Physics Ran An Experiment On Time

---TRANSCRIPT--- In July 2025, physicists at the National Institute of Standards and Technology built the most accurate clock ever made. It measures time to the 19th decimal place. To put that in perspective, the previous record was the 18th decimal place, and that already felt like an absurd level of precision. But here is the part that nobody talks about when they report this story. The moment this clock became accurate enough to actually matter, it revealed something uncomfortable. The fundamental constants of nature, the fixed numbers that hold all of physics together, might not be fixed. They might be drifting. Slowly, almost imperceptibly, but measurably. So you build the most precise instrument in human history,

[music] and the first thing it tells you is that the ground underneath physics is not as solid as you thought. That is [music] where we are starting tonight. Not with a theory, with a measurement, and with a result [music] that does not add up. Before we go further, there is something worth sitting with. What is a clock actually measuring? Not how, we know how. Pendulums, atomic oscillations, electron transitions, but what? When you glance at your watch, you assume it is reading some property of the universe that exists [music] independently of the watch itself. Something called time. Something that flows, that passes, that cannot be stopped or reversed. And that assumption feels so obvious it barely registers as an assumption at all. But here is the thing. Physics does not have a clean definition of time that is not circular. Every formal definition either uses time to define time or borrows from everyday experience and dresses it up in mathematics. Newton called it absolute, flowing uniformly everywhere, independent of anything external. Einstein dismantled that in 1905. But what Einstein replaced it with is not a definition either. It is a description of how time behaves under certain conditions. How it stretches, how it compresses, how it relates to space. That is different from knowing what it is. Carlo Rovelli, one of the physicists who has spent more time on this question than almost anyone alive, put it plainly. We know an enormous amount about how time behaves. We do not know what it is. Those are not the same thing, and that gap between behavior and identity is exactly where the experiments start producing results that nobody [music] expected. So when that NIST clock suggested the constants might be drifting, the unsettling part was not the drift itself. It was the question underneath it. 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? The first crack showed up in 1971, and it was not subtle. Joseph Hafele and Richard Keating loaded cesium atomic clocks onto commercial passenger flights, not specialized research aircraft, just regular planes, [music] flew them around the world, and compared them to identical clocks that stayed on the ground. The clocks that flew came back different. Not broken, not malfunctioning, just different. They had measured a slightly different amount of time than the ones that stayed still. The airborne clocks lost about 59 nanoseconds relative to the ground clocks, [music] matching almost exactly what Einstein’s general and special relativity predicted. Almost exactly. There were small discrepancies which the experimenters noted carefully, [music] and which have been quietly interesting to a handful of physicists ever since. But the main result was clear. Time is not the same everywhere. Move through space, accelerate, change your position in a gravitational field, and you move through time at a different rate than someone who [music] stayed put. This is not a metaphor. This is not an approximation. Two objects, same starting point, same ending [music] point, different amounts of elapsed time. And if you are the kind of person who finds that straightforward, [music] yes, relativity, we know this. I want to ask you something. Do you actually feel the weight of what that means? Because what it means is that there is no single [music] universal clock ticking in the background of reality. There is no master time that everyone shares. There are only local clocks, local rates, local experiences of duration. The universe does not have one timeline. It has as many timelines as it has objects moving through it. That was

What came after only made it stranger. Here is the part that makes this impossible to dismiss as abstract [music] physics. Your phone knows where you are right now because of GPS satellites orbiting at about 20,000 km altitude. Those satellites carry atomic clocks, and those clocks, because they are moving fast and sitting higher in Earth’s gravitational field, run faster than clocks on the ground, by about 38 microseconds per day. 38 microseconds sounds like nothing. Less than a blink, less than anything your nervous system can register. But GPS works by calculating your position from tiny differences in the arrival times of signals from multiple satellites. The math is exquisitely sensitive to [music] timing. If engineers did not correct for that 38-microsecond daily drift, if they simply ignored the fact that time [music] runs differently up there than down here, the position errors would accumulate at roughly 11 km per day. Your navigation would be useless within hours. So every GPS device on Earth, every time you ask your phone to route you somewhere, [music] is silently applying a correction derived from the fact that time is not uniform. The engineers did not debate the philosophy of time before building the system. They did the calculation, applied the correction, and it worked. Which means the weirdness we are talking about is not a theoretical [music] curiosity. It is already baked into the infrastructure of modern life. The question is not whether time behaves this way. We know it does. The deeper question, the one that keeps showing up in every experiment, is whether what we are measuring when we measure time is actually time or something else that we have been calling time because we did not have a better word [music] for it. Something that should have been impossible happened in 2021, and the fact that most people never heard about it says something about how poorly we communicate the genuinely strange results in physics. A team [music] at JILA in Colorado measured the difference in the flow of time between two layers [music] of atoms, separated by 1 mm. Not 1 km, not 1 m, 1 mm. The thickness of a credit card. The difference was real, measurable. The atoms at the top of the sample experienced time passing at a fractionally faster rate [music] than the atoms at the bottom, exactly as general relativity predicts for objects at different heights in a gravitational field. The precision required to detect this was 7.6 * 10 to the power of -21. Right now, as you read or listen to this, time is passing slightly faster at the top of your head than at your feet. Your head is aging at a different rate than your shoes. The difference over a human lifetime is too small to notice, but it is not zero. It is a real, physical, measured fact. And here is where it gets uncomfortable in a way the GPS example does not quite capture. The GPS case tells us time runs differently at different altitudes. Fine, we [music] accept it. But the JILA experiment pushes into territory that feels genuinely vertiginous. 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? Is there one thing called time, or just a collection of local rates that we bundle together under one name for convenience? And if it is the latter, if time is a useful fiction we construct from our immediate surroundings, then what does it mean to say that the universe has a history? What does it mean to say that something happened before something else? I find myself with no clean answer to that. And the experiments that follow make the question sharper, not easier. Here is something that physics has known for a hundred and thirty [music] years, and has never quite figured out what to do with. Take any equation that describes how the physical world behaves.

