Physics Ran an Experiment on TIME — The Results Don't Add Up

ELI5 / TLDR

We just built the most accurate clock in human history, and the first thing it whispered back was that the fixed numbers holding physics together may not actually be fixed. A century of other experiments says the same thing from different angles. Clocks run at different speeds depending on where you stand. The direction of time appears to come from the act of watching, not from the laws of physics. And the most complete equation we have for the universe has no time in it at all. We know a lot about how time behaves. We still have no idea what it is.

The Full Story

The clock that told on physics

In July 2025, the National Institute of Standards and Technology in Boulder finished building an aluminum ion clock accurate to the 19th decimal place. The previous record was the 18th. That one extra digit was not a rounding flex. It was the point at which the instrument became sharp enough to check whether the so-called fundamental constants of nature — the fine structure constant, the gravitational constant, the fixed numbers threaded through every equation in physics — are actually fixed.

Early indications suggest they might be drifting. Slowly. Almost imperceptibly. But measurably.

If the constants are changing, what are they changing with respect to? What is the thing they are moving through? And if that thing is time, what is time?

That is the question the whole video circles. And the uncomfortable answer is: nobody really knows.

Physics, for all its precision, does not have a clean definition of time that isn’t circular. Newton called it absolute. Einstein dismantled that in 1905 and replaced it with a description of how time behaves — stretching near mass, compressing at speed — without ever saying what it is. Carlo Rovelli puts it plainly. We know an enormous amount about how time behaves. We do not know what it is.

Clocks on planes, clocks in orbit, clocks one millimeter apart

Think of time like water. You’d assume there’s one ocean and all of us are bobbing on its surface at the same rate. The experiments keep showing that isn’t true.

In 1971, two physicists named Hafele and Keating strapped cesium atomic clocks into commercial passenger flights, flew them around the world, and compared them to identical clocks that stayed on the ground. The airborne clocks came back about 59 nanoseconds behind. Not broken. Just different. They had measured a slightly different amount of time than the ones that stayed still.

Your phone relies on this every second. GPS satellites orbit at about 20,000 km. Their clocks, moving fast and sitting higher in Earth’s gravitational field, run about 38 microseconds per day faster than clocks on the ground. If engineers didn’t correct for that daily drift, your phone’s sense of where you are would walk off by roughly 11 km a day. Navigation would be useless within hours. The weirdness is already baked into your commute.

Then in 2021, a team at JILA in Colorado measured a time difference between two layers of atoms separated by one millimeter. The thickness of a credit card. The atoms at the top aged ever so slightly faster than the ones at the bottom, exactly as general relativity predicts for objects at different heights in a gravitational field. Right now, the top of your head is aging faster than your feet. The difference over a human lifetime is too small to notice. But it is not zero.

If time is that sensitive, that local, that dependent on exact position, then what exactly is the thing we are measuring when we say time?

The arrow that isn’t in the physics

Here is the part that has quietly haunted physics for 130 years. Take any fundamental equation — Newton, Maxwell, Schrödinger, Einstein. Run it forward in time. Now run it backward. The math works identically in both directions. The universe, as far as its own equations are concerned, has no preference.

So why does everything feel like it moves forward? Why do broken eggs never reassemble? Why is the past fixed and the future open?

Ludwig Boltzmann nearly broke himself trying to answer this in 1895. His conclusion: the arrow of time isn’t a law. It’s a statistical bias. Think of shuffling a deck. There are way more messy orderings than neat ones, so random shuffling drifts toward mess. The universe started in an extraordinarily ordered low-entropy state, and everything we call “the flow of time” is just the slow drift from that tidy beginning toward mess.

Which raises a question nobody has answered. Why did the universe start that tidy in the first place?

Observation as a thermodynamic event

This is where it gets strange. In 1961, Rolf Landauer at IBM proved that erasing a single bit of information — resetting a switch, clearing a memory register — must release a tiny amount of heat. Not because our computers are sloppy. Because the laws of thermodynamics require it. Information is physical. Memory is not free.

