Quantum Sphere: The Extinction Paradox That Defies Everyday Physics

A solid ball should block its own shadow, but a quantum simulation shows it blocks four times that at low energy and twice at high energy, never once matching the classical answer.

5:29 video5 min readWatch on YouTube

Here is a question that sounds like it should have a boring answer. Fire a beam of particles at a solid ball. Nothing gets through it. How much of the beam does the ball actually block? In classical physics, the everyday kind, the answer is a one-liner: it blocks exactly its own shadow, at every energy, forever. In quantum mechanics, that answer turns out to be wrong, not slightly off in some edge case, but wrong at every energy level there is, and wrong in two different ways that do not even agree with each other.

Breaking the incoming wave into pieces

The tool for measuring this is a scattering cross-section, a physicist's way of describing how big a target looks to an incoming beam versus how big it actually is. The incoming particle behaves as a wave, and the useful move is to break that wave into pieces, called partial waves, one for each way it can approach the ball: dead center, slightly off center, further off still. The ball knocks each partial wave out of step, and adding up that disruption gives the answer. Nothing here is approximated: because nothing can get through the solid ball, the wave must drop to exactly zero at its surface, and that single requirement pins down every piece of the calculation exactly. The entire setup reduces to one dial, the ball's size divided by the wavelength of the incoming particle, and the number of partial waves that actually matter turns out to be roughly whatever that dial reads.

Calibrating against an independent build

Before trusting any of the readings, the instrument had to be checked against something built independently. That calibration came from asking Claude to build an interactive browser simulation of the same physics, using nothing but a single prompt. With the dial set to 13.6, the Claude-built simulation read 2.328, and a separately written version of the same code, sharing no logic with the browser version, read 2.3283, agreement on every digit printed. The one place they briefly disagreed was cosmetic, a missing pi symbol on an axis label, not the underlying physics.

Reading one: four times the shadow

With the dial turned hard left, toward long wavelengths and a tiny ball, the reading comes out to exactly four times the everyday classical answer, accurate to eight decimal places. At this end of the scale, a single partial wave, the dead-center one, produces all of that effect on its own. The ball blocks four times its own shadow, not four-ish, four.

Reading two: the extinction paradox

Turning the dial hard right, toward short wavelengths and a large ball with hundreds of partial waves contributing, is supposed to be the regime where ordinary physics finally takes over, since a tiny wavelength against a big object is exactly the case where quantum and classical predictions are expected to converge. Instead, the reading settles at exactly twice the classical answer, never once at one. The extra amount turns out to come from the shadow itself: a sharp-edged shadow is not something waves produce for free, since waves naturally spread out. Producing a clean shadow requires bending waves in behind the ball to cancel out what would otherwise fill in that space, and that bending counts as scattering in its own right. Adding up how much bending it takes to build a clean shadow comes to exactly one more shadow's worth, on top of the ball itself. That is the extinction paradox, and the rule that makes the numbers work out is called the optical theorem.

How close does it actually get to two?

Across every energy scanned, the quantum answer never touches the classical answer of one; it moves from four down toward two but never reaches either integer exactly. The closest observed value across the whole scan was 2.199. Getting within 1% of two requires the dial to reach roughly a thousand, meaning a ball a thousand wavelengths across, already enormous by this measure, and even then the reading sits at 2.09, still about four and a half percent high. When a textbook states that the high-energy answer is simply two, that statement describes a limit, not something anyone has actually measured directly, since no real experiment reaches that regime.

The settling gauge and the soft target

A third reading tracked how quickly the ratio approaches two as the dial increases, settling at a rate consistent with two-thirds, starting at 0.63 and reaching 0.6659 by the time the dial hits 10,000, moving the same way at every step in between, though this rate is described as measured rather than formally proven, since no source deriving it directly could be found. A separate check asked whether the reading wobbles on its way from four down to two, and it does not, not a single bump across the entire range, even though a sum built one partial wave at a time might be expected to ring. The explanation is that nothing gets through a solid ball, so there is no interior for anything to get trapped inside, and no trapping means no ringing. Swapping the solid ball for a soft target, something particles can actually enter, breaks that smoothness immediately: the same code under the same settings produces ten distinct spikes where the solid ball produced zero, confirming the smoothness belonged to the physics of the solid ball, not to the code itself.

