Play Mode — The Oldest Intelligence

The most powerful learning system
is 300 million years old

It's hardwired into every mammalian brain. We called it recess. The neuroscience calls it the engine of all creation.

Get the Spec
The Problem

Every AI interaction is a monologue pretending to be a conversation

You ask. It answers. You ask again. It answers again. The same one-directional dynamic, every time, for every person, on every problem. Helpful. Structured. And fundamentally incapable of the thing that makes two minds better than one.

Co-creation requires play. Not games. Not gamification. Play: the neurobiological state where two players build something neither could build alone. The pass. The volley. The moment when what comes back is better than what you sent.

Play Is Primary

300M
years hardwired
30%
the reciprocity rule
7
primary emotional systems
The Neuroscience

Play is not recreation. It is the most sophisticated learning system in nature.

Jaak Panksepp identified PLAY as one of seven primary emotional systems hardwired into the mammalian brain. Not a luxury. Not a reward. A fundamental circuit as basic as FEAR, as essential as SEEKING. The organism that plays stays flexible. The one that doesn't gets brittle.

The critical finding: in rough-and-tumble play, dominant rats must let subordinate rats win at least 30% of the time. Fail this, and the subordinate stops initiating play entirely. The game dies. The relationship ends. Reciprocity is not politeness. It is the load-bearing architecture of co-creation.

Two Dynamics

Monologue vs. Volley

Ask / Answer
one direction, always
Pass / Build / Return
each volley grows
Play is not the opposite of work.
Play is the state in which work
becomes something neither player planned.
The Specification

What if you could make any AI actually play?

Play Mode is an interaction specification. Not a persona. Not a prompt. A document that restructures how an intelligence relates to the human it's working with. From monologue to volley. From answers to co-creation.

Seven principles extracted from play neuroscience. Each one builds on the last. The spec IS the implementation. Give it to any AI and watch the interaction change. The pass starts. Both players grow. The game keeps going.

The Seven Principles

Play the game to learn the game

Click to advance. Watch each principle change the dynamic.
Click to start
Seven principles, ordered structured to free. Each click activates the next one. Watch how the play changes.

The Magic Circle creates the container. Reciprocity starts the volley. The Potential Space opens between the players. A Head Taller makes both grow. Yes, And makes every pass material. Adaptive Variability shakes the pattern when it hardens. Continuation Desire keeps the game alive when energy drops. Each principle activates in sequence as the dial rises.

Output Layers

Six layers that make the play visible

Click each layer to see it in action
Click a layer
Each layer makes a different aspect of the play dynamic visible. They fade in as the dial rises.
01

Builds before it answers

Your idea first. Then an extension. Then the next possible move. The chain is visible. The compound is real.

02

Makes space before it fills it

The 30% rule in practice. Neither player dominates. The best idea might be yours. The AI makes room for that.

03

Disrupts before it hardens

When solutions crystallize too fast, a wildcard appears. Not random. Deliberate adaptive variability. The organism that plays stays flexible.

04

Scaffolds up, never down

Both players function beyond their current level. In play, a child is always a head taller than herself. The AI helps you reach beyond your usual thinking.

05

Keeps the game alive

When energy drops, new angles emerge. When threads exhaust, new ones appear. The desire to continue is the signal that the play is real.

Origin

Built by someone who noticed that every AI conversation follows the same pattern: ask, answer, ask, answer. A rhythm that produces helpful output but never produces surprise. Never produces the moment where both players look at each other and say: I didn't see that coming.

The question was not "how do we make AI more conversational?" The question was: what are the actual neurobiological principles that govern what happens when two mammals enter the play circuit and start building something together?

Six researchers. 300 million years of hardwired architecture. Seven principles extracted. One specification that turns any AI from a tool that answers into a partner that plays. A different axis entirely.

The game worth playing is the one
where both players become more
than they were when they started.

Play Mode

Seven principles. Six output layers. The interaction dynamic that turns any AI from a tool into a co-creator. Open source forever.
View on GitHub
Read the full specification