[music] Newton’s laws, Maxwell’s equations, quantum mechanics, general relativity. Run them forward in time. Now run them backward. The mathematics works identically in both directions. Not approximately, not with small corrections, identically. Which means that nothing in the fundamental laws of physics requires time to move in the direction you experience [music] it moving. The universe, as far as its own equations are concerned, has no preference. Forward and backward are the same. Ludwig Boltzmann saw this in 1895, and it nearly broke him. He spent years trying to explain why we perceive a direction. Past to future, cause before effect, broken eggs never reassembling. When the laws governing every particle in the universe do not contain that direction at all. What he eventually concluded is that the arrow of time is not a law. It is a statistical bias. The universe started in an extremely ordered, low entropy state. From that starting point, disorder increases, not because it has to, but because there are overwhelmingly more disordered states than ordered ones. And random processes tend toward the more probable. That is the entire explanation for why time has a direction. Not a fundamental law. A statistical tendency. Which raises a question Boltzmann never answered and nobody has answered since. Why did the universe start in such an ordered state in the first place? That initial condition, that improbably tidy beginning, is doing all the work. Remove it, and the arrow of time disappears entirely. And to me, that is one of the most unsettling facts in all of science. Everything you experience [music] as the passage of time, memory, aging, causality, the sense that the past is fixed and the future is open, all of it is downstream of one unexplained coincidence at the beginning of everything. So, if the arrow of time comes from statistics [music] and not from the laws themselves, you might expect that at small scales, at the quantum level where individual particles live, time would already start to look symmetric. And that expectation turned out to be correct in a way that is harder to dismiss than most people [music] realize. Andrea Rocco and Thomas Goff at the University of Surrey [music] were looking at what physicists call open quantum systems. Quantum systems that interact with their environment, which is basically everything real, since nothing in nature is perfectly isolated. They wanted to understand why time appears to flow forward in these systems, even though the underlying equations are symmetric. What they found in February 2025 is that even after applying standard simplifying assumptions, even after accounting for the way energy and information dissipate into the environment, the equations still behave the same way whether time moved forward or backward. The mathematical object at the heart of the description, what they called the memory kernel, turned out to be symmetric in time. [music] The arrow was not emerging from the physics of the system. It was simply not there. Goff described his reaction as surprise. That is a careful word for a physicist to [music] use, because it means the data went somewhere the math was not expected to go. What this tells us is that the direction of time we experience is not being generated by the quantum processes underneath our reality. It is being generated by something sitting above the level of the equations. And the obvious candidate, the thing that keeps showing up in result [music] after result, is the observer. The thing doing the looking. Which raises a question that feels almost too strange to ask seriously. [music] What if the direction of time is not a feature of the universe? What if it is a feature of what it means to be aware of the universe? What if it is us? That question, whether the observer produces the arrow of time, is not [music] metaphysics. It has a physical answer. And to get to it, we need to understand something Rolf Landauer proved at IBM in 1961. He showed that erasing one bit of information, resetting a switch from one to zero, clearing a single memory register, necessarily releases a tiny amount of heat into the environment. Not because of engineering inefficiency, not because our computers are imperfect, but because the laws of thermodynamics require it. is physical. Manipulating it costs energy. The minimum cost scales directly with the temperature of the environment. Hotter surroundings, more expensive the erasure. Thinking is more expensive in a hot environment than a cold one. Memory is not free. And this connects to time in a way that took [music] decades to fully appreciate. If the arrow of time is related to entropy, to disorder increasing, and if every act of observation, every act of recording information, every act of building a memory, necessarily increases entropy, then observation is not a passive window onto reality. Observation is a thermodynamic event. [music] It produces heat. It generates irreversibility. It creates entropy. And entropy is the only thing in physics that distinguishes [music] the past from the future. So, observation does not just register the arrow of time, it participates in producing it. What I find remarkable about this is that it means every time you remember something, every time you store information about the past, you are not just recording time. You are, in the most literal thermodynamic sense, generating it. [music] Which raises a question that deserves more attention than it usually gets. If memory creates the arrow of time, and the arrow of time is what makes the past different from the future, then what came first? Did time produce memory, or did memory produce time? In November 2025, a team at Oxford tested exactly how far that logic goes. The result Vivek Vadhwa and Natalia Ares got at Oxford is one of those things that takes a moment [music] to fully land. And I think most people who encounter it move past it too quickly. They built a microscopic clock, a double quantum dot, two tiny regions that a single electron can jump between. Each jump is a tick. The clock works. It keeps time in the most minimal sense possible. Then they asked a question that sounds almost too basic. How much energy does it cost to run this clock? Compared to how much it costs to read it, [music] to observe it, to detect the ticks, to turn quantum events into a record. The answer was not what anyone expected. Reading the clock required up to a billion times more energy than running it. Not twice as much. Not 10 times. A billion. And when they analyzed where all that energy was going, they found it was producing entropy. The act of measurement was generating irreversibility. It was creating the thermodynamic conditions that make one direction of time distinguishable from the other. When the two quantum dots reached the same temperature, when the system was perfectly balanced, the electron jumped forward