LLMs on Mac: The school to run them from scratch on your Apple
·2 min read·Intermediate
“
Got a Mac and wonder how AI models can run without an internet connection? Someone finally built a complete guide to understand and do it yourself.
In 30 seconds
01Alvaro Videla launched Inference School, a course to run AI models (LLMs) on Apple silicon chips.
02The project features 48 hands-on lessons, exercises, and a dedicated macOS Studio app.
→
💡
What this means for you
For the average person, this means anyone with a Mac and a bit of curiosity can learn to run AI models directly on their computer, without relying on external services or the cloud.
It feels like a new AI concept pops up every single week. Amidst all the hype, there's one thing every language model has, but few people truly grasp.
·2 min·1·Intermediate
03It teaches local AI inference, eliminating reliance on external cloud services.
0101
Who wants to run AI on their own Mac?
While everyone downloads AI apps that perform magic with a click, some of us wonder how they actually work. More than that, some want to get their hands dirty, understanding how a language model, like the ones answering your questions, can run directly on a Mac. Yes, without sending your data all over the internet.
This is somewhat the dream for those seeking autonomy or simply curiosity. Instead of using online services, the idea is to make your computer's chip do the work. And by chip, we mean specifically the ones Apple puts in its Macs, from the M1 onwards. A nice advantage, wouldn't you say?
0202
How does this "school" for Apple chips work?
Alvaro Videla, a well-known developer, created " School" on GitHub to teach inference on Apple silicon. It's a complete, structured course, not just a bunch of scattered code. The goal is to help you grasp the "first principles," meaning the basics of how AI models process information on your hardware.
📬 Enjoying this article?
Get the best AI news every week, straight to your inbox.
The package is quite substantial: it includes 48 guided lessons, complete with runnable exercises. There's also a native macOS "Studio" app, which likely helps you visualize and manage your experiments. For the diligent students, a companion book is also provided. All designed to turn you into a small local AI wizard.
0303
Why is it important to learn from scratch?
Learning to run AI from scratch on your Mac means not just understanding, but also gaining control. You don't depend on external servers or costly subscriptions to experiment. You can test, modify, and understand the limits of models directly on your machine, using Swift and Metal, Apple's languages and technologies for computational power.
This "hands-on" approach is crucial for anyone who wants to go beyond passive AI use. Whether you're a developer or just a curious enthusiast, it gives you the tools to explore a complex world. Plus, it's a great way to squeeze every last drop out of your Mac's M-series processor, using it for more than just your everyday apps.
Bob Starr had built a cute little app in just a few hours: follow the vibe, let intuition guide you, and the code flows out naturally. What he didn't know was that his masterpiece was quietly harboring a security hole that could've unlocked everything to anyone.
