Dr. Arthur Gazaryants, DOM
Hello everyone,
Thank you for being part of the Artupuncture family. Many of you have asked about the new paper I just finished: “Participatory Anti-Entropic Selection in Conscious-Agent Networks.” It is written in dense academic language because I wanted every step proven mathematically for scientists to check. Today I am giving you the plain-English version, exactly what the paper says, told like we are sitting together in the treatment room.
No PhD needed. Just curiosity and an open mind.
The Big Question Behind the Paper
Here is a question that has fascinated philosophers and scientists for centuries: Is the physical world the deepest layer of reality, or is there something underneath it?
Most of modern science assumes matter comes first and consciousness somehow emerges from it, neurons firing in the right pattern produce your experience of the color red, the taste of coffee, the ache in your lower back. This is called physicalism and it has been enormously productive.
But there is a serious alternative, developed over decades by cognitive scientist Donald Hoffman and others, called Conscious Realism. The core idea: conscious experiences are fundamental and the physical world we all share is more like an interface, a kind of shared desktop, generated by networks of conscious agents interacting with each other.
Think of the icons on your computer screen. The little folder icon is not the actual file, it is a useful picture that helps you interact with something deeper. Hoffman’s proposal is that physical objects, including our bodies, play a similar role: they are useful representations, not the ultimate reality.
This idea has faced a fair criticism: it sounds intriguing, but where is the math? Can you actually build a working model and show that it produces something interesting? That is what my paper set out to do.
What I Actually Built
I constructed the simplest possible version of Hoffman’s framework that you can write down completely and check every number. Here is the setup:
Two agents (think of them as two observers, you and me, for example) share access to one piece of hidden information called z. You can picture z as a light that is either on or off. Neither of us sees z directly. We each get a slightly noisy message, correct 99% of the time, flipped 1% of the time, like a faint bit of static on an old TV.
Based on what we each perceive, we independently choose an action: “I vote the light is on” or “I vote the light is off.”
Then the shared light updates based on our votes. If we agree, it usually moves to match our vote. If we disagree, it mostly stays where it was. A tiny bit of noise (1%) keeps the system from getting permanently stuck.
That is the entire model. Two agents, one shared bit, noisy perception, independent voting, a simple update rule. Small enough to fit on a napkin.
The Key Ingredient: Awareness Amplitude
Here is where it gets interesting. I added a single dial to the model called awareness amplitude, written as the Greek letter α (alpha).
When α is zero, the agents choose their votes with a small amount of randomness, they mostly match what they perceived, but not with any extra effort toward clarity.
When α is turned up, each agent becomes increasingly biased toward the action that will make the future state of the shared light as predictable as possible. In technical language, they minimize the predicted entropy of the next state. In plain language: they try harder to keep the shared picture clear and orderly.
This is not a new invention. The mathematical structure is identical to what neuroscientists call active inference, a well-tested framework developed by Karl Friston in which biological agents minimize surprise about their future sensory states. My α is equivalent to what physicists call an inverse temperature parameter. The math is known. What is new is showing that this mechanism works explicitly inside the Hoffman conscious-agent framework, with every number derived and checkable.
What the Math Shows
The paper proves three things rigorously:
- As α increases, the agents converge toward perfect agreement. At α = 0, the two agents already agree about 96% of the time (because they start from the same noisy source). By α = 5, agreement rises to about 97.3% and the mutual information between their actions climbs from 0.76 bits toward a theoretical maximum of about 0.86 bits. These are small numbers in a tiny model, but the trend is what matters: more awareness, more shared order.
- In the limit (as α goes to infinity), each agent’s vote becomes a perfect copy of what they perceived. The decision collapses to a point: no randomness, no wobble. The paper proves this analytically—it is not just a simulation result.
- When α = 0, the system is maximally disordered. I proved that at zero awareness amplitude, the system’s resting state is the uniform distribution, every possible configuration is equally likely. There is no built-in bias. All the order that emerges comes from turning α up.
The key underlying insight (Lemma 1 in the paper) is that the action matching your percept always produces lower future uncertainty than the mismatching action. This is intuitive, if you saw the light is on, voting “on” keeps the future more predictable, but the paper proves it holds for every possible noise level, not just the specific numbers I chose.
