Small rules.
Unfamiliar life.
A game about breeding algorithmic organisms—and discovering what they can do together.
You direct the research. Your agent runs the experiments. Tiny programs become cells, colonies, and eventually the building blocks of a larger world.
A game in design. This site is an invitation to explore the proposal.
Begin with a wound.
A line of cells is trying to sort itself. Each knows only its neighbors. Then some stop acting.
Can the colony restore order?
Concept illustration · not a simulation result
Breed a fast lineage with a resilient one. Inspect the offspring. Sometimes a useful move makes the world look less ordered before it gets better. The interesting part is finding out why—and whether it works somewhere new.
Walk through the first fifteen minutesThe conversation is the interface.
Bring your own agent. Give it a question worth investigating.
An example of intended play. The Rust engine and agent skills are still to be built.
The engine will own the state, rules, and evidence. Your agent will operate the laboratory and help you understand it. Every creature must work without the agent inside it.
How the agent and engine fit togetherWhat you master becomes a beginning.
The proposed world changes scale. A colony you once struggled to understand becomes one part of your next organism.
The cell
Give a small program a body. Watch what its local rules make possible.
The wound
Mix lineages. Disable cells. Discover which colonies can recover.
The tissue
Let a successful colony become a component of something larger.
The ecology
Build partnerships between specialists. Meet environments that expose their weaknesses.
The frontier
Evolve the processes that discover organisms, representations, and algorithms.
The first prototype starts with cells and damage recovery. Later chapters are a direction to explore.
A frontier worth competing for.
Discover organisms that solve harder problems, survive unfamiliar conditions, and do more with the same resources.
Smarter agents and more search can help you find them. Ranked organisms will face equal execution limits and unseen challenges. Their results must be independently reproducible.
Read the proposed competition rulesPlay that leaves something useful behind.
Inspired by Michael Levin’s work on minimal collective systems, Platonik asks how much behavior a small set of rules can express.
A better algorithm. A surprising failure. A conjecture that survives a test. These can all be discoveries worth keeping. The long-term research ambition reaches toward computational complexity; a game score is evidence about its challenges, not a proof about P versus NP.
Explore the research foundationsThe world starts with a few rules.
The design is open. Start with the first creature, or go deeper into the experiments behind it.
Read the Platonik field guide