and backward with equal probability. The clock had no direction. It was equally past and future. [music] Time in that state had no arrow. The arrow appeared only when the measurement apparatus recorded the jumps. Only when something outside the quantum system built a memory of what had happened. The direction of time in this experiment was not a property of the clock. It was a property of the act of looking at the clock. So, here is the question I cannot get out of my head after sitting with this result. If you removed every observer from the universe, every device, every memory, every record of any event, would time still have a direction? The experiment does not answer that. But it makes it impossible to say yes with any confidence. The mechanism inside that Oxford experiment is where the strangeness really lives. [music] And it is worth slowing down to look at it carefully. The electron sits between two quantum dots. It can jump left or right. At the quantum level, without any measurement happening, these jumps are symmetric. The electron has no preference. [music] Forward and backward are indistinguishable. There is no tick. There is just motion. Now the measurement apparatus turns on. It records each jump. [music] It builds a sequence. Left, right, left, left, right. And that sequence has a direction. It has a before and an after. It has a past [music] and a future. The arrow of time appears the moment the record exists. And here is the part that should give you pause. The arrow is in the record, not in the electron. The electron does not know it is being observed. [music] It does not change its behavior because of the measurement. But the entropy, the irreversibility, the thing that makes past different from future, is generated entirely in the measurement apparatus. In the act of writing the result down. Which means that if you imagine a universe with no records, no memories, no structures that retain information about previous states, that universe would have electrons jumping back and forth, energy moving around, things happening, but no arrow of time, no past, no future, [music] just an eternal symmetric present. That is not science fiction. That is what the Oxford experiment implies. And it connects directly to something John Wheeler argued for decades. That information is not just a description of physical reality. That at its deepest level, the universe is not matter and energy, but bits, records, questions answered. If that is right, then time itself is a record. A memory the universe keeps of itself. And the question of what is doing the remembering does not have a clean answer yet. If the arrow of time lives in the record, rather than in the physics, [music] the next question arises almost on its own. What does quantum mechanics actually say time to begin with? And the answer is stranger than most people realize. Because quantum mechanics barely says anything about time at all. In every other area of physics, time and space are treated as partners. General relativity weaves them into space-time, a single fabric that bends and curves. But in quantum mechanics, that partnership does not exist. Space is an observable. [music] You can measure position. You can put position into the equations as something the system itself possesses. [music] Time is not an observable. It is a parameter, a number you feed into the equation from outside, like setting a dial before you run the experiment. The Schrödinger equation treats time as something that simply flows in the background, uniform and external, untouched by anything the quantum [music] system does, which is almost exactly what Newton said in 1687. Absolute time, flowing equably, independent of external. Quantum mechanics, for all its revolutionary strangeness, smuggled Newton’s time back in through the back door. John von Neumann formalized this in

Time in quantum mechanics is not an operator. It is a given. Which means that every quantum experiment ever run, [music] every result we have, every prediction we have tested, was conducted inside a framework where time was assumed rather than derived. We have never measured time from inside a quantum system. We have only ever measured quantum systems from inside time. Whether those are the same thing is a question that turns out to matter enormously. And in 2020, a team in Vienna decided to stop treating it as a theoretical problem and actually push on it in a laboratory. What Julia Rubino’s team in Vienna was testing sounds [music] like it belongs in a philosophy seminar rather than a physics lab. They wanted to know whether the sequence of events in a quantum system has to be fixed. [music] Whether A always comes before B if A causes B. In classical physics, causality is absolute. The order of events is written into the structure of reality. What they found is that this is not how quantum systems work. They set up a situation where a quantum particle passes through two operations. Call them A and B. And the order in which it experiences them [music] is in quantum superposition. Not randomly one or the other. Both simultaneously. A before B and B before A at the same time in the same experiment [music] for the same particle. And this was not a theoretical result. It was measured. The experiment [music] produced outcomes that could only be explained if causal order was genuinely indefinite. Not unknown, not random, but superposed. So, here is what that actually means. Causality, the principle that causes precede effects, that the past shapes the future and not the other way around, is not a fundamental feature of quantum reality. It is something that emerges at larger scales. The way temperature emerges from the motion of molecules underneath it. At the quantum level, before and after are not fixed. The universe does not have a built-in before. It has a before only when something forces the order to become definite. Which, if you have been following the thread from Boltzmann through Landauer through the Oxford experiment, sounds familiar. The order of time, like the direction of time, seems to need something to fix it. And that something keeps pointing back toward the same place. Toward the observer. Toward the record. Toward whatever is doing the measuring. Whether that is a satisfying answer or the beginning of a much deeper problem is something I genuinely do not know. So, we have two theories. General relativity and quantum mechanics. Between them, they account for essentially everything we can observe in the physical universe. Every particle, every force, every structure from atoms to galaxy [music] clusters. Both have been tested to extraordinary precision. Both work. And yet, when you sit them down next to each other and look at what they say about time, they contradict each other so completely that they cannot both be right about what time fundamentally is. In general relativity, time is dynamic.