Stack that against Boltzmann. The arrow of time comes from entropy increasing. Every act of observing, recording, or remembering necessarily increases entropy. Which means observation is not a passive window onto reality. Observation is a thermodynamic event. It produces heat. It generates the very irreversibility that distinguishes past from future.

In November 2025, a team at Oxford took this logic into the lab. They built a microscopic clock out of a double quantum dot — two tiny regions an electron can hop between. Each hop is a tick. Then they asked a small, devastating question: how much energy does it cost to run the clock versus how much it costs to read it?

Reading the clock required up to a billion times more energy than running it. Not twice as much. Not ten times. A billion. And when they analyzed where that energy went, it was producing entropy. When the two quantum dots were perfectly balanced, the electron hopped forward and backward with equal probability — the clock had no direction. The arrow only appeared when something outside the system built a memory of what had happened.

The arrow is in the record, not in the electron.

Imagine a universe with no memories, no records, no structures that retain information about previous states. Electrons jumping. Energy moving. Things happening. But no arrow of time. No past. No future. Just an eternal symmetric present. That is what the Oxford experiment implies.

Quantum mechanics smuggled Newton back in

In every other area of physics, space and time are partners. General relativity weaves them into spacetime, a single fabric that bends and curves. But in quantum mechanics, space is an observable — something the system itself possesses, which you can measure and put inside equations. Time is not. It is a parameter. A dial you set from outside before running the experiment. The Schrödinger equation treats time as a uniform external thing that flows in the background, untouched by anything the quantum system does.

That is almost exactly what Newton said in 1687. Absolute time, flowing equably, independent of external. Quantum mechanics, for all its revolutionary strangeness, quietly reimported Newton’s time through the back door.

In 2020, a team in Vienna led by Julia Rubino pushed on this. They set up a quantum particle to pass through two operations, A and B, where the order of those operations was itself in quantum superposition. A-before-B and B-before-A happening simultaneously for the same particle. And the experiment worked. Causality — the principle that causes precede effects — turned out not to be a fundamental feature of quantum reality. It’s something that emerges at larger scales the way temperature emerges from the motion of molecules.

The equation where time disappears

If you try to write down the quantum mechanics of the entire universe — not a particle, not an atom, the universe — you run into a tension between general relativity (which describes how spacetime geometry evolves in time) and quantum mechanics (which describes how states evolve in time). Both need a single combined equation.

In 1967, John Wheeler and Bryce DeWitt wrote one. And time disappeared. Not approximately. Not in some limiting case. The variable T that appears in every other equation in physics was simply absent from the final result.

The universe as a whole, described by the most complete equation physicists could construct, does not evolve in time. It just is. Static. Timeless. One quantum state with no before and no after.

This isn’t a fringe result. The Wheeler-DeWitt equation is the starting point for every serious attempt at quantum gravity. And it says time is not a fundamental feature of the universe. Time has to emerge from a timeless underlying reality. The question is how.

Time as something you only feel from inside

In 2013, a team led by Ekaterina Moreva in Turin tested an idea Page and Wooters proposed in 1983. They built a tiny quantum system of two entangled photons. One acted as a clock. The other acted as an observer of that clock.

Then they looked at the system two ways.

From the outside, treating the whole two-photon system as a single object — the system appeared static. Timeless. Exactly what Wheeler-DeWitt predicts.

From the inside, using one photon to measure the other — time existed. The clock ticked. Events had sequence.

Same system. Same physical setup. Two completely different realities depending on where you stood.

Time may not be a property of the universe. It may be a property of being inside the universe.

And here’s the twist: the universe, by definition, has no outside. No observer can ever occupy the external perspective where time is absent. The only description any real observer can ever access is the internal one. The one where time exists.

