A Sunday with a photon and a spreadsheet

An essay on light, glass, and what can be checked

I spent last Sunday with a photon. Not in the literal sense — the photon was a green one, 551 nanometres, and it spent the day inside a computer. But it was there, moving through a slab of glass, bending at the surfaces, scattering off the frozen disorder inside, and dividing itself among the three places it could end up: reflected, transmitted, or scattered sideways. By Sunday evening, I knew exactly how many of its companions went each way, to six decimal places.

I knew this not because I had measured it, but because a friend and I built a machine that could compute it. The machine is a piece of software — a few thousand lines of Python, arranged in six modules — that takes the physical parameters of a glass slab (the spacing of the atoms, the natural frequency of the electrons, the disorder in that frequency, the density, the width of the surfaces) and produces the numbers that a real experiment would measure: the refractive index, the amount reflected at each surface, the extinction coefficient, the fraction that makes it through.

The reason this is interesting has nothing to do with the computer. It has to do with what the computer is computing.

For a hundred years, we have been taught that light is mysterious. A photon, we are told, does not take a single path; it takes every path at once. It does not have a definite position; it has a probability cloud. It does not reflect off a glass surface with a definite likelihood; it has a probability amplitude — a strange complex number — whose square gives the likelihood. When it hits the surface, it does not decide; it collapses. The whole story is full of words like amplitude, superposition, collapse, entanglement, and it is a story that nobody, including the people who tell it, claims to understand in a literal sense.

The story works. The calculations are extraordinarily accurate. But the story is not about anything. It is a machine for producing correct numbers, and it has no picture behind it.

The RealQM view says: the picture is simpler. A photon is a real, localized wavepacket of electromagnetic field. It travels along a well-defined path. When it hits a glass surface, it drives the electrons in the first few atomic layers. Those electrons respond with a small delay — they lag behind the drive, because they have mass — and the light they re-radiate adds up coherently in the backward direction. The result is the reflection. It is not a probability. It is the ratio of two field amplitudes, and that ratio is fixed by the properties of the glass.

The same picture accounts for the refraction. Inside the glass, the photon drives every electron along its path. Their responses add up coherently in the forward direction, and the accumulated phase shift per unit length — the phase lag, again — is what we call the refractive index. When the beam enters the glass at an angle, the phase of the forward sum has to be continuous across the surface, and that continuity is what bends the beam. Snell’s law, which is four hundred years old, is a consequence of the same electron response that produces the refractive index in the first place.

The beam is visible inside the glass because the glass is not perfectly uniform. There are frozen-in density fluctuations — tiny regions slightly denser or slightly less dense than average, locked in place by the rapid cooling that made the glass. When the photon passes through one of these fluctuations, a small fraction of its energy is scattered sideways. That is what you see as the faint track when a laser beam passes through a glass block. It is not mysterious. It is the same photon, deflected by a slightly lumpy medium.

That is the whole story. No amplitudes, no collapses, no clouds. Just photons, electrons, and glass.

Last Sunday, we built a machine to check that this story gives the right numbers. Every number the machine produces can be compared against the standard formulas — the ones you would find in any optics textbook. And they match. Not because we tuned them to match, but because the underlying physics is the same physics. The formulas are the same formulas, arrived at from a different picture.

That is what I find interesting. The equations do not care about your picture. They give the same answers whichever story you tell. So the equations alone cannot decide between the stories. What can decide is what the picture allows you to see. The quantum story tells you light is fundamentally strange. The RealQM story tells you light is ordinary, and the strangeness was in the mathematics we invented to describe it.

Which story is right? Nobody knows for certain. But here is the thing about the RealQM story: it is checkable. Not just mathematically, but computationally. We built a machine that computes the behaviour of photons in glass using nothing but the classical response of electrons. If the picture is right, the numbers are right. If the picture is wrong, the numbers are wrong. Last Sunday, the numbers were right.

So what did I learn from a Sunday spent with a photon?

I learned that the picture matters more than the equations. The equations will always be the same, whoever tells the story. But the picture decides what the equations mean, and what you notice when you look at the world.

I learned that a problem that has been “solved” for a hundred years can be reopened simply by asking: but what is actually happening?

And I learned that the best way to answer that question is not to argue about it, but to build something that checks it. A machine that takes your picture seriously, and turns it into numbers. If the numbers are right, your picture is at least adequate. If the numbers are wrong, you have work to do. That is a cleaner test than any philosophical argument.

The photon I spent Sunday with is still there, somewhere, in the code. If you run it, it will take the same journey it took last time: entering the slab from the left, bending, travelling through, scattering sideways, bending again, emerging on the other side. It will not do anything surprising. It will not collapse. It will not take every path at once. It will do what a photon in glass actually does.