[music] It is a dimension of space-time, a real physical thing that curves in response to mass and energy. That stretches near massive objects. That is woven into the geometry of the universe. It is not a background. It is a participant. In quantum mechanics, time is a fixed external parameter. It does not curve. It does not respond to the quantum system. It just flows in the background, providing the stage on which quantum events occur. It cannot be measured from inside the system. It cannot be put into superposition. It is, in the quantum framework, essentially Newtonian. These descriptions are not just different. They are structurally incompatible. General relativity needs time to be a dynamic physical thing. Quantum mechanics needs time to be an inert external given. You cannot have both simultaneously. Which means one of the two most successful theories in the history of science has an incomplete picture of something as basic as time. And here is what bothers me about how this is usually discussed. We talk about it as a technical problem waiting for a technical solution. But what if the incompatibility is not a gap to be filled? What if it is telling us something about the nature of time that the vocabulary to say directly? In 1967, one calculation tried to follow that logic to its end. The answer was not what anyone was hoping for. John Wheeler and Bryce DeWitt were not trying to do something exotic. They were trying to do the obvious thing. Write down the quantum mechanics of the universe as a whole. Not the quantum mechanics of a particle or an atom. The universe. All of it. One equation. When you attempt this, you immediately run into a technical problem. General relativity describes how the geometry of space-time evolves over time. Quantum mechanics describes how quantum states evolve over time. To combine them, you need a single equation that handles both. When Wheeler and DeWitt wrote that equation in 1967, time disappeared. Not approximately. Not in some limiting case. The time variable, the 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 [music] physicists could construct, does not evolve in time. It just is. Static. Timeless. [music] A single quantum state with no before and no after. This is called the Wheeler-DeWitt equation, and it is not a fringe result. It is the starting point for every serious attempt at quantum gravity. And what it says, with complete mathematical clarity, is that time is not a fundamental feature of the universe. It is something that has to emerge somehow from a timeless underlying reality. The question is, how? How do you get time from a universe that the equations say has none? There are several candidate answers, and none of them are fully satisfying. But in 2013, one team decided to stop debating the candidates and test one of them in a laboratory. What they found is one of the most quietly unsettling [music] results in modern physics. And it almost never comes up outside specialist discussions. The experiment Ekaterina Moreva and her colleagues [music] built was designed to test an idea Page and Wooters had proposed back in 1983. The idea is called the relational interpretation of time, and it works like this. If the universe as a whole has no time, as the Wheeler-DeWitt equation says, then time must arise from the relationship between parts of the universe. Specifically, between a clock and an observer of that clock. To test this, Moreva’s team created a simple quantum system. Two entangled photons. One acted as the clock. The other acted as the observer. Then, [music] they looked at the system in two different ways. First, from the outside, treating the whole two-photon system as a single object [music] and measuring it from without. From this perspective, the system appeared static. Timeless. No evolution, no change, no before or after. [music] Exactly what the Wheeler-DeWitt equation predicts. Then, they looked from the inside, using one photon to measure the other. Letting the internal observer track the internal clock. From this perspective, time existed. The clock ticked. Events had a sequence. The internal observer experienced time passing. Same system. Same physical setup. Two completely different realities depending on where you stood. If you were outside, the universe had no time. If you were inside, it did. What this means is almost too large to hold in one thought. Time may not be a property of the universe. It may be a property of being inside the universe. It may be what [music] it feels like to be part of something rather than an observer of the whole. And if that is true, then every question we thought we knew how to ask about the beginning of time, the end of time, the direction of time, all of them need to be reframed entirely. Not because the questions are wrong, because the thing they’re asking about is not where we thought it [music] was. What the Maricopa experiment actually implies is something most summaries of it miss. If you are inside a system, you experience time. If you are outside, you do not. And the universe, by definition, has no outside. There is no position you can occupy that is not inside it. Which means the timeless description, the Wheeler-DeWitt equation, the static quantum state, the universe with no before and no after, is a description from a perspective that physically cannot exist. No observer can ever stand there. The only description any real observer can ever access is the internal one. The one where time exists. And here is where it gets more specific and more unsettling than just saying time is an illusion. This does not mean time is not real. It means something more precise. It means time is real, but only locally. Only from inside. Only for participants. The universe does not have a single time everyone shares. It has a collection of local times, [music] each one arising from the relationship between a particular observer and a particular clock. What we call time, the thing we measure, feel, structure our entire existence around, [music] is something that emerges from entanglement. From quantum correlations between parts of a system that has no global time at all. What I find genuinely vertiginous about this [music] is the question it opens at the boundary. If time emerges from the relationship between observer and system, if it is not a backdrop, but a product, what happens at the edges of that relationship? What happened to time before observers existed? Was there a then at all? And is that question even coherent, given that the word before already assumes the thing we are trying to explain? While theorists work through what the Maricopa result means, other physicists are doing something more immediate. They are pushing on general relativity with a precision it has never faced before, looking for the exact point where it starts to show a crack. The ACES mission, Atomic Clock Ensemble in Space, is running right now on the International Space Station. It carries two clocks. PHARAO, a cold atom cesium clock developed by CNES in France, [music] and SHM, a hydrogen maser built by Swiss institutions. Together, they generate a time signal with a fractional frequency stability of 1 * 10 to the power of -16. At that level of precision, even the smallest deviations from general relativity’s predictions become visible. The clocks on the ISS are compared continuously with networks of atomic clocks on the ground through a microwave link that corrects for all known sources of noise and drift. What the mission is