Time you can engineer a boundary in

In December 2025, Hadiseh Moussa’s team at City University of New York demonstrated something predicted in the 1960s but never cleanly done. They built a system that reflects electromagnetic waves in time rather than in space. They suddenly changed the electrical properties of the medium a wave was passing through, and the wave partially reflected — not backward through space, but backward through time. A time-reversed copy of the signal began retracing its own history.

What this revealed isn’t exotic new physics. It’s that Maxwell’s equations — the most tested relationships in physics — already contain the capacity for time reversal. The asymmetry we experience, the fact that waves propagate into the future and not the past, isn’t written into the laws of electromagnetism. It’s written into the initial conditions. The choice of a uniform starting medium.

Observation can also freeze time

There’s a flip side to the Oxford result. It’s called the quantum Zeno effect, first proposed in 1977 and confirmed many times since. A quantum system — a decaying particle, an atom between energy levels — evolves on some characteristic timescale. But if you measure it frequently enough, the evolution slows down. In the limit of continuous observation, the system stops evolving entirely. You can hold a quantum system in its current state indefinitely simply by measuring it fast enough.

Observation creates the arrow of time. Observation also controls its rate. And both cost energy (Landauer). Which means controlling time at the quantum level has a thermodynamic price tag. You pay to slow it. You pay to give it direction.

The cracks in cosmic time

While quantum experiments were pulling on time from below, cosmology has been pulling on it from above.

The Hubble constant — the rate at which the universe is expanding — should be one number. Two independent methods keep giving two different answers. The cosmic microwave background says 67.4. Measurements using Cepheid variable stars say 73.2. The gap is five sigma, which in physics is the threshold for calling something a discovery.

In 2024, the Dark Energy Spectroscopic Instrument added something worse. Dark energy — the force driving cosmic expansion, the thing Einstein called the cosmological constant — may not be constant. Early data suggests it’s weakening. If the engine driving cosmic expansion is changing, then the rate at which cosmic time flows is also changing. Our model of the universe’s 13.8-billion-year timeline is built on assumptions that may be wrong.

The BaBar result nobody talks about

In 2012, the BaBar experiment at SLAC measured a direct asymmetry in the transformation rates of B mesons into their antiparticles. State A going to state B happened at a different rate than state B going back to A. In 2020, the T2K experiment in Japan found a related asymmetry in neutrinos. Both are real. Both are statistically robust. Both show that the arrow of time isn’t just a statistical tendency at large scales — it’s built into the behavior of individual particles at a structural level. The standard model can describe this asymmetry mathematically but cannot explain why it exists from first principles.

Rovelli vs Smolin — two honest readings of the same data

Carlo Rovelli concludes from all of this that time is not a river. It’s more like heat. Heat doesn’t exist in a single molecule. It emerges from the collective behavior of enormous numbers of particles interacting. Remove the large numbers and heat loses its meaning. Time, Rovelli argues, works the same way. At the level of fundamental physics, there is no time. Time is what the universe looks like from the inside when it is big enough and complex enough to have a thermodynamic arrow.

Lee Smolin, looking at the same data, reaches the opposite conclusion. If the laws of physics are themselves products of cosmic history — if they evolve, if black holes spawn new universes with slightly different constants, if the laws we live under are the ones that survived — then time must be prior to the laws. You cannot derive time from laws that are themselves inside time. Time has to be the one thing that doesn’t emerge from anything else.

Two serious physicists. Identical experimental results. Incompatible conclusions. That disagreement is the honest summary of where the field actually stands.

David Albert and the past hypothesis

David Albert identifies what he calls the past hypothesis — the single unexplained assumption doing all the work in our experience of time. The universe began in a state of extraordinarily low entropy. Remove that one fact and the arrow of time disappears entirely. Memory becomes meaningless. Causality collapses.

