Build AI product sense
one run at a time
Interactive lessons where you run real models, step inside them in 3D, and push them until they break. Then take on the challenges and climb the leaderboards.
Step inside the model
Immersive, full-screen journeys through what happens inside an AI model
Inside an Embedding Model
Watch a search become tokens, then 1,024 numbers, then a point in meaning space, and find its nearest neighbors.
Inside Next-Token Prediction
Watch a model write: every candidate for the next token, the roll of the dice, and what temperature does to the odds.
Inside an Agent Loop
Ride along as an agent re-reads everything, calls tools, and goes around again, and see what each lap costs.
Inside RAG
Watch a question find its nearest docs, see them pasted into the prompt, and trace each citation back to its source.
Inside an Eval
Watch 8 tickets graded three ways, read a failure, and see a bigger model score worse.
Inside a Prompt Injection
Watch a hidden instruction in an email hijack an inbox agent, then try the defenses and see which ones stop it.
Challenges
Beat the goal on real model runs, or call what a model did, then share your best
Get all 12 requests right at the lowest cost per 1,000 requests.
Predict what real models did in ten recorded runs, then see if you called it.
Review ten real model outputs: ship it, or flag what went wrong.
Foundations
How models behave: tokens, prompts, cost, and when they make things up
Building
Putting models to work in a product: tools, retrieval, agents
Shipping
Making it measurable, safe and affordable
Bring AI PM Lab into your assistant
Connect it to Claude, ChatGPT, Cursor or VS Code, and your assistant can review AI specs, price features and design eval suites with what the lessons teach.
- “Review this AI feature spec and tell me what's missing.”
- “What would 50,000 requests a day cost on gpt-4.1-mini?”
- “Design an eval suite for our support bot.”