looking for is any departure from what Einstein’s equations predict for how time runs at that altitude and velocity. Not to confirm relativity. Relativity has been confirmed enough times. What they are looking for is where it fails, because everyone in the field knows it must fail somewhere. General relativity and quantum mechanics cannot [music] both be completely correct. And at some level of precision, one of them has to show a crack. ACES is sensitive enough that if the crack exists at the scale [music] physicists expect, it should be visible in the data. The results are still coming in. And what I find striking about that, genuinely striking, is that we are living at the exact moment when the most precise test of time ever conducted is running, and the answer is not in yet. What do you do with that? Most people have no idea this experiment exists. [music] The crack, if it is there, is being measured right now. Nobody knows yet which side of it we are on. What the ACES mission cannot do is test general relativity and quantum mechanics simultaneously in [music] a single measurement. Those two theories live at different scales, and every experiment before now has tested one or the other, [music] but never both at once. That changed in July 2025. Igor Pikovski at Stevens Institute, Jacob Covey at the University of Illinois, and Johannes Borregaard at Harvard published a protocol in PRX Quantum that does something I find almost audacious in its ambition. Their design puts an atomic clock into a superposition of two different heights above Earth’s surface. The same clock at the same moment is simultaneously higher and lower. In quantum mechanics, both are real. In general relativity, the two heights correspond to two different rates of time flow. So, the clock in superposition is simultaneously experiencing two different rates of time. And the interference between those two time flows, the quantum signature of the superposition, can be measured. This has never been done before, because it requires both quantum coherence and gravitational precision at a level that only recently became technically feasible. What makes this matter is not just the precision. It is what the result is forced to include. Every experiment before this tested quantum mechanics in flat space-time or general relativity without quantum effects. This is the first protocol that makes both theories show up in the same number at the same moment. And if that number does not match what either theory [music] predicts individually, if the combined quantum gravitational behavior of time is something neither framework anticipated, then we will have direct experimental evidence that both are approximations to something we do not yet have, a theory without a name, a theory nobody has written down yet. What would it even look like? That is not a rhetorical question. It is the actual open problem. There is a detail about quantum transitions that does not fit comfortably [music] into any of the frameworks we have been building. The event has a duration. Not a duration measured by an external clock. A duration written into the geometry of the material itself. Hugo Dil’s team at EPFL showed this in February 2026. [music] They measured how long it takes an electron to absorb a photon and jump to a higher energy state without using any external reference. [music] Every previous measurement of this kind relied on a clock sitting outside the quantum system, timing it from without. What Dil’s team did instead was use the quantum system’s own internal structure to record the duration. When an electron absorbs a photon, it can follow multiple quantum pathways simultaneously. These pathways interfere with each other, and the interference leaves a signature in the spin of the emitted electron. By analyzing that signature, the team could determine how long the transition took, entirely from information encoded inside the process [music] itself. What they found is that quantum transitions are not instantaneous. They have a measurable duration. But the duration depends on the atomic structure of the material. Change the symmetry of the crystal, and the transition takes longer. Lower symmetry, longer time. Higher symmetry, shorter time. The geometry of the material determines how long the quantum event lasts. Not an external clock. Not the flow of time in the background. The shape of the thing itself. And this sits in the same family as the Wheeler-DeWitt result [music] and the Maricopa experiment. If the duration of a quantum event is determined by internal geometry rather than external time, what exactly is flowing when we say time flows? [music] Is it possible that what we call time is just a way of describing [music] the internal geometry of processes? That duration is not something events happen inside, but something that is a property of the events themselves. I do not know how to make that idea comfortable, but I also do not know how to dismiss it. And back to that NIST clock from the beginning. The same month, Pikovski’s team published their quantum network protocol, the NIST team in Boulder announced something that connects all [music] of these threads to a question most physicists consider almost too large to ask directly. Their aluminum ion clock, accurate to the 19th decimal place, 41% more precise than anything before it, is now sensitive enough to detect changes in the fundamental constants of nature. The fine structure constant, [music] which determines how strongly electrons interact with light. The gravitational constant, [music] which sets the strength of gravity. These numbers appear throughout physics as fixed values [music] written into the equations as given. But they are not derived from anything deeper. They are measured. And if they are measured quantities rather than mathematical necessities, there is no fundamental reason they have to stay constant. They could drift slowly, almost beyond detection, but measurably. If you have a precise enough instrument, the NIST clock is now precise enough to check. Not assume, actually check. And if the constants are drifting, if the fine structure constant today is not exactly what it was a billion years ago, then the laws of physics [music] as we know them are not eternal truths. They are snapshots, descriptions of the universe as it happens to be right now, at this moment in cosmic [music] history. And if the laws are changing, then time, the dimension along which they change, [music] is more fundamental than the laws themselves. Time stops being a consequence of physics and becomes the thing physics happens inside, which inverts everything we thought we knew about the relationship between time and physical law. And that inversion connects directly to a crack that has been widening at the center of cosmology for about a decade. A crack that does not involve quantum mechanics at all, but is just as hard to explain away. There is a number that has been sitting at the center of cosmology for about a decade, refusing to resolve. And it matters for exactly the reasons the NIST result matters. It is the Hubble constant, the rate at which the universe is expanding. It should be one number. The universe is one universe. But when physicists measure it two independent ways, they get two different answers. The first method uses the cosmic microwave background, the faint afterglow of the early universe, mapped with extraordinary precision by the Planck satellite in 2018. This gives 67.4 km per second [music] per