Every experience you have ever had — every memory, every plan, every sense that there was a yesterday and will be a tomorrow — rests on one initial condition that nobody chose and no law required. Penrose tries to explain it with cyclic cosmology (the ordered beginning was the disordered end of a previous eon). Hawking tries to dissolve it with imaginary time (the beginning was a smooth transition from a state where time and space were equivalent). Both push the question back one step. Neither closes the circle.

It from bit

John Wheeler, late in his life, kept pushing a phrase most of his colleagues found too strange to engage with: “it from bit.” The idea that information isn’t a description of physical reality — it’s what physical reality is made of. Every particle, every field, every physical quantity derives its existence from answers to yes or no questions.

If that’s right, then the question of who or what is doing the asking isn’t a side issue. A bit is not a bit in the absence of something that distinguishes between its two states. Wheeler’s framework makes the existence of observers a necessary feature of reality rather than an accidental one. And if observation is what generates the arrow of time, as the Oxford experiment suggests, then the observer is not just a witness to time. The observer is part of what produces it.

Wheeler’s delayed choice experiments went one step further. They suggested the act of observation can, in a precise quantum mechanical sense, reach backward and determine how a particle behaved at an earlier moment. The past in those experiments is not fixed until something in the present forces it to be fixed. Which means the past might not be a record of what happened. It might be a construction assembled backward from the present, shaped by every act of observation that has ever occurred.

Where we actually are

We have built the most accurate clocks in the history of measurement, and they are telling us the constants of nature may not be constant. We have experiments showing the arrow of time is not in the laws of physics but in the act of observation. We have a mathematical description of the entire universe that contains no time variable. We have laboratory evidence that causality is not absolute, that quantum transitions have durations determined by geometry rather than the flow of time, and that time can be reflected at an engineered temporal boundary. We have two cosmological measurements of the same quantity disagreeing by five sigma. We have one theoretical framework in which time emerges from thermodynamic complexity and a competing framework in which time is the only fundamental thing, and both remain consistent with the data.

And we have an unsolved problem at the center of all of it. Why did the universe begin in such an ordered state.

What we do not have is a clear, consistent, experimentally confirmed account of what time actually is.

Key Takeaways

The 2025 NIST aluminum ion clock is accurate to 19 decimal places (41% more precise than anything before) and is now sensitive enough to detect whether fundamental constants like the fine structure constant are drifting over time.
Time is not uniform. Clocks run measurably faster at higher altitudes. GPS applies a 38-microsecond-per-day correction; without it, navigation would be off by ~11 km per day.
In 2021 at JILA, a time difference was measured across atoms separated by just 1 mm — the top of your head ages faster than your feet.
The fundamental equations of physics are time-symmetric. Nothing in the laws themselves requires forward flow. The arrow of time is, per Boltzmann, a statistical consequence of the universe having started in a very ordered low-entropy state.
Landauer (1961): erasing one bit of information releases heat. Information is physical. Memory isn’t free.
The Oxford 2025 double quantum dot experiment: reading the clock cost up to a billion times more energy than running it. The arrow of time emerges in the act of measurement, not in the quantum system itself.
Quantum mechanics treats time as an external parameter, not an observable — essentially Newton’s absolute time. This is incompatible with general relativity, where time is a dynamic dimension.
The Wheeler-DeWitt equation (1967): when you write the quantum mechanics of the whole universe, the time variable vanishes. The universe as a whole does not evolve.
The Moreva 2013 experiment: a two-photon system looks timeless from the outside and time-flowing from the inside. Time may be a property of being inside a system.
Vienna 2020: causal order itself can be placed in quantum superposition. A-before-B and B-before-A can happen simultaneously for the same particle.
The quantum Zeno effect: observing a quantum system frequently enough freezes its evolution. Observation controls both the arrow and the rate of time.
Moussa 2025 at CUNY: demonstrated a temporal boundary that reflects waves backward through time. Time reversal was already implicit in Maxwell’s equations — the asymmetry is in our initial conditions.
BaBar (2012) and T2K (2020): time asymmetry is not just statistical but structural, built into how individual particles behave. No first-principles explanation exists.
The Hubble tension: CMB gives 67.4, Cepheids give 73.2, 5-sigma gap. DESI (2024) suggests dark energy may be weakening. Cosmic time itself may not be flowing at a constant rate.
Penrose’s conformal cyclic cosmology: the low-entropy Big Bang is the geometrically-equivalent continuation of the high-entropy end of a previous eon. Time loses scale when only massless particles remain.
Hawking’s no-boundary proposal: via Wick rotation, the four dimensions become equivalent and there is no temporal beginning — just a smooth geometric transition into spacetime.
Rovelli’s view: time is like temperature. It emerges from thermodynamic complexity and has no meaning at fundamental scales. Smolin’s view: time is the only fundamental thing because the laws themselves evolve.
David Albert’s “past hypothesis”: every sense of a past you’ve ever had rests on one unexplained low-entropy initial condition. Remove it and time’s direction disappears.
Black holes may destroy information (Hawking radiation is thermal). If so, the past itself becomes ambiguous wherever something has crossed an event horizon.
Scott Kelly returned from the ISS ~7 milliseconds younger than his twin. There is no universal “now.” The relativity of simultaneity makes the present a local feature of a reference frame, not a global fact.
Wheeler’s “it from bit”: information is the substrate of reality, not a description of it. Observers may be a necessary feature, not accidental.

