Recursive Learningthe games

We're here to do recursive eco-improvement. A thing that improves itself still has to adapt with its environment, or adapt the environment — so improvement that ignores the ground it stands on tends to stall. These are two small places where that hypothesis gets played rather than argued: you move something, and you watch what the move costs.

The Walk

Conditional probability, made into steps you actually take. A probability is never a property of the world on its own — it is a property of the world given a path. Move the allocation and the number moves.

Power-first, and Doom comes upabout 54%
Stop-first, and the Stop comes upabout 94%

Take the walk

As-If — HOT POTATO edition

Draw a character and a hypothesis, upright or reversed, and hold it as if true. You cannot turn the card over until you have written what changes if it is. The reveal is earned, not clicked.

The deck it plays over is still a private grammar, so this door opens on an explainer, not a round. It becomes playable the moment the deck is published — nothing else has to change.

See what it will be

The grammars in this channel

This repo is one channel of recursive.eco, held in git. Every grammar below can be opened, starred onto your own channel, forked, or corrected — the JSON is right here.