megaparsec. The second method uses Cepheid variable stars as distance markers, led by Adam Riess and his team, and gives 73.2 km per second per megaparsec. The difference is five sigma. In physics, five sigma is the threshold for calling something a discovery. This gap is large enough that it almost certainly cannot be explained by measurement error. Both teams have checked their work repeatedly. The numbers do not converge. Now, here is why this belongs in a conversation about time, rather than just about space. The Hubble constant is not just a rate of expansion. It is a rate of change in the geometry [music] of space-time itself. It tells you how fast the fabric, including the time dimension, is evolving at cosmic scales. If two independent methods of measuring that rate disagree by five sigma, it means we do not fully understand how the geometry of space-time is behaving on the largest scales. And if we do not understand that, our entire picture of cosmic time, the 13.8 billion-year timeline, the sequence of cosmic events, the age of structures, is built on something that is not fully resolved. What I find hard to accept about this is how casually it is sometimes described as a measurement problem, rather than a physics problem. Five sigma is not a rounding error. It is a signal. And in 2024, a second instrument made that signal considerably louder. The Hubble tension was already unsettling enough on its own. Then the Dark Energy Spectroscopic [music] Instrument added something worse. DESI spent years mapping the distribution of galaxies across enormous volumes of the universe, using the patterns of that distribution to measure how cosmic expansion has changed over time. What it found in 2024 is that dark energy, the force driving the acceleration of expansion, the thing Einstein called the cosmological constant, may not be constant at all. The data suggests it is weakening, slowly, but measurably. It was stronger in the early universe and has been decreasing since. Dark energy is the engine driving the expansion of space. And the expansion of space is the mechanism by which the universe has a cosmic history, a sequence of events, an earlier and a later, a past and a future at the largest scales. If the engine is not constant, then the rate at which cosmic time flows is not constant, either. The universe is not just expanding. It is expanding at a rate that is itself changing in a way we did not predict and do not fully understand. And the uncomfortable implication of the DESI result, sitting alongside the Hubble tension, is that our model of cosmic time is built on assumptions about dark energy that may be wrong. If dark energy is evolving, the timeline shifts. Events we thought we had dated may have occurred at different moments. The history of the universe, told in time, depends [music] on understanding what time is doing at cosmic scales. And right now, two of the most important measurements in cosmology are telling us something we do not yet know how to read. I think this gets underreported because it is too large to sensationalize [music] cleanly. But the ground is moving. And if the engine of cosmic time is changing, that raises a question nobody in the field has a confident answer to. Changing toward what? There is a risk that all of this starts to feel like trouble at the edges, exotic problems that do not touch the solid ground of what we already know. But the solid ground has its own strangeness, and it has been measured clearly enough that there is no room to dismiss it. In 1959, [music] Robert Pound and Glen Rebka did an experiment at Harvard that is about as clean and direct as physics gets. They measured whether light changes frequency as it moves through a gravitational field, whether light loses energy climbing up and gains energy falling down. They did this by shining gamma rays up a tower 22.5 m tall and measuring whether the frequency of the light at the top differed from what was emitted at the bottom. It did. The shift matched exactly what general relativity predicted. A prediction Einstein had made in 1907, 52 years before anyone measured it, using nothing but the principle that acceleration and gravity are equivalent. What strikes me about this is not the confirmation. It is the gap. 52 years between prediction and measurement, and the measurement arrived exactly where the theory said it would. Time really does run slower at the bottom of a tower than at the top. Not by much, a few parts in 10 to the power of 15, but unambiguously, repeatedly, in a building in Cambridge, Massachusetts. Everything since then [music] has been a deeper version of the same question. Why does the geometry of space-time make the rate of time depend on position? And the deeper you push that question, through the Giga millimeter result, through the Oxford quantum clock, through the Wheeler-DeWitt equation, the stranger the territory becomes. The strangeness [music] is not at the edge. It was in the building in 1959. We just did not fully see it yet. Something physicists had been predicting for decades, but had never managed to produce cleanly in a laboratory, finally happened in December 2025. And the result is one of those things that sounds like science fiction until you look at what it actually demonstrates. Hadiseh and his team at the City University of New York built a system that reflects electromagnetic waves in time, rather than in space. [music] Normally, when a wave hits a boundary, a wall, a surface, a transition between two materials, it reflects back through space. Moussa’s experiment created a temporal boundary instead. By suddenly and uniformly changing the physical properties of the medium the wave was traveling through, shifting the electrical permittivity of the system instantaneously, [music] they created a moment in time where the wave experienced a discontinuity. And the wave reflected. Not backward through space, backward through time. The signal began retracing its own history. It went back through the same medium it had already passed through, reproducing the past states of the wave in reverse order. The team also observed frequency translation. The reflected wave emerged at a different frequency than the original, which is the temporal equivalent of a wave changing direction when it enters a different medium. The result was clean, repeatable, and confirmed a theoretical prediction that had existed since the 1960s without ever being clearly demonstrated. The experimenters are careful to note that time outside the system continued normally. What changed was the behavior of the wave inside an engineered temporal boundary. But that note of caution, while scientifically accurate, does not fully dissolve the question the experiment raises. [music] If the boundary between past and future is something that can be engineered, something that responds to the physical properties of a medium, what exactly is that boundary? What is it separating? The mechanism of Moussa’s experiment reveals something about time that was always true [music] and had simply never been demonstrated this clearly before. When a wave hits a spatial boundary, a wall, what is conserved is its frequency. The pitch stays the same, the direction changes. The temporal boundary works the opposite way. When the medium changes suddenly in time rather than in space, what is conserved is the direction. The wave keeps moving forward through space, but its frequency changes, shifting to a value determined by the properties