Can a machine learn presence without watching a person?
Artificial intelligence has become remarkably good at understanding language, generating images, and interpreting enormous amounts of digital information.
Yet it still knows surprisingly little about the physical world.
Today's AI mostly experiences reality through cameras, microphones, documents and prompts. It rarely develops a continuous relationship with the environments it inhabits.
We believe another approach is possible.
Rather than teaching machines to observe more, we want to explore whether they can perceive differently.
Resonance Machine is an experimental research project investigating whether a collection of inexpensive, privacy-conscious sensors can produce a meaningful representation of presence, movement, rhythm and the interaction between a person and physical space.
This is not a surveillance system.
It is not a medical device.
It is not another camera.
It is an attempt to ask a different question.
This project begins with inexpensive hardware, modest objectives and experimentally verifiable questions.
Can an artificial system learn the continuity of physical presence without relying on conventional visual surveillance?
Instead of asking:
"Who is this person?"
we ask:
"How has the world changed because someone is here?"
Every physical environment possesses rhythm.
Morning light.
Changing shadows.
Air pressure.
Temperature.
Radio reflections.
Movement.
Silence.
Breathing.
The cadence of footsteps.
Most of these signals remain below conscious perception.
Most intelligent systems ignore them entirely.
Our goal is to explore whether these weak signals, synchronized together, can form a richer understanding of presence than any individual sensor alone.
Modern intelligent systems increasingly depend on cameras.
While cameras are powerful, they are also intrusive.
We believe the future of embodied AI should include systems capable of understanding environments through subtle physical interaction rather than continuous observation.
Our working hypothesis is straightforward:
Multiple weak signals may together contain more meaningful information than one dominant sensor.
Instead of building another vision system, we want to construct a multimodal sensing instrument composed of inexpensive and widely available hardware.
Examples include:
None of these sensors alone defines a person.
Together they may produce information that no individual sensor can capture alone.
We use the word resonance deliberately.
Traditional sensing attempts to classify isolated events.
Resonance attempts to understand continuity.
A person changes a room.
A room changes a person.
Neither can be fully understood in isolation.
The machine should learn this relationship rather than merely recording isolated measurements.
The first prototype is intentionally modest.
Its objectives are:
The purpose is not perfect recognition.
The purpose is discovering whether synchronized multimodal sensing produces stable representations.
Privacy is a design constraint, not an afterthought.
It is the foundation of the project.
Our principles are:
We believe:
A system capable of sensing presence must be designed with consent before capability.
We are not interested in building another chatbot.
We are not interested in replacing people.
We are interested in building a new scientific instrument.
A machine capable of carefully observing the physical world without reducing it to images.
Build the first synchronized multimodal sensing prototype.
Collect a carefully controlled volunteer dataset.
Develop baseline models capable of recognizing:
Explore whether synchronized sensing can produce a stable high-dimensional representation of person-space interaction over time.
Publish our findings, limitations and future research directions.
Initial funding supports:
Our immediate objective is not to build a company.
It is to build a credible scientific instrument.
Fundamental research should not be limited to billion-dollar laboratories.
Important ideas can emerge from small independent research teams willing to ask unusual questions.
We believe this question is worth investigating.
Humanity has already given machines its books.
Its images.
Its language.
The next step may not be giving machines more information.
It may be giving them a careful, ethical and privacy-conscious way to encounter the physical world itself.