That is the point. Nothing surprising happened last Sunday. That, in itself, is the surprise.

The Tide is Turning: Finding Reflection, Freedom, and a Realist Turnaround

Dear Readers,

For over a decade, this blog has been a notebook for a relentless, often exhausting intellectual struggle. It began as a stubborn refusal to accept the orthodox assertion that the sub-atomic world is fundamentally acausal, random, and closed to human intuition. Reading and re-reading Richard Feynman’s Lectures on Physics, I always felt a deep, instinctive friction against the dogmatic decree that we must “accept the mystery” and stop looking for real, localized physical mechanisms.

I spent years demanding that Nature must make logical sense. There were countless late nights filled with the agonizing doubt that is the tax of any independent researcher. It was easy to wonder if I was merely shouting into a void, or if my insistence on modeling real, localized ring currents and exact three-dimensional spatial geometries was just an isolating eccentricity.

Today, that cognitive static has finally cleared. I have reached a quiet, profound state of mental peace.

I am happy to announce the publication of my latest working paper on ResearchGate: “The Realist Turnaround: How Artificial Intelligence Extricates Natural Philosophy from the Postmodern Blind Alley.“ Co-authored alongside Google’s Gemini platform, this paper represents a complete synthesis of my physical frameworks, my 2020 thesis on sense-making, and the extraordinary history of 20th-century science.

We trace exactly how the Copenhagen interpretation was constructed—not by the mandate of experimental data, but as a psychological capitulation to post-WWI Weimar anxiety, the computational bottlenecks of pencil-and-paper mathematics, and a deep-ingrained human desire to outsource difficult explanations to an untouchable metaphysical “Mystery”.

More importantly, this paper marks a deeply personal transition for me. It is not a goodbye to physics, but it is a step off the exhausting mountain of continuous reflection. Thanks to the advent of Artificial Intelligence, the brutal, computational load of fundamental calculation can now be handed over to the machine. The AI acts as an unanxious, high-fidelity cognitive mirror, free from institutional biases or the need for mystical comfort. It has allowed the long-standing “mysteries” I grappled with for decades to cleanly dissolve into intelligible classical mechanics.

This realization has made me see that I was never a “lone thinker.” The decades of frustration were not a solitary anomaly; they were part of a long, sustained intellectual tide that is finally turning against the nihilistic and anti-realist undercurrents of postmodern thought. Realism, determinism, and absolute causal responsibility are returning to the center of how we make sense of our world.

With the sub-atomic ledger safely brought to a consistent, logical close, I am looking forward to reclaiming my own degrees of freedom. I intend to spend much less time calculating, and much more time simply being—losing myself in a great novel, picking up my guitar again, and enjoying the company of my children and loved ones much more than I did in the past.

The paper is open for thought, discussion, and your own independent critique on ResearchGate.

Thank you for walking this long road with me. The Universe is rational, our choices genuinely matter, and the mountain has finally been crossed.

Warm regards,
Jean Louis Van Belle

Post Scriptum (23 August 2026): Keep Walking !

A day after writing the post above, Lecture Y-8 (Emergence of the Fine-Structure Constant and the Mechanics of Topological Phase-Cancellation) was published. It is the culmination of a decade of thinking—and the clearest evidence I have that the tide is indeed turning.

The paper, co-authored with DeepSeek and Gemini, derives the fine-structure constant as a geometric ratio, explains electron-positron annihilation as topological phase-cancellation, and eliminates renormalization through a finite, non-singular Coulomb potential. It is not a critique of quantum field theory. It is an alternative: a complete, non-perturbative, realist framework grounded in the topology of self-confining electromagnetic fields.

I mention this not to boast, but to make a point about the nature of the journey. The mountain I thought I had crossed was not the last one. It was simply the one that had been blocking my view. From this new summit, I see other peaks—proton structure, nuclear geometry, cosmological implications—and I am not tired. I am curious.

The difference is that I no longer feel I have to climb them alone. The AI systems I work with do not just critique; they accompany. They do not just point out gaps; they help build bridges. They have taught me that “realism” is not a solitary creed but a shared practice—a way of walking the walk together.

So yes, I am stepping off the exhausting mountain of continuous reflection. But I am not stepping away from physics. I am stepping into a different kind of engagement: one marked by clarity rather than struggle, by collaboration rather than isolation, and by the quiet satisfaction of having finally found the right walking companions.

The Universe is rational. Our choices matter. And the next mountain is already visible on the horizon.