of the new medium. And simultaneously, a time-reversed copy of the wave is generated, [music] propagating backward. This is not a property of exotic quantum systems. It follows from Maxwell’s equations, which are among the most tested relationships in all of physics. Which means the capacity for time reversal is not something that has to be injected into physics from outside. [music] It is already there, in the structure of the equations, waiting for the right physical conditions. Moussa’s team did not discover something new. They revealed something that was always present. And that is a different kind of result. It says that the asymmetry of time, the fact that waves travel forward and not backward, that signals propagate into the future and not the past, is not written into the laws of electromagnetism at all. It is written into the initial conditions. Into the fact that we start with a wave going one direction and a medium that is uniform. Change the medium suddenly and the equations immediately produce both directions simultaneously. [music] The arrow of time, even in classical physics, is not a law. It is a choice of starting conditions. And this brings back the question from Boltzmann, now sharper. Who chose the starting conditions? Or more precisely, was it a choice at all? Or was it the only option available? [music] That is not a rhetorical question. It is one of the open problems in physics. That question, why the initial conditions were what they were, connects to results from particle physics that I find deeply puzzling, and that do not get enough attention in discussions about time. What the BaBar experiment at SLAC measured in 2012 was direct evidence of time reversal violation in subatomic particles. [music] They were studying B mesons, unstable particles that can transform into their antiparticle equivalents. The question was whether the process of going from state A to state B was statistically identical [music] to the process of going from state B back to state A. If time is truly symmetric, those rates should be equal. They were not. The transition rates differed by a measurable amount. Then, in [music] 2020, the T2K experiment in Japan found a related asymmetry in neutrinos, matter and antimatter versions of the same particle oscillating between different types [music] at different rates. Both results are real. Both are statistically robust. And both show that the arrow of time [music] is not just a statistical tendency at large scales, the way Boltzmann described it. It is built into the behavior of individual particles in ways that the standard model can accommodate mathematically, but cannot explain from first principles. We can describe the asymmetry. [music] We cannot say why it exists. What bothers me about this is the gap between the explanations we have and the explanation we actually need. Boltzmann gave us thermodynamics, the arrow of time as a statistical bias toward more probable states. [music] But BaBar and T2K are showing something at the particle level that is not [music] statistical at all. It is structural, built in at the level of individual interactions. And a structural asymmetry demands a structural explanation. We do not have one. Which means the deepest source of time’s direction remains unknown, even as our measurements of it get more precise. The BaBar and T2K results [music] show that the direction of time is built into individual particle behavior. But there is a result from quantum mechanics that adds something almost opposite [music] to this picture. Not that time has a forced direction, but that observation can stop it entirely. It is called the quantum Zeno effect, [music] first proposed by George Sudarshan and B N. Misra in 1977, [music] and confirmed experimentally multiple times since. The idea is this. A quantum system evolves over time, a particle decays, an atom transitions between energy levels, a quantum state changes. [music] This evolution has a characteristic time scale. But if you measure the system frequently enough, if you observe it fast enough, the evolution slows down. The more often you look, the slower it changes. And in the limit of continuous observation, the system stops evolving entirely. It freezes. You can hold a quantum system in its current state indefinitely simply by measuring it fast enough. Now, put this alongside the Oxford experiment. Observation creates the arrow of time. The Zeno effect says observation also controls the rate of time, how fast quantum evolution proceeds. Frequent observation slows quantum time. No observation removes the arrow entirely. And the act of observing costs energy, as Landauer showed. Which means that controlling time at the quantum level has a thermodynamic price. You pay to slow it. You pay to give it direction. Time, in this picture, [music] is not a river flowing independently of everything around it. It is something closer to a process that requires participation to sustain [music] and participation to direct. Which raises a question I find genuinely difficult to resolve. If observation sustains time, if the rate and direction of time depend on the presence and frequency of measurement, what was time doing before there was anything in the universe capable of observing anything? Was it doing anything? And if the answer is no, then what exactly started The fact that observation controls time at the quantum level makes the question of measurement precision suddenly feel less like engineering and more like philosophy made [music] physical. Because the more precisely you can measure time, the more deeply you can probe these [music] effects. And precision took a step in 2024 and 2025 that changes what [music] is possible. Multiple teams at PTB in Braunschweig and [music] at JILA in Boulder demonstrated working nuclear clocks based on thorium 229. [music] The difference between these and conventional atomic clocks is not [music] just precision. It is what is doing the oscillating. Atomic clocks use the transitions of electrons in the outer shells of atoms, which interact with the surrounding electromagnetic environment and can be disturbed by electric and magnetic fields. Nuclear clocks use transitions inside the atomic nucleus itself, shielded from almost everything outside by the electron cloud surrounding the nucleus. The result is a stability orders of magnitude better than atomic clocks, and a sensitivity to effects that atomic clocks cannot reach. Specifically, to changes in the strong nuclear force, the force holding protons and neutrons together. If the strong force has a coupling constant that varies even slightly over cosmological time, a nuclear clock will eventually detect it. And here is why that matters beyond the obvious. [music] If fundamental constants are changing, then the laws of physics are historical rather than eternal. They have a before and an after, [music] which means they are inside time. And if the laws are inside time, time cannot be derived from the laws. Something has gone in a circle. In physics, a circle usually means a missing piece. This circle has been sitting here for decades. [music] The nuclear clock is the first instrument precise enough to start looking for it. What it finds, or does not find, will either close the circle or force us to redraw it entirely. That circle, time cannot be derived from laws that are themselves inside time, runs directly into something happening right now at the level of international standards. And I think the way this is usually