Worlds that grow. Lives that remember.

Living Worlds provides worlds: simulation, training and the lives an AI lives in them.

9 registered results met. Each is registered before it runs and scored once on data held back.

A glass tree with lives in its crown and the worlds they grow from beneath

The idea

Experience, not just data.

A model can read about a grasp, a storm or a traffic jam. It learns more by living through them.

In Living Worlds anything can live. A robot arm repeats a task a million times. A service lives through a quiet day, a surge and a failure. A grid lives a storm season. A population lives thirty years. Each keeps one record of everything that happened to it.

The same goes for a forest across four seasons, a river and its fish, a chair on a test line, a bridge in a storm, a piece of music or a painting taking shape. Living systems and built systems share one world, and each keeps its record.

That record is lived experience: a history an agent learns from, a dataset is built from and a digital twin is checked against. Ages and lives mean exactly this, experience gathered over simulated time. People are one kind of life among many.

What was learned does not stay in the world. Ipnops provides the inputs, integrations, processing and compute that carry it out to the real machine, the real program and the real data, and reads what happens back in.

Areas

Simulations, training, predictions, intelligence and interfaces.

What Living Worlds provides for physical and digital AI, area by area.

Simulations

Rooms and acoustics, crowds and people, populations and their lives, weather and earth, physics and robots, worlds and scenes. Simulated motion is checked for plausibility: free fall, ground contact and energy are tested against the declared measurement noise.

Training

Fitting and fine-tuning on lived runs, and evaluation on separate test sets.

Predictions

Forecasts, fast stand-ins for slow simulations, and uncertainty: a forecast carries its interval, and the interval is scored for coverage against what happened.

Intelligence

A rule-and-evidence core, and agents that carry a life record and remember what a person told them.

I/O and interfaces

Readings in and results out, between these worlds and robotics, machines, infrastructure and software.

Games and living worlds

Living populations and worlds for games: characters that are real selves, each with one life and its own memory carried across sessions, in worlds that keep living whether or not anyone is playing. Embodied in Unreal Engine and streamed to any screen, a console among them.

What it is for

Industries and areas.

Each area says what the simulations do and gives one example. Every result is in Research.

Acoustics and audio

Rooms, venues and studios: how a space sounds at positions nobody measured.

For example: How a hall sounds at every seat before load-in.

Cities and public space

Crowds and pedestrians in streets, squares and shopping centres.

For example: A crowd moving through a shopping centre.

Driving and streets

Vehicles driven through simulated town streets, with front-camera frames and trajectories recorded. A short path is drawn from a front-camera picture and a speed, and a short front-camera clip is rendered from a map-style still and a weather prompt.

For example: A vehicle driven through a town for a few seconds, frame by frame.

Buildings and design

Building models and CAD parts are read, checked, converted and meshed for simulation.

For example: A building model checked for validity, and a part meshed for a solver.

Robotics and machines

Robots and machines at work among people.

For example: People around a working robot, on days nobody had opened.

Weather and environment

Weather forecasts scored against measurements, and daylight at a desk.

For example: A forecast of the coming days at a place, scored against what was measured.

Populations and demography

Simulated people lived month by month: life courses and births, women's and men's life expectancy twenty years on in the countries tested, and how groups answer survey questions.

For example: Life expectancy twenty years on, in a year nobody had read.

Sensing and edge environments

Live readings from sensors brought in as measurements and forecast, and body pose read from an image.

For example: A room's temperature, air and light, streamed and forecast.

Molecules and materials

Structures and energies of molecules, proteins and crystals; relaxed crystal lattice constants are compared with measured values.

For example: A crystal relaxed to its lattice constant, and a small molecule relaxed to its lowest-energy shape.

What it uses

What it runs on and draws on.

Compute

Living Worlds runs on compute that Ipnops operates.

Physical and digital AI

Robotics, machines, infrastructure and software agents connect to these worlds through Ipnops interfaces.

Data

Public statistics and surveys, public room and pedestrian recordings, measured buildings and power networks, and weather and earth models as inputs.

Models

Rule-based engines for lives and rooms, learned surrogates and forecasters, and language models for voices, each scored against the real world before it is used.

A glass tree on a disc, with worlds in the soil and small scenes on its limbs

How it fits together

Worlds, lives and what they learn.

The ground

Places, physics and the rules they obey.

Lives and memory

Each life keeps one record and carries it forward.

Many kinds of life

Societies, cities, minds and machines in one world.

What they learn and show

Results checked against the real world, then carried out through Ipnops.

What you do

Build it. Simulate it. Train in it. Carry it out.

One world, the same steps whether it holds a robot cell, a crowd or a population.

01

Build the world

Describe it, or bring in what is real: scans, plans, scenes and simulation-ready assets.

02

Simulate it

Physics for robots, vehicles, cloth and fluids. Rooms that sound right. People and software agents that act.

03

Train and test in it

Policies, skills and behaviours learned across thousands of parallel runs, then tried against what could go wrong.

04

Carry it out

Through Ipnops, to the real robot, machine or program. Read what it does back in, and the world stays honest.

Carry it out

From the world to the real thing, with Ipnops.

Ipnops runs the compute and is the interface between these worlds and robotics, machines, infrastructure and software.

Simulate

From a question to a world.

Ask what you want to understand. Living Worlds builds the world, fills it with lives, runs it for as long as it takes, and shows what happened and why. Each use below says what it is built for and what it has already shown; every test and its result is on the research page.

Simulated crowds in streets, squares and venues, scored against recorded pedestrians.

01

Describe it

A place, a time, a population and the question, in plain words.

02

Bring in what is real

Your records, plans, scans and measurements, carried across.

03

Run it again

The same seed gives the same world.

04

See what held

Every life keeps its record; every result can be checked.

Play runs a small town on the engine's own rules, in your browser.

Results so far

Checked against the real world.

A few of the results. Each one opens in Research with its protocol, result and scorer.

30 of 31

Countries split by place and age

Given each country's national answer, Living Worlds split 1,556 questions it had never seen by city, countryside and age, inside the survey's own noise in 30 of 31 countries.

7 of 8

People around a working robot

How close real people came to a real robot in a room full of people carrying loads, predicted better than a robot-blind crowd in 7 of 8 recordings.

40 feeders

Power through feeders without a meter

The power through 40 unmetered feeders, every fifteen minutes for a year it had never seen: ahead of physics and a learned model in every season.

Topics

Every simulation, by topic.

One engine underneath. Four families of worlds, from people to machines, and products already built on them.

Populations

Societies · Civilizations · Markets · Minds

Machines, software and twins

Physics · Machines · Software · Digital twins

Places and phenomena

Cities · Acoustics · 3D · Environment

Building on it

Code

A sapling and its older selves, with a life's scenes in the trunk rings

One life

Any one of them will tell you its story.

Every life keeps one record of what it lived through, and changes only when the change holds. Follow one life and you see the world it belongs to.

Born

Into a place, a family and a time, with a temperament of its own.

Shaped

By the people around it, the work it does and the events that reach it.

Remembered

Every moment written to one record that travels with it, and that anyone can check.

Met

Ask it what happened and why. It answers as itself.

A glass tree whose crown holds many small worlds like fruit

Which world will you live in?

What grows here is meant to reach the real world. Start with a world of your own.