What This Means
What the paper demonstrates: In a minimal two-agent model, an entropy-minimizing bias (awareness amplitude) produces measurable increases in intersubjective agreement and shared order. The mechanism is mathematically rigorous, fully reproducible and consistent with both active inference and the Hoffman conscious-agent framework.
This is how science works. You start with the smallest possible case, prove it rigorously, and then carefully extend. The two-agent model is a proof of concept, a demonstration that the framework can produce quantitative results, not a finished theory of everything. The exciting work now is scaling up: more agents, more complex states and eventually, contact with measurable biology.
Why I Find This Inspiring Anyway
Even with all those caveats, I find this work deeply exciting and this is why.
For as long as I have practiced medicine, I have noticed something that the standard biomedical model struggles to explain: the quality of a patient’s awareness during treatment seems to matter. Patients who are mentally present during acupuncture, who notice the sensation of the needles, who breathe with intention, who participate rather than just lying there scrolling their phones, tend to respond better and hold their improvements longer. This is a clinical observation, not a controlled trial. But it is a pattern I have seen consistently over nearly 25 years.
The paper gives me a formal analogy for why that might be the case. If something like awareness amplitude exists in biological systems, if the degree to which a system actively minimizes future uncertainty affects how coherently its parts coordinate, then conscious participation would not be a mystical add-on. It would be a measurable parameter in the coordination dynamics.
I want to stress the word analogy. The model does not prove this is what happens in your body. It shows that in a simple mathematical world, the principle works. Whether it scales to biology is an open question and one I mention in the paper’s future directions section as requiring substantial additional work.
Practices That Resonate with This Principle
With that framing in mind, here are practices I already recommend to patients that resonate with the principle of participatory awareness, not because the paper proves they work through this mechanism, but because the analogy is suggestive and the practices are beneficial on their own well-established merits:
Present-moment body awareness. Sitting quietly for five minutes and noticing sensations without judgment. This is well-supported by mindfulness research independently of my paper. The model offers a suggestive parallel: each moment of noticing is structurally similar to what the agents do when they attend to their perceptual signal rather than acting randomly.
Conscious participation during treatment. When you receive acupuncture, notice the quality of the needle sensation. Breathe into the areas being treated. This kind of engaged presence is something practitioners across many traditions have long recommended. The model provides a mathematical metaphor for why participation might matter, but the clinical recommendation stands on its own.
Shared intention. When you and your practitioner (or you and a loved one) hold the same intention during a healing encounter, you are doing something the model would recognize: two agents coordinating toward the same outcome. We don’t know yet whether this mechanism applies directly to human interaction, but research on therapeutic alliance already suggests that shared intention between patient and practitioner improves outcomes.
Gentle, attentive movement. Tai chi, qigong, slow walking with full sensory attention. These practices have robust evidence bases for pain reduction, balance and wellbeing. The model gives me a way to think about whyattention might be the active ingredient, but the practices are worth doing regardless.
What Comes Next
The paper is under review at Frontiers in Complex Systems. If it is accepted, it will be the first fully explicit, mathematically reproducible worked example of the Hoffman conscious-agent framework with derived quantitative statistics.
The next step I am most excited about is extending the model from two agents to many, dozens, hundreds, eventually thousands. I suspect (and the paper conjectures) that above a certain threshold of awareness amplitude, large networks will show sudden phase transitions into coherent consensus, similar to how water suddenly freezes or how a room full of people suddenly starts clapping in unison. Proving this is a substantial project and I will keep you posted.
In the longer term, I hope to connect this theoretical work to measurable biomarkers, the kind of data we already collect through our functional medicine panels. Can we find signatures of coordination coherence in your labs? That is speculative today, but it is the direction I am heading.
If you are curious about the full paper, I am happy to share it. If you want to experience what participatory awareness feels like in practice, book a session and we will explore together not because a two-agent binary model told us to, but because 25 years of clinical experience and a growing body of research both point the same way: how you show up matters.
You are not a passive observer of your health. That much, I believe the science supports.
With gratitude,
Dr. Arthur Gazaryants, DOM
Artupuncture • Calabasas, CA