Coming in November: How LLMs work, a 23-lesson video course Other

Our new video course, How LLMs work, comes out in November. It goes from the first artificial neuron to building your own LLM, in 23 animated lessons with hosts Anna and Max. Along the way, Akami builds his own chatbot, AkamiGPT, one piece in every lesson.

The course is part of the Has Brains subscription. The first lesson, the two-minute intro, is free for everyone.

What you will learn

• How it runs: the GPUs and data centers behind models, and the full journey of one prompt from the app to the GPU and back.
• How to use it well: the context window, good prompts, RAG, prompt security, and how to evaluate models.
• How to run it yourself: quantization, and local models on your laptop and phone.
• What is inside: the math, attention, the transformer, pretraining, fine-tuning with LoRA, RLHF and reasoning.
• Beyond text: how AI makes pictures, voices, music and video, and what may come after LLMs.
• Building for real: design patterns for AI agents, training and serving on thousands of chips, and MLOps.
• The finale: Akami builds his own LLM, step by step, with a runnable Google Colab kit.

The 23 lessons

Hands-on

Several lessons include practice you can follow along with:

• A small search engine built with RAG in Google Colab.
• Red-teaming your own bot with promptfoo.
• Running a model offline with Ollama, and on a phone.
• A real PyTorch training loop and a nanoGPT Shakespeare run.
• A LoRA fine-tuning script run in Google Colab.
• A first agent in Claude Code.
• The finale kit: LoRA on Qwen2.5, a 4-bit GGUF file and an Ollama model you can chat with.

Every lesson comes with a transcript.

We will post the exact release date here and in the Weekly AI newsletter.