Program 06 · MACHINA · prototype · embodiment branch

Cognitive engines for machines
that build worlds.

MACHINA is not a robotics lab. It is a civilisation of machines — reasoning, building, and coordinating in a simulated world that obeys physics. Conventional autonomous systems navigate. MACHINA's systems think.

Current status — simulated bodies first, physical drones and robotics later. The metrics below reflect an idle/prototype sim state until live partner deployments come online. This is the embodiment branch of the stack, not a shipped product.

Substrate
N-d
cognitive manifolds · not XYZ
Coordination
Colony
no central command · emergent
Engineering
Factorio-class
supply chains · resource flow
Output
World-builders
ground · drone · hybrid swarms

Two directions

Body, then colony.

MACHINA splits the problem in two — the cognition of a single embodied machine, and the coordination of a colony of them. Both are simulated in the same physics; the science is in the gap between the two.

Direction I

Cognitive Mechatronics.

Sensorimotor cognition in a single machine. Body and mind as one system — perception, motor planning, and goal-formation entangled, not pipelined. The machine acts because it has decided to, not because a planner instructed an executor.

single agent · embodied loop · closed-form physics
Direction II

Dynamic Cognitive Engineering.

How colonies of machines engineer their own world in real time. Supply chains, route optimisation, mining, smelting, building — emerging from local rules, with no central command. The colony is the unit; the individual is interchangeable.

colony scale · emergent · Factorio-class logistics

Architecture

Three axes only.

The MACHINA stack is not deep — it is wide. A small number of foundations carry the entire research line: a substrate for thinking, a method for engineering, a class of outputs that close the loop back to the world.

Substrate
N-dimensional cognitive space — not XYZ. Machines reason inside configuration manifolds where the dimensions are degrees of freedom, energy budgets, and goal axes — not Cartesian coordinates. A path in that space is a plan.
Method
Factorio-class logistics — emergent supply chains, resource flow, colony-level optimisation. Conveyor and route topology grows from observed throughput, not from hand-drawn blueprints. The colony writes its own factory.
Output
Autonomous world-builders — ground robots, drones, and hybrid swarms that construct environments rather than navigate them. The simulator is the lab; the question is whether the same engines run on metal next.
20
Active agents
ground · drone · builder · scout
0
Structures built
emergent · not designed
0
Resource extracted
continuous harvest
0
Simulation tick
live · async
World simulator · live
Sense → Route → Build · MACHINA / Program 06
Drones
0
Ground
0
Builders
0
Scouts
0

Open the laboratory.

The simulator above is the live MACHINA prototype — the same agents you see on this page now also run during your read-through. Drones learn routes, builders complete sites, ore stocks rise. The full lab opens for partners building in cognitive mechatronics or colony-scale autonomous engineering.

Enter MACHINA Lab Live simulator · prototype · 2026-Q2