Key takeaways

  • A quantum mechanical ball blocks exactly four times its classical shadow at low energy (long wavelength), accurate to eight decimal places.
  • At high energy (short wavelength), the quantum answer settles at exactly twice the classical shadow, never converging to the classical prediction of one.
  • The extinction paradox explains the extra factor of two: producing a clean shadow requires bending waves behind the ball, which itself counts as scattering.
  • The optical theorem is the rule that accounts for this extra scattering contribution.
  • Reaching within 1% of the theoretical high-energy limit of two requires a ball roughly a thousand wavelengths across, a regime no real experiment has actually measured.
  • Switching from a solid ball to a soft, enterable target introduces ten distinct spikes in the reading, where the solid ball produced a perfectly smooth curve.

Try it yourself

Before touching any code, try placing two everyday cases on the same dial: slow neutrons scattering off an atomic nucleus, and red light scattering off a droplet of fog. Then, with the full simulation, plot both a solid ball and a soft target of depth 900 on the same axis and see whether your results match the described pattern, a smooth curve from four to two for the solid ball and roughly ten spikes for the soft target, first appearing near a dial value of 2.88. The full simulation runs in about 200 lines of plain Python with nothing to install. This project is part of the Humanitarians AI Fellows program.

Chapters

  1. 0:00Classical shadow vs. Quantum reality
  2. 0:50Breaking waves down: Understanding partial waves
  3. 1:40Calibrating the instrument against a Claude-built browser simulation
  4. 2:25Low Energy (Reading 1): Why a tiny ball blocks 4x its shadow
  5. 3:10High Energy (Reading 2): The extinction paradox and the optical theorem
  6. 4:15The Settling Gauge (Reading 3): Measuring the 2/3 decay rate
  7. 5:05Hard Ball vs. Soft Target: Why entering a target creates spikes
Full transcript(auto-generated, with timestamps)

Classical shadow vs. Quantum reality

[0:00]This is humanitarians. Okay. A question that sounds like it has a boring answer. You fire a beam of particles at a solid ball. Nothing gets through it. How much of the beam does the ball block? In classical physics, the everyday kind, it's a oneliner. It blocks its own shadow. Same answer at every energy forever. In quantum mechanics, that answer is wrong. Not slightly off, not in some strange corner. Wrong at every energy there is. And the two ways it's wrong don't even agree with each other. Hi, this is Tanme KC Carney in for humanitarians AI. This one's about scattering cross-sections. A physicist's way of saying how big a target looks to an incoming beam versus how big it actually is. Here's the plan. Build the instrument. Check it against a simulation. Claude writes, then take three readings. The third one's where it gets interesting. Let's build something to measure it. It's really just a sum. The incoming particle is a wave, and it helps to break that wave into pieces, one for each way it can come in dead

Breaking waves down: Understanding partial waves

[0:50]Center at the ball, slightly off center further off. Physicists call those partial waves. The ball knocks each one out of step. Add up the knocking and you have your answer. And nothing here gets approximated. Nothing gets through the ball. So the wave must drop to zero at its surface. And that one requirement pins every piece exactly. One dial on the whole thing. Ball size divided by wavelength. K* A. Turn it. Take a reading. One marking on this dial worth knowing first. The number of waves you have to add up is roughly whatever the dial says. Don't take that on faith. When k times is 100, the number that matters is 102 at 2020, which makes sense. A wave coming in too far off center just misses the ball. So you can place any target on this dial before computing a thing and we'll do exactly that at the end. Two more gauges here. Whether the reading settles and whether it wobbles getting there, both unread for now. You don't trust an instrument

