NOUMENAX
02 / RESEARCHWorking questions · Implementation notes

Questions worth building for.

This is a working research archive: the questions behind my projects, the architecture I can document, and the next things I need to prove. These are development directions, not claims of published research.

01Multi-arm robotics

Coordination without conflict.

Multi-Arm Robotics is a ten-arm initiative in development at Ethernovia. It explores how networking, centralized AI compute and articulated arms can support cooperative manipulation; coordinated operation has not been demonstrated.

How can independent arms share space and make useful progress without obstructing each other?

Related engineering

02Computational accessibility

Useful simulation on ordinary hardware.

HOUSE explores household-agent perception, decision-making and task completion in repeatable CPU-first environments. It combines rule-based physical approximations, configurable perception and deterministic evaluation; full rigid-body dynamics remain outside its current scope.

What can we evaluate before depending on an expensive GPU or a physical robot?

Related engineering

03Observation & verification

Evaluation an agent cannot bluff.

The public prototype uses limited observations, independent goal checks and replay verification. Those are inspectable engineering choices, not evidence of learned general intelligence.

How do we know a policy solved the task without leaking the answer through its observations?

Related engineering
SYSTEMArchitecture explorer

Inside the symbolic prototype.

Documented prototype architecture · Public technical notes

Household world

Completed

Procedural rooms, entities, spatial relations, residents and household state form the environment. Seeded generation makes initial conditions repeatable.

The world uses symbolic state and grid/box geometry, not continuous collision meshes or a photorealistic scene.

Component descriptions follow the public HOUSE repository.

Full engineering case study

Research progress

Built. Building. Next.

Completed work

  1. HOUSE: world kernel, perception, 16 domainsCompleted

    HOUSE — CPU-First Robotics Simulation & AI-Agent Evaluation Framework

  2. Evaluation suite and leaderboardCompleted

    HOUSE — CPU-First Robotics Simulation & AI-Agent Evaluation Framework

  3. Temporal evaluation, dialogue and memory updateCompleted

    HOUSE — CPU-First Robotics Simulation & AI-Agent Evaluation Framework

In development

  1. Author and verify more catalog scenariosIn Development

    HOUSE — CPU-First Robotics Simulation & AI-Agent Evaluation Framework

  1. Requirements definitionIn Development

    MULTI-ARM ROBOTICS

  2. Mechanical designIn Development

    MULTI-ARM ROBOTICS

  3. Component selectionIn Development

    MULTI-ARM ROBOTICS

  4. Initial assemblyIn Development

    MULTI-ARM ROBOTICS

  5. Electrical and networking integrationIn Development

    MULTI-ARM ROBOTICS

  6. AI software integrationIn Development

    MULTI-ARM ROBOTICS

  7. Individual arm testingIn Development

    MULTI-ARM ROBOTICS

  8. Coordinated system testingIn Development

    MULTI-ARM ROBOTICS

  9. Final demonstration preparationIn Development

    MULTI-ARM ROBOTICS

Future objectives

  1. Extend and validate physical approximations and geometric-camera behaviorPlanned

    HOUSE — CPU-First Robotics Simulation & AI-Agent Evaluation Framework

  2. Publish reproducible CPU and memory benchmarksPlanned

    HOUSE — CPU-First Robotics Simulation & AI-Agent Evaluation Framework

  3. Record real task execution and replay demonstrationsPlanned

    HOUSE — CPU-First Robotics Simulation & AI-Agent Evaluation Framework

  4. Publish a versioned release with evidence-backed documentationPlanned

    HOUSE — CPU-First Robotics Simulation & AI-Agent Evaluation Framework

Dates appear only where recorded in the public project history. Future objectives have no implied completion date.