here.

The quick version: I curate and share the most important Python articles, news, resources, podcasts, and videos.

Think the Pareto Principle (80/20 rule) meeting the Python world. I give you the 20% that will get you 80% of the results.

If you're a long time reader, welcome back old friend.

Alright, let's not waste any valuable time and jump right into this month's updates.

Here's what you missed in June 2024 as a Python Developer…

Where is William!? Python Edition 🧔🏼‍♂️

This may be the most fun way for you to learn about hashable objects in Python. But seriously, "WHERE'S WILLLIAAAM?".

PyCon US 2024 Recap 🐍

A bunch of Pythonistas, Pydudes and Pydudettes got together for the annual conference. What did they talk about? What did you miss out on?

Here is the recap of the big week long event to keep you up to date with the latest.

Python CLI Time 🖍

Did you know that some Python modules can double-up as handy command-line tools? For example, you can run Python's webbrowser module from the command-line to open up a given URL in your default web browser:

$ python -m webbrowser https://pym.dev/p
Opening in existing browser session.

Learn more of Python's many command-line utilities you can use like the one above.

NVIDIA Loves Python ❤️‍🔥

NVIDIA (now the most valuable company in the world) loves GPUs and loves Python. So they released a new library: Warp.

It's a Python framework for writing high-performance simulation and graphics code. Warp takes regular Python functions and JIT compiles them to efficient kernel code that can run on the CPU or GPU.

Best Discussion of the Month 🤳

A great discussion around the topic: Why do message queue-based architectures seem less popular now?

It turns into revealing that "just because Big Tech/Google does it, doesn't mean all startup/projects should do it". I highly recommend reading some of the top comments.

An Alternative AI Universe 👽

Here is an alternate perspective on AI and how useful it is to companies. As with all things in life, it's never black and white. This article, however you may receive it, is hilarious though...

News Around the World 🗺

• Top engineers (and co-founder) of OpenAI have left to create this company for the sake of humanity and AI safety.

• Julian Assange has reached a plea deal with the U.S., allowing him to go free after many years. This story needs to be made into a movie.

• Adobe got sued for making cancellations really hard for their product. Scummy behaviour is not cool.

Big Tech News 🏢

Apple's AI Day was the big talk this month especially because of what it means for the industry moving forward and Apple's plans for generative AI: ‘Apple Intelligence’.

I personally think that as usual, Apple is right on the money on how to use the power of A.I.. My favourite part was their ideas around AI privacy for individuals.

You can read the full article on the Apple AI strategy here which is super fascinating, or read the summary below.

Benedict Evans (a VC) summarized the whole event nicely:

1. A LLM (SLLM?) on the device, with variants tuned for different capabilities, will power a wide range of new features within existing Apple apps, and available to third party apps. Tasks that need more compute will be passed to a full-sized Apple foundation model running in an Apple cloud on custom Apple Silicon (sorry Nvidia) with a new and sophisticated security architecture meaning that Apple itself never sees the data. Apple provided benchmarks claiming that both models have comparable quality to competing on-device and large models from competitors (sorry OpenAI).
2. Most of these features are about suggestion, image generation, autocorrect and complete etc - ‘rewrite this mail to be more friendly’. As I’ve said, I think GenAI is a tech that can enable new features much more than it is a product in itself. Meanwhile there are no open-ended prompts, and there’s an image generator but it doesn’t do photorealism - Apple is trying to close off the obvious paths for abuse or misunderstanding.
3. Apple is also very focused on personal context. LLMs mean that Siri will now actually (maybe) work - you can ask ‘how long will it take to get to the restaurant my mother mentioned the other day?’ There’s a (secure, private) index of your activity and content on the device, an expanded system of ‘intents’ (with an API) to track which apps can do what, and a brokerage system to work out which apps to combine for your request (you could call this an agent, if you like). Obviously, a cloud LLM like ChatGPT has none of this context today, Google Search and Meta have bits of it, and arguably only the device (iOS/Android/Windows) has this holistic view of your context: so far only Apple has really articulated this model.
4. An OpenAI partnership: for tasks that Apple describes as needing a ‘world’ model - ‘suggest recipes for this photo of my groceries’ - your device will suggest sending the query to ChatGPT, for free, with no account needed and no data retained. This seems to have confused some people: to make it clear, none of the other features use ChatGPT at all.
5. All of this needs a lot of local compute and RAM and so only the iPhone 15 Pro (last year’s model) and M1 and later Macs and iPads (the last few years) will get these features. You can argue whether this will drive a refresh cycle, but the underlying hardware justification is plausible.

• Not to get outdone by Apple, Meta is working on its own AI strategy: Create models and open source them to the world. Here is how they train their models at scale.

• Microsoft on the other hand had an AI stumble: Microsoft postpones Windows Recall after major backlash — will launch Copilot+ PCs without headlining AI feature.

• Google, although they introduced the initial paper that propelled this boom in LLMs and AI tools, is now a little behind. They are shifting strategies: Google DeepMind Shifts From Research Lab to AI Product Factory... because they need to make money off this thing. The company combined its two AI labs to develop commercial services, a move that could undermine its long-running strength in foundational research.

• U.S. Clears Way for Antitrust Inquiries of Nvidia, Microsoft and OpenAI... makes sense considering last month Nvidia became the most valuable company in the world.

Completely useless to your career but still great 🙃

• Try drawing the most ridiculous iceberg and see what happens.

• The most insane 0.35 seconds Rubiks Cube solver. Human or robot? Can you guess before watching this video?

• Hacking Millions of Modems (and Investigating Who Hacked My Modem). This is a great read and a great demonstration of problem solving.

• How Online Privacy Is Like Fishing - a pretty fresh read.

• Make some beats with this cool little website.

• The most epic game of Pong you will ever play.

• Did you know that we don't know exactly how water freezes?

Best Resource of the Month 🥽

When Google gives their opinions on software best practices, usually the industry listens. This time they share their thoughts on how software engineers should use AI tools.

I think they are right on the money with this one. If you work in software, you need to read this one.

Trick of the Month 🎩

The most useless but fun thing you have ever had on your desktop while working. You're welcome from making you procrastinate even more at your job.

Thanks for reading!

See you next month everyone... also share this with your friends... pretty please! ❤️

By the way, I teach people how to code and get hired in the most efficient way possible as an Instructor at the Zero To Mastery Academy. You can see a few of our courses below or see all ZTM courses here.