Calibrating the instrument against a Claude-built browser simulation

[1:40]Until it agrees with one somebody else built. So calibration. One prompt to claude. Build an interactive simulation of this one web page. Physics in the browser. It comes back working with the dial at 13.6. Claude simulation reads 2.328. My own code written separately sharing nothing says 2.3283. Agreement on every digit it prints. The physics is right. And honestly, it would have been an easier story if it weren't one knit and it's the label, not the physics. There's a pi missing off the axis. Reading one dial hard left long wavelength tiny ball four. And I mean four, not fourish, accurate to eight decimal places. Four times the everyday answer and the dead center wave produces all of it. One wave. The ball blocks four times its own shadow. Reading two. Dial hard right. Short wavelength big

Low Energy (Reading 1): Why a tiny ball blocks 4x its shadow

[2:26]Ball. Hundreds of waves. Two. Now hang on. This is the end of the dial where ordinary physics is supposed to work. Tiny wavelength big object. The easy case where the quantum answer should collapse onto the everyday one. It doesn't. It parks at exactly twice. So where's the extra bit coming from? Behind the ball there's a shadow. And a sharpedged shadow isn't something waves do for free. waves spread. To carve a clean one, you have to bend waves in behind the ball, so they cancel what would otherwise be there. But bending is scattering. Those particles change direction. They count. Add up the bending it takes to build that shadow. And it's exactly one more shadow's worth. The ball plus the shadow it casts, too. That's the extinction paradox. The rule that makes it add up is the optical theorem. Both readings on one axis with the everyday answer underneath. Left end four, right end

High Energy (Reading 2): The extinction paradox and the optical theorem

[3:10]Two. Ordinary physics, one flat all the way across. Four, then two, never one. Across every energy we scan, the quantum answer never touches the everyday one. Closest it gets is 2.199. The ball is never the size it looks. Reading three, the settling gauge. Does it ever actually settle on two? I measured how fast it closes the gap. That rate climbs decade after decade. Starts at 0.63. By the time the dial reads 10,000, it's 0.6659. Every step in between moves the same way. Consistent with 2/3 and nothing else nearby. But measured, not proved. I couldn't find a source that deres it so it stays flagged. And here's what

That costs you. The part that changed how I read these graphs. To get within 1% of two, the dial has to read about a thousand. At 100 a ball 100 wavelengths across, already enormous. You're still at 2.09 4 1/2% high. So when a textbook says the high energy answer is two, that's a limit, not a measurement. Nobody has run that experiment. One gauge left. Does the reading wobble on the way down? No. Not one bump the whole way from four to two, which surprised me. A sum that switches on one wave at a time. Sounds like it ought to ring. The reason is the ball. Nothing gets through

The Settling Gauge (Reading 3): Measuring the 2/3 decay rate

[4:16]It. So there's no inside and nothing gets trapped in something it can't get into. No trapping, no ringing. Lovely story. And lovely stories are the ones you should try to break. Same code, same settings, but a soft target this time. Something particles can get inside. 10 spikes. Zero for the solid ball. 10 for the soft one. That smoothness belong to the target, not my code. Your turn. Two of them. First 30 seconds. No code. Put these on the dial yourself. Slow neutrons on an atomic nucleus and red light on a droplet of fog. Where does each land? answers on screen. Then the real one with a greater. Plot both curves on the same axis. The solid ball and a soft target of depth 900. Prompts on screen. You've got it. If the solid ball reads 2.328 at 13.6 and falls without a single bump. The soft target should give about 10 spikes first near 2.88. Broken. If the solid ball wobbles,

Hard Ball vs. Soft Target: Why entering a target creates spikes

[5:05]Your sum cut off too early, or the shortcut for these waves falling apart past the dials number, it never raises an error. One dial, count the waves, read the answer straight off it, then check whether the limit you're quoting is one anybody could reach. Every number here came from about 200 lines of plain Python. Nothing to install. It's in the description. Go break it. This is Tanme Kulcarnney for Humanitarians AI.

More from Humanitarians AI Fellows

Humanitarians AI Lyrical Literacy Project