Why Quizzing?

The research

Decades of cognitive science research converge on one finding: active recall is the most effective way to learn. Not re-reading, not highlighting, not summarizing. Testing yourself.

This is sometimes called the testing effect — the finding that retrieving information from memory strengthens that memory far more than passively reviewing it. Andrew Huberman covers the key studies in his Optimal Protocols for Studying & Learning episode. Andrej Karpathy recently explored this same idea in the context of AI-assisted learning.

The Feynman Technique takes this further: if you can explain a concept simply, from scratch, as if teaching someone else, you understand it. If you can't, you've found your gaps.

Why not flashcards?

Traditional flashcards (Anki, Mochi) work well for rote memorization — vocabulary, dates, formulas. But they're limited for deeper understanding:

• Static questions produce static answers. You end up memorizing the wording of the card rather than understanding the concept.
• Cards don't adapt. A flashcard can't probe a gap it doesn't know about.
• Recognition isn't understanding. Picking the right answer from a prompt is very different from explaining something from scratch.

AI changes this. An AI assistant can rephrase questions, follow up on weak spots, ask you to explain why rather than just what, and evaluate your explanations for accuracy and depth. It's the difference between a multiple-choice test and a conversation with a knowledgeable tutor.

We think the best format for AI-powered quizzing is still an open research question. Static flashcards may be a thing of the past — but what replaces them? The Feynman technique and Socratic dialogue are strong starting points, and we're actively exploring what works best.

As Andy Matuschak explores in How Might We Learn?, the tools we use for learning are far behind the tools we use for everything else. FeynmanLM is our attempt to close that gap.

Why not a note-taking app?

There are plenty of great note-taking apps — Notion, Obsidian, Roam. FeynmanLM is deliberately not one of them.

Note-taking apps are designed around capturing and organizing information. FeynmanLM is designed around tracking sources and testing your understanding of them. The difference matters:

• Source tracking, not note management. FeynmanLM tracks what you've read, watched, and listened to — the source material itself, not your notes about it.
• AI-powered quizzing, not review. Instead of re-reading your notes, you quiz yourself with an AI that has access to the original source.
• Understanding over organization. The goal isn't a perfect knowledge base. It's actually understanding and remembering what you've learned.

Your AI assistant (Claude, ChatGPT) already has access to your sources through MCP. That's a better "note-taking" system than any app — it can search, summarize, compare, and quiz you across everything you've saved.