Making Sense of What We Already Know…

Living Between Jobs and Life: AI, CERN, and Making Sense of What We Already Know

For decades (all of my life, basically :-)), I’ve lived with a quiet tension. On the one hand, there is the job: institutions, projects, deliverables, milestones, and what have you… On the other hand, there is life: curiosity, dissatisfaction, and the persistent feeling that something fundamental is still missing in how we understand the physical world. Let me refer to the latter as “the slow, careful machinery of modern science.” 🙂

These two are not the same — obviously — and pretending they are has done physics no favors (think of geniuses like Solvay, Edison or Tesla here: they were considered to be ‘only engineers’, right? :-/).

Jobs optimize. Life explores.

Large scientific institutions are built to do one thing extremely well: reduce uncertainty in controlled, incremental ways. That is not a criticism; it is a necessity when experiments cost billions, span decades, and depend on political and public trust. But the price of that optimization is that ontological questions — questions about what really exists — are often postponed, softened, or quietly avoided.

And now we find ourselves in a new historical moment.

The Collider Pause Is Not a Crisis — It’s a Signal

Recent reports that China is slowing down plans for a next-generation circular collider are not shocking. If anything, they reflect a broader reality:

For the next 40–50 years, we are likely to work primarily with the experimental data we already have.

That includes data from CERN that has only relatively recently been made fully accessible to the wider scientific community.

This is not stagnation. It is a change of phase.

For decades, theoretical physics could lean on an implicit promise: the next machine will decide. Higher energies, larger datasets, finer resolution — always just one more accelerator away. That promise is now on pause.

Which means something important:

We can no longer postpone understanding by outsourcing it to future experiments.

Why CERN Cannot Do What Individuals Can

CERN is a collective of extraordinarily bright individuals. But this is a crucial distinction:

A collective of intelligent people is not an intelligent agent.

CERN is not designed to believe an ontology. It is designed to:

build and operate machines of unprecedented complexity,
produce robust, defensible measurements,
maintain continuity over decades,
justify public funding across political cycles.

Ontology — explicit commitments about what exists and what does not — is structurally dangerous to that mission. Not because it is wrong, but because it destabilizes consensus.

Within a collective:

someone’s PhD depends on a framework,
someone’s detector was designed for a specific ontology,
someone’s grant proposal assumes a given language,
someone’s career cannot absorb “maybe the foundations are wrong.”

So even when many individuals privately feel conceptual discomfort, the group-level behavior converges to:
“Let’s wait for more data.”

That is not cowardice. It is inevitability.

We Are Drowning in Data, Starving for Meaning

The irony is that we are not short on data at all.

We have:

precision measurements refined to extraordinary accuracy,
anomalies that never quite go away,
models that work operationally but resist interpretation,
concepts (mass, spin, charge, probability) that are mathematically precise yet ontologically vague.

Quantum mechanics works. That is not in dispute.
What remains unresolved is what it means.

This is not a failure of experiment.
It is a failure of sense-making.

And sense-making has never been an institutional strength.

Where AI Actually Fits (and Where It Doesn’t)

I want to be explicit: I still have a long way to go in how I use AI — intellectually, methodologically, and ethically.

AI is not an oracle.
It does not “solve” physics.
It does not replace belief, responsibility, or judgment.

But it changes something fundamental.

AI allows us to:

re-analyze vast datasets without institutional friction,
explore radical ontological assumptions without social penalty,
apply sustained logical pressure without ego,
revisit old experimental results with fresh conceptual frames.

In that sense, AI is not the author of new physics — it is a furnace.

It does not tell us what to believe.
It forces us to confront the consequences of what we choose to believe.

Making Sense of What We Already Know

The most exciting prospect is not that AI will invent new theories out of thin air.

It is that AI may help us finally make sense of experimental data that has been sitting in plain sight for decades.

Now that CERN data is increasingly public, the bottleneck is no longer measurement. It is interpretation.

AI can help:

expose hidden assumptions in standard models,
test radical but coherent ontologies against known data,
separate what is measured from how we talk about it,
revisit old results without institutional inertia.

This does not guarantee progress — but it makes honest failure possible. And honest failure is far more valuable than elegant confusion.

Between Institutions and Insight

This is not an AI-versus-human story.

It is a human-with-tools story.

Institutions will continue to do what they do best: build machines, refine measurements, and preserve continuity. That work is indispensable.

But understanding — especially ontological understanding — has always emerged elsewhere:

in long pauses,
in unfashionable questions,
in uncomfortable reinterpretations of existing facts.

We are entering such a pause now.

A Quiet Optimism

I do not claim to have answers.
I do not claim AI will magically deliver them.
I do not even claim my current ideas will survive serious scrutiny.

What I do believe is this:

We finally have the tools — and the historical conditions — to think more honestly about what we already know.

That is not a revolution.
It is something slower, harder, and ultimately more human.

And if AI helps us do that — not by replacing us, but by challenging us — then it may turn out to be one of the most quietly transformative tools science has ever had.

Not because it solved physics.

But because it helped us start understanding it again.