Rapid Fire
Rapid Fire
Compression, not learning. Use this after you understand the material, not instead of understanding it.
Prerequisites
Ideally the corresponding full section. These decks are retrieval practice — they strengthen recall of things you already understand and quietly fail you on things you don't.
How to use these decks
Every entry is a question with a three-line answer. The point is not to read them.
- Cover the answer. Read only the question.
- Say your answer out loud, in full sentences, as if explaining it to someone.
- Uncover and compare. Not "did I know it?" — did I say it as cleanly as that?
- Mark the misses. Anything you couldn't say fluently goes on tonight's reread list, and its full page is one click away.
Common misconception
Recognizing an answer feels almost identical to being able to produce one, and the two come apart the moment you have to actually explain something. Reading these decks silently will make you feel fluent without making you fluent. Say them out loud — that's the whole mechanism.
The decks
| Deck | Full treatment |
|---|---|
| LLM Fundamentals | Section 01 |
| Embeddings & Search | Section 02 |
| Vector Databases | Section 03 |
| RAG | Section 04 |
| Agents | Section 05 |
| MCP | Section 06 |
| Production | Section 07 |
| Model Landscape | Section 08 |
| Multimodal | Section 09 |
| Tradeoff One-Liners | — the 30-second versions of the big decisions |
Where this connects
- How to reason about AI system design — pair the decks with the framework; recall plus structure covers most of what you need on demand.
- Start here — the "night before" reading order puts these in sequence.