Full Course · Community Access

Open-Source AI: The Ultimate Guide

7 modules. 30 lessons. Understand open-source AI models — what they are, how they work, when to use them, and how to run them yourself. No engineering background needed. Just practical literacy for professionals who want to make informed decisions.

Not Hype. Practical Understanding.

Most open-source AI content is either too technical or too shallow. This course gives you the knowledge to evaluate models, understand trade-offs, and make real decisions — without becoming a machine learning engineer.

  • Privacy, Independence, Cost The real reasons open models matter — and when they don't. An honest assessment, not a sales pitch.
  • How Models Actually Work Dense vs sparse, quantization, context windows, thinking models — explained through analogies, not equations.
  • Read Benchmarks Without Getting Fooled How benchmarks lie, what they actually measure, and how to evaluate models based on your real work.
  • Run Models on Your Own Machine What hardware you need, what software to use, and how to try models for $5 before investing in anything.
  • A Decision Framework That Lasts Models change monthly. The skill to evaluate them doesn't. Learn the framework, not the leaderboard.

7 Modules. Zero Jargon Required.

Each module builds your understanding step by step. You start with why open models matter and finish with a personal strategy for choosing and running them. Every module ends with a hands-on exercise.

01

The Open Source AI Landscape

Why open-source models matter. Who makes them and why. Why the core technology is more stable than headlines suggest.

  • Why Open-Source Models Matter
  • Who Makes Open Models and Why
  • The Car Engine That Hasn't Changed
  • Exercise: Map Your AI Usage
02

How Models Are Built

Architecture decisions explained through analogies. Dense vs sparse, quantization as MP3 compression, context windows, and thinking models.

  • Dense vs Sparse: The Big Architecture Choice
  • Quantization: The MP3 of AI
  • Context Windows: The Desk Analogy
  • Thinking Models: Speed vs Intelligence
  • Exercise: Read a Model Card
03

How Models Are Trained

The three stages of training. Why some models outperform others. What “open” actually means — and what you really get.

  • The Three Stages of Training
  • Why Some Models Are Better Than Others
  • Open vs Closed: What You Actually Get
  • Exercise: The Cost Comparison
04

Understanding Benchmarks

What benchmarks measure, how they lie, and how to actually evaluate a model for your use case.

  • What Benchmarks Are and Why They Exist
  • How Benchmarks Lie
  • How to Actually Evaluate Models
  • Exercise: Benchmark Detective
05

Running Models Yourself

Try before you buy. Software platforms for local models. Hardware reality check — what you actually need and what you don't.

  • Try Before You Buy
  • Software to Run Local Models
  • Hardware Reality Check
  • Exercise: Try an Open Model
06

Choosing The Right Model

The decision framework. Small models for local use. Large models through APIs. Our curated picks for both.

  • The Decision Framework
  • Small Models: Run Locally
  • Large Open Models: API Alternatives
  • Exercise: Find Your Model
  • Our Picks: Models for Your Own Computer
  • Our Picks: Models Through Online Services
07

The Fast Moving Landscape

How to stay current without drowning. What’s coming next. Build a learning system that outlasts any model release.

  • How to Stay Current
  • What Is Coming Next
  • Exercise: Your Learning System
7 modules · 30 lessons · Exercises in every module

Practical Literacy, Not Engineering Degree

What you get here

  • Concepts explained through analogies, not equations
  • A decision framework that survives model releases
  • Honest trade-offs — not “open source is always better”
  • Hands-on exercises in every module
  • Curated model picks, regularly updated
  • Community of professionals making the same decisions

What you find elsewhere

  • Technical papers written for researchers
  • Hot takes that expire in two weeks
  • Hype without honest downsides
  • Theory with no practice
  • Outdated model recommendations
  • Reddit threads with contradicting advice

Is This Course Right for You?

Built for you if:

  • You use AI daily and want to understand what you're using
  • You hear “quantization” and “MoE” and want to know what they mean
  • You want to evaluate models yourself instead of following hype
  • You’re considering running AI locally for privacy or cost
  • You make technology decisions for yourself or your team

Not designed for:

  • ML engineers who train models — this is user-level, not research-level
  • People who want a single “best model” answer — the course teaches frameworks, not hot takes
  • Anyone looking for a coding bootcamp — no programming required or taught

Your Instructor

S

Sjoerd Tiemensma

Freelance developer and AI automation specialist. Writes the “Use AI” newsletter (2,000+ subscribers), certified Make.com Expert, and builds AI-powered systems for clients daily. Sjoerd turns complex AI concepts into practical, accessible explanations — and designed this course to do the same for you.

Common Questions

Do I need a technical background?

No. The course is designed for professionals who use AI, not build it. Every concept is explained through analogies and plain language. If you can install an app, you can follow this course.

Do I need special hardware to follow along?

No. Module 5 shows you how to try open models through a browser for a few dollars before investing in anything. Local hardware is optional — the course teaches you when it makes sense and when it doesn't.

Won't this be outdated in a few months?

Specific model names change. The frameworks for evaluating them don't. The course teaches you how to think about models, read model cards, evaluate benchmarks, and make decisions — skills that survive any release cycle. We also keep the curated picks updated.

How is this different from watching YouTube explainers?

YouTube gives you 10-minute overviews of single topics. This course builds a complete understanding, step by step, with exercises that make you apply what you learn to your actual work. Plus direct access to the instructor and community.

I already run local models. Is this still useful?

If you already run Ollama or LM Studio, modules 1-2 might be review. But the benchmarks deep-dive, the decision framework, and the curated picks in modules 4-6 go beyond surface-level knowledge. Most experienced users find blind spots they didn't know they had.

Stop Guessing. Start Understanding.

Join Autonomee for the full course, direct support, and a community of professionals navigating the open-source AI landscape together.

Join Autonomee