How it works

From “what is attention?” to reading the paper yourself

Scholar Quest is a loop: it finds the next thing for you to read, writes it for you, notices when you finish, and checks that it stuck.

  1. Tell us what you want to learn

    Pick a topic and take a short placement check. We note what you already know so nothing is pitched too low or too high.

  2. Get editions written for you

    Lessons and research paper summaries arrive on your Kindle, e-reader or inbox, with the background you're missing woven in.

  3. Read, highlight, move on

    Highlight the last line when you finish. That's the signal: your skill tree updates and the next edition is lined up.

  4. Prove it stuck

    Quick review questions bring ideas back before you forget them, and build challenges show you can use them.

What “written at your level” means

Your learner profile records what you know: from your placement answers, the lessons you've finished and the challenges you've passed.

When an edition is written, the profile goes with it. If you already know linear algebra, it won't explain matrix multiplication. If you've never met softmax, a short background box appears right where it's used, and a background chapter at the start covers anything bigger.

Three depths for every paper

  • Summary: about 10 minutes. The problem, the idea, why it matters.
  • Overview: 25 to 40 minutes. How it works, with the key equations explained.
  • Deep: an annotated walkthrough of the whole paper, section by section.

Any depth can be read in plain-language mode for a first pass.

How finishing works

The last line of every edition says “Highlight this line to mark this lesson complete.” Highlight it on your Kindle. When your highlights sync, your skill tree updates and the next edition is lined up. There's nothing to tick off.

How review works

A few days after you finish, short questions written from that edition come back to you. Answer them in a few minutes. Get them right over growing intervals and the skill's Recall tier unlocks.

Read your way into a new field

We're opening the private beta to 5 to 10 people first. Starting topic: AI and machine learning.

Join the beta