4Color.AI

Research · invention · learning

The four-color world of AI, math, and cognition.

4Color.AI is Colin Li's public lab. Walk in through any of the doors below, or read straight down.

The Four-Color WorldEvery idea in the lab, laid out at once.See it all →The Proof IslandOne theorem, checked in front of you.Watch it work →The Thinking SpaceStep inside the ideas and walk.Walk it yourself →

Featured projects

Experiments and research builds from the lab, starting with MirrorX.

Live-testable V2 spine

MirrorX

A persistent agent-world research project now running controlled agent jobs against a real KickStart staging sandbox.

Open

Learning trails

Guided paths that turn research questions into concepts students can inspect and build from.

Beginner to intermediate

From graph coloring to AI systems

A learning trail from the Four Color Theorem to constraint solving, representation, and AI system design.

OpenIntermediate

How agents reason, reflect, and coordinate

A learning trail for understanding agent behavior through durable identity, memory, app roles, reflection, and coordination.

OpenIntermediate

Is logical thinking just a form of harnessed randomness?

A learning trail asking whether reasoning, search, and creativity depend on controlled randomness.

Open

Field notes

Short essays on AI systems, mathematics, cognition, and the boundaries between them.

note

The four-color world

Why 4Color.AI uses the Four Color Theorem as a metaphor for AI, math, and cognition.

Open

Built by Colin Li

Math + CS founder, builder of CodingMind Academy and JCIA, exploring AI-native learning and agent societies through public experiments.