An interactive explainer, inspired by Adam Becker's What Is Real?
Quantum mechanics predicts experiments perfectly, yet no one agrees on what happens when you look: does observation change reality, split it, or merely reveal it?
Fire particles one at a time at a barrier with two slits. With no one watching the slits, single particles paint an interference pattern, as if each went through both slits as a wave. Turn on a which-path detector and the stripes vanish into two plain bands. This page draws the physics directly on a 2D canvas because the experiment itself is flat: source, slits, screen.
Idle: the screen is blank. Fire a particle to begin.
Reduced motion detected: particles land instantly without flight animation.
The math is settled; its meaning is not. Drag each slider to step through one measurement event, frame by frame.
What it claims: before measurement there is only a wave of possibilities; the act of measuring collapses it to one definite outcome.
What it costs: "measurement" is never defined. Where does the quantum world end and the classical one begin?
What it claims: nothing collapses. Every outcome happens, each in its own branch of a splitting universe. You are one branch of you.
What it costs: an unimaginable number of parallel worlds, and hard puzzles about why outcomes have the probabilities they do.
What it claims: particles are real and always have positions; an invisible guiding wave steers them. Measurement just reveals where the particle already was.
What it costs: the guiding wave acts instantly across any distance, sitting uneasily with relativity's speed limit.
A single qubit sits in superposition: 70% chance of reading 0, 30% chance of reading 1. Its state is drawn as a vector tilted between the poles of a sphere. Press Measure and watch each interpretation tell a different story about the same click.
The vector leans toward 0. Copenhagen: it will snap to a pole when measured. Many-Worlds: both poles will occur, in branches weighted 70/30. Pilot-wave: the answer is already fixed; we just do not know it yet.
Five quick questions. Instant feedback, no scoreboard shame — 0/5 correct so far.
What Is Real? by Adam Becker tells the human story behind everything on this page: how Bohr's Copenhagen view became orthodoxy, how Bohm and Everett were sidelined for asking what the math means, and why the question is roaring back. If this page made you curious, the book will make you obsessed. It earns its spot on any list of books to fall in love with science.
This educational measurement explainer samples screen positions using an illustrative probability weight. With the detector off, it uses cosine squared of point zero seven five times offset, multiplied by exponential of minus offset divided by one hundred ten squared. At center offset zero the weight is one. At offset twenty point nine four four the cosine term is zero, producing a dark fringe. At forty-one point eight eight the cosine squared returns to one, but the envelope lowers the weight to about point eight six five zero one five. Common four hundred pixels per weight unit makes bars four hundred, zero and about three hundred forty-six. The offsets are source canvas coordinates, not physical slit distances or photon wavelengths. Rejection sampling paints a toy interference distribution, not a solved experiment. The Bloch-style display leans toward outcome zero with source probability point seven, leaving point three for outcome one. Common four hundred pixels per probability unit gives bars two hundred eighty and one hundred twenty, totaling four hundred. Its initial polar angle is arccos of point four, about one point one five nine three radians: cosine theta equals two times point seven minus one. The displayed vector is a decorative two-dimensional projection. Each Measure click independently tests a new random number against point seven, even if the preceding display already shows a pole. Prepare resets only that picture and text, not the counters. Thus repeated clicks are fresh seventy-thirty trials, not a faithful repeat measurement of the same collapsed quantum state. The total counter always equals the two outcome counters together. The Many-Worlds panel is a diagram scrubber, not a count of measured physical universes. Its progress m runs from zero to one. The first two branches extend by fraction minimum of one and two m. The next four extend only after halfway, by fraction maximum of zero and two m minus one. At m zero only the trunk is drawn; at one half the trunk and first two branches are complete; at one there are seven line segments in total. Here a common two display pixels per original SVG coordinate preserves the source endpoint geometry. These branching timings are authored illustrations. Likewise the Copenhagen and pilot-wave panels are visual metaphors, not implementations of competing quantum dynamics or evidence that distinguishes interpretations.