Claude’s Take

This is one of the better “state of the field” physics videos I’ve seen — patient, honest about what’s unresolved, willing to sit with uncomfortable conclusions rather than resolve them artificially. The voice (presumably scripted, narrated, maybe partially AI-assisted given the Acronium channel’s style) is unusually slow and reflective for a YouTube video on fundamental physics. It doesn’t posture. It doesn’t oversell. When it doesn’t know, it says so.

The experimental results cited are real and mostly well-chosen. Hafele-Keating, Pound-Rebka, JILA’s millimeter-scale gravitational time dilation, BaBar’s T-violation, the Moreva-Page-Wooters test, the Vienna causal superposition work, Landauer’s principle, Wheeler-DeWitt — these are all legitimate and fairly represented. The Oxford double quantum dot clock is a real 2023 paper by Ares and collaborators (the video’s date of November 2025 may be slightly off or refer to a follow-up). The NIST 19-decimal-place aluminum ion clock is real (July 2025). DESI’s dark energy results and the Hubble tension are ongoing real puzzles. The Moussa temporal boundary work is genuinely a recent result, though “reflecting in time” is a somewhat dramatic framing of a more modest photonic result.

Where I’d push back mildly: the interpretive glue between experiments is stronger than the experiments themselves justify. The Oxford result doesn’t prove the arrow of time requires observers. It shows measurement is thermodynamically expensive and that irreversibility lives in the record. The Moreva experiment is a proof-of-principle for a very small two-qubit toy system — extrapolating “time is only real for inside observers” from it to the whole universe is a philosophical move, not a measured one. The video is mostly careful to flag this, but a viewer could walk away thinking consciousness literally creates time, which is stronger than the physics supports.

The drift of fundamental constants is also still genuinely contested. NIST’s clocks are precise enough to look, but positive detections of drift are not established. Presenting it as a near-confirmed result overstates what’s known.

Still, the overall picture — that general relativity and quantum mechanics describe time in structurally incompatible ways, that the arrow of time is an open problem at the level of first principles, that the Wheeler-DeWitt equation genuinely contains no time variable, and that the cosmological tensions are real and unresolved — is all accurate. The framing of “we know how time behaves, we don’t know what it is” is the most honest summary available.

Score: 8/10. A notch below the best physics explainer work because the editorial voice occasionally conflates interpretation with result, but genuinely rare for how carefully it refuses to pretend this is settled.