Module 02 · AIFrom a model to an app
Watch: The problems it solved, and the one it did not quite →▶ Watch the video
“Agents and Tool Use”, the board that closes the How LLMs Work track. It draws the loop that every coding agent is running, including whichever one you are reading this in.
Before you start. You need one fact from Inference (sampling, context, KV cache, hallucination): a language model is a text predictor. You hand it a string, it hands back the next stretch of text, and it has no memory, no network and no hands. Everything here is built out of that limitation.
Continue reading with a plan
The first three lessons of every track are free. This lesson is included with a plan, which unlocks every lesson, board, video, lab and paper for one price.
The problems it solved, and the one it did not quite · also What it brings, what it broke, what changes for you, Agents from first principles: build the loop yourself
Related paper
Connects to
Saved in this browser, and to your account when you sign in.