[February 2025] Python Monthly Newsletter 💻🐍

Welcome to the 63rd concern of Python Month-to-month!

If it’s your first time right here, welcome, I such as you already. If you would like the total again story on this month-to-month e-newsletter, head here.

The fast model: I curate and share an important Python articles, information, assets, podcasts, and movies.

Assume the Pareto Precept (80/20 rule) assembly the Python world. I provide the 20% that can get you 80% of the outcomes.

For those who’re a very long time reader, welcome again outdated good friend.

Alright, let’s not waste any useful time and soar proper into this month’s updates.

This is what you missed in February 2025 as a Python Developer…

The Mutable Entice 🎣

Ever had a Python operate behave surprisingly, remembering values between calls when it shouldn’t? You’re not alone! That is considered one of Python’s sneakiest pitfalls—mutable default parameters.

Learn about them here.

Weekend Undertaking Thought 🧰

An ideal solution to spend your weekend: compile the original Python 1.0!

Python is now 31 years outdated, so it is time to look again in time and comply with this information and do the inconceivable: see what Python actually seemed like when the primary steady launch got here out.

CUDA Programming for Python Builders 🧀

NVIDIA’s CUDA (Compute Unified System Structure) is a platform and C++ extension that lets us write packages to run on the GPU.

CUDA offers a programming mannequin and APIs so builders can write code that runs instantly on the GPU, unlocking vital efficiency positive aspects by offloading parallelizable work from the CPU to the GPU.

In essence, you break your downside into many small items that may be solved on the similar time (like giving every GPU core a tiny job).

Be taught the way it all works utilizing Python in this great introductory guide.

PyPi Wheels 🛹

PyPi now supports iOS and Android wheels, making it simpler for Python builders to distribute cell packages.

Change My Thoughts 🍻

Software development topics I’ve changed my mind on after 10 years within the business… effectively not me precisely, however the writer of this publish.

I are likely to agree with most of those. It is vital to be versatile and keen to vary your opinions all through your profession.

Developer LLM Setup 🤹

I’ve been constructing so many small merchandise utilizing LLMs. It has been enjoyable, and helpful. Nonetheless, there are pitfalls that may waste a lot time. Some time again a good friend requested me how I used to be utilizing LLMs to write down software program. I assumed “oh boy. how a lot time do you’ve!” and thus this publish.

This is a great read and good insights even for those who do not use any AI instruments.

SQL Homicide 🥀

Wish to study SQL in a enjoyable manner? Extra enjoyable than the ZTM SQL Bootcamp?

No such factor.. however right here is the second most suitable choice: SQL Noire. Remedy a homicide thriller with SQL and have enjoyable.

TL;DR Of LLMs 🦥

This month, Andrej Karpathy (well-known AI researcher and educator) launched a video titled “Deep dive into LLMs like ChatGPT.” It’s a goldmine of knowledge, however it’s additionally 3 hours and 31 minutes lengthy.

This guy watched the whole thing and made a TL;DR version for anybody who desires the important takeaways with out the massive time dedication.

Developer Philosophy 🎍

A nice short read on some good developer ideas and philosophy from a senior. My favorite one is “There may be normally an easier solution to write it”.

Information Across the World 🗺

Massive Tech Information aka AI Information 🏢

• Final month DeepSeek was the massive story. To make the most of that HYPE, as a result of let’s be trustworthy, these corporations want to remain on high of the HYPE prepare, now you can discover the phrase “deep” in two new LLM merchandise:

1. Open AI launch Deep Research. It is the standard case of the product advertising being far more spectacular than the precise product, however some have stated this solves the “get the ten greatest blue hyperlinks” once you search one thing that Google was good at.
2. Perplexity launched, you guessed it, Deep Research too.

• Talking of DeepSeek, they’re on a mission to proceed open sourcing a bunch of their work. They launched 5 big open source projects this month.

• Google introduced the neighborhood launch of Gemini 2.0: their greatest mannequin but for coding efficiency and complicated prompts.

• OpenAI launched GPT 4.5. Massive warning indicators right here: It sounds prefer it’s so costly and the distinction in usefulness shouldn’t be a lot better. They even counsel they don’t seem to be going preserve serving it within the API for lengthy. Progress is flattening…

• Anthropic had a really large month. First they launch a report on how their AI instruments are getting used: turns out it’s mostly the software industry. Then they launched most likely one of the best coding AI on the market: Claude 3.7 Sonnet and Claude Code. If Claude Code actually will get going, it will probably change the way you code dramatically… it seems to be very spectacular.

• X introduced Grok3 and it guarantees to be the unfiltered model of AI.

• Microsoft had a giant Quantum Computing breakthrough this month: Majorana 1, a quantum chip utilizing topological superconductors and theoretical particles made actual. The eight-qubit chip goals to scale to 1 million qubits, doubtlessly creating computer systems extra highly effective than all present ones mixed.

Fully ineffective to your profession however nonetheless nice 🙃

Greatest Useful resource of the Month 🥽

That is an 18 lesson slide presentation by 2 professors. It asks an vital query about LLMs and AI: Are they modern day oracles or are they bullshit machines?

I promise that for those who take the time and undergo the 18 classes, you’ll come out the opposite finish higher knowledgeable about AI instruments and the way they work. This was my favorite useful resource of the month!

Do not let the title idiot you although. The teachings do an excellent job balancing out skepticism with practicality of utilizing these instruments.

This lesson is very vital:

AI chatbots are designed to be anthropoglossic: capable of converse, write, and converse in human-like vogue. Once we work together with anthropoglossic programs, we naturally assume they’ve the total vary human capabilities. They don’t.

In relation to bullshitting anybody about something, an LLM has an enormous benefit over any human.

Individuals use language in ways in which sign belonging. In a social state of affairs, one would possibly use explicit kind of slang; in a tutorial paper, one would possibly depend on sure types of jargon. However every of us has restricted expertise and experience. We solely belong to a couple social teams; we’re solely skilled in a number of domains. We do not have the insider data to talk within the codes of teams we do not belong to.

ChatGPT has entry to lots of the codes of many various teams. Consequently, it’s higher than a human outsider at mimicking the modes and patterns of speech, the dialects and slang and jargon of the teams which can be well-represented in its coaching set.

Therein lies its superhuman bullshitting skill. I’ve one perspective; it has been skilled on hundreds of thousands of views. I can’t go bullshit a bunch of radiologists at a radiologist conference, however ChatGPT presumably might. At the least it might persuade an adjoining group—surgeons, say—that it was an skilled radiologist.

Trick of the Month 🎩

• New Python IDE is in town.. it is a paid IDE although… bummer.

• Goose – Open Supply, and regionally run: your on-machine AI agent, automating engineering duties seamlessly.

• Cool idea in case you are on the lookout for a enjoyable undertaking. You’ll find the GitHub hyperlink in there so you may create one in your personal life: My Life in Weeks

• Wish to be a freelancer? Here is how to do it.

• Learn how to change your settings to make your self much less useful to Meta. Takes 2 minutes.

See you subsequent month everybody… additionally share this with your folks… fairly please! ❤️

By the best way, I educate folks code and get employed in probably the most environment friendly manner attainable as an Teacher on the Zero To Mastery Academy. You possibly can see a number of of our programs under or see all ZTM courses here.