Lessons
Video lessons with interactive transcripts
Each lesson pairs a video with a clickable transcript and chapter markers. Click any line in the transcript to jump the video to that exact moment.
Use the chapter list to skim what’s covered. Use the transcript to find the exact sentence you want to revisit — clicking a line seeks the video. The active line highlights as the video plays.
Lessons follow the Roadmap — grouped by step below.
Step 1 First use of AI for coding
Set up your cohort tooling and get oriented. Start with the welcome video before diving into model inference.
Lesson 0 · Step 1
Welcome to First Break AI
A guided tour of the cohort — the six stages from commit to capstone, who this cohort is for, the roadmap walkthrough, the checklist, and the AI setup. Watch with an interactive transcript and chapter markers.
Step 2 Run a model locally
Run Qwen3 0.6B on your laptop and trace every operation from tokenization to token output — HuggingFace as a supply chain, then GGUF, memory mapping, and pure C inference.
Lesson 1 · Step 2
HuggingFace Beyond Upload
Why an open model is not a file — it’s a supply chain you can inspect. The mental shift from from_pretrained to provenance, datasets, security (trust_remote_code), evals, runtime routing, and agent-readable manifests. Builds on the May 8 office hours.
Lesson 1b · Step 2
Qwen3 Fundamentals
How a language model lives on disk and loads into memory — byte by byte. Part A video is live (run Qwen3 locally in pure C); Parts B and C coming soon. Written lesson covers GGUF, memory mapping, SafeTensors, and the conversion pipeline.
Step 3 Inference deep dive
Beyond running a model — inference engines, OpenAI-compatible serving, and wiring a local coding agent. First lesson: llama.cpp server with Qwen2.5 Coder and Cline.
Lesson 2a · Step 3
Run a Coding AI on Your Laptop
Serve Qwen2.5 Coder with llama.cpp server, connect Cline via OpenAI-compatible API — free, private, no API key. First Step 3 lesson.
More lessons will be published here as the cohort progresses.