Intelligence is what high-dimensional systems do to resist decoherence.
Brains have more internal degrees of freedom than any observer can track. We derive what follows from that—mathematically, with simulations, and in peer-reviewed papers.
Drag a particle on each side.
Why AI can write a sonnet but can't catch a ball
Both AI and biological intelligence coordinate many semi-independent parts. The difference is how.
AI runs on a global clock. One loss function synchronizes everything from outside. You can read every weight and gradient. High-dimensional in representation, low-dimensional in dynamics.
Brains synchronize from within. Billions of semi-independent oscillators align through phase coupling—slower, messier, but the dynamics themselves are high-dimensional. The information lives in a space that measurement collapses.
That gap has consequences: for how fast brains can think, why psychedelics dilate time, and where current AI architectures hit ceilings that biology doesn't.
A 4D object projected to your 2D screen. Drag to rotate. Information is lost at every projection.
The substrate matters
A brain has ~1014 continuously coupled degrees of freedom. A GPU has one clock. Scaling parameters doesn't add dimensions to the dynamics—it just makes a bigger low-dimensional system.
Coordination is expensive
Getting distributed oscillators to align takes time that grows exponentially with how many you need to coordinate. That's why thinking is slow and why visual binding takes 30–50 ms.
Measurement destroys
Every observation projects a high-dimensional state onto something lower-dimensional. A spike train. An fMRI voxel. You can't observe the coordination without collapsing it.
How codes form
DNA, neural signals, language—all are low-dimensional representations of high-dimensional processes. The code carries meaning precisely because it discards most of the details.
This isn't a bug. It's the only way to get a signal out of a system with more internal states than any channel can carry.
Drag the bandwidth slider to compress the code channel and watch B's dynamics collapse.
Papers
4 published, 3 under review.
Intelligence as High-Dimensional Coherence: The Observable Dimensionality Bound and Computational Tractability
Intelligence emerges whenever a system maintains high-dimensional internal dynamics that collapse to coherent outputs — and this architectural principle, not clever algorithms, explains the million-fo
Minimal Embedding Dimension for Self-Intersection-Free Recurrent Processes
You need at least 3 dimensions for continuous cyclic dynamics — below that, trajectories collide and the system is forced into discrete categories. This is a hard geometric constraint with implication
Coherence Time in Biological Oscillator Assemblies Bounds the Rate of State Registration
Distributed biological computation is bottlenecked by coherence time — the waiting time for multiple semi-independent oscillator modules to align within a tolerance window before a state transition ca
A Thermodynamic Foundation for the Second Law of Infodynamics
Vopson's "second law of infodynamics" proposes that information entropy decreases over time, with high-symmetry states representing minimum information entropy. This paper provides the thermodynamic m
The thesis
The full argument in one page.
Read →All papers
4 published, 3 under review across five disciplines.
Browse →Simulations
Interactive demos. Play with the ideas.
Explore →