reported buries what is actually interesting about it. The international community is preparing to redefine the second. A coordinated comparison of optical clocks across six countries, published in June 2025, was a major step toward making this formal sometime around 2030. The current definition, based on cesium oscillation since 1967, [music] is being replaced with something based on optical transitions. Oscillating at frequencies roughly 10,000 times higher than cesium, measurable with far greater resolution. Now, here is what gets left out of most reporting on this. The redefinition is not just a technical upgrade, it is an admission. An admission that the previous ruler was not precise enough to measure the thing we are actually trying to understand. The new definition will allow physicists to test whether fundamental constants drift, to probe quantum gravitational behavior of time, to run experiments that the cesium standard simply could not support. But here is what I keep coming back to. Every definition of the second, the old one and the new one, is a definition of how time behaves, not what time is. The standards committee is not in the business of answering that. Nobody is. And the remarkable thing about rebuilding the ruler right now is that the new ruler will be precise enough to show us things we have never seen before. Whether what it shows us will tell us what time actually is or just give us more precise descriptions of something we still do not understand. [music] That is the question the redefinition process deliberately avoids. But it may be the only question that matters. And I wonder sometimes whether the answer, when it comes, will look anything like what any of us are expecting. If you try to follow all of these results to a single conclusion, the NIST clock, the Hubble tension, DESI, the Wheeler-DeWitt equation, the Mariva experiment, [music] the Oxford quantum clock, what you run into is the same wall from every direction. Time is not what we thought it was, and the frameworks we have do not agree on what it actually is. Carlo Rovelli’s answer to this, developed with Lee Smolin in loop quantum gravity since 1988, is to say that time is not a background. It is not a stage on which events occur. It is a description of how events are related. Two events are related if one influences the other. If a signal can pass between them, if they share a causal connection. Time in loop quantum gravity is nothing more than the ordering of those relations. It is not a thing. It is a description. Rovelli makes a point that is easy to understand and [music] hard to accept. When you remove the thermodynamic approximation, when you look at a small enough system with few enough particles that statistical effects vanish, time disappears from the equations. The equations have no preferred direction, no flow, no before and after. Time, he argues, is what happens when you zoom out. When you have a large enough system with enough interacting parts that statistical behavior emerges and entropy starts to increase, time is a macroscopic phenomenon, like temperature. Temperature does not exist for a single molecule. It exists for large collections of molecules as a measure of their average kinetic energy. Time, in this view, exists for large collections of events as a measure of their average causal ordering. Which means that asking what time is doing at the smallest scales, at the Planck length, at the level of individual quantum gravity events, is like asking what the temperature of a single proton is. The question may simply not apply. And if time does not apply at the smallest scales, the next question writes itself. What exactly does it emerge from? If loop quantum gravity dissolves time into relations between events, Roger Penrose takes the same problem in a completely different direction. His framework, conformal cyclic cosmology, does not try to explain where time comes from. It tries to explain why the universe started in such an improbably ordered state [music] in the first place. That is the real mystery underneath everything Boltzmann identified. The arrow of time points from low entropy to high entropy. Fine. But why was entropy so low at the beginning? Why did the universe start in extraordinary order when the overwhelming majority of possible initial states are disordered? [music] The standard answer, that it just did and we should not ask why, has always felt like a placeholder. [music] Penrose refuses to accept it. His proposal is that the Big Bang was not a beginning. [music] It was a transition. In conformal cyclic cosmology, the universe passes through an infinite sequence of eons. Each eon starts with something like a Big Bang and ends with something like a heat death, maximum entropy, cold, empty, dark. But here is where it gets strange. In the far future of an eon, when all matter has decayed and all black holes have evaporated, only massless particles [music] remain. Photons and gravitons moving at the speed of light. [music] And massless particles do not experience time. They have no internal clock. The distinction between a nanosecond and a trillion years does not exist for them. At that point, Penrose argues, the universe loses track of scale. It cannot distinguish between very large and very small. In that scale-less [music] state, the cold empty end of one eon and the hot dense beginning of the next become geometrically equivalent. The heat death and the Big Bang are the same moment seen from different sides. The end of time and the beginning of time are the same event. Which means the low entropy beginning of our universe is not an accident requiring explanation. It is a mathematical consequence of the high entropy end [music] of the previous one. I find this answer either deeply satisfying or deeply unsatisfying depending on the day. What I cannot do is dismiss it. The geometry works. Stephen Hawking arrived at a similar dissolution from a completely different mathematical direction. Working with James Hartle in 1983, he proposed what they called the no-boundary condition. A way of describing the quantum state of the universe that does not require an initial condition at all. The approach uses a mathematical technique called a Wick rotation, which replaces real time with imaginary time, multiplying the time coordinate by the square root of negative [music] one. This sounds like a formal trick. In the context of quantum gravity, Hawking and Hartle argued it was more than that. In imaginary time, the distinction between time and space disappears. The four dimensions of space-time become four spatial dimensions, [music] all equivalent, none playing the special role time plays in ordinary physics. And in this four-dimensional spatial geometry, there is no boundary, no edge, no point you can call the beginning. Just as the surface of the earth has no edge, you can travel in any direction indefinitely without reaching a wall, the universe in imaginary time has no beginning. The Big Bang, when you rotate back from imaginary time to real time, appears not as a moment where time start from zero, but as the point where a spatial geometry transitions into a space-time geometry. A smooth rounded cap, like the bottom of a sphere. Time did not begin there. It emerged [music] there from a region where the distinction between time and space did not yet apply. What I find striking about both Penrose and Hawking here is that they are pointing at the same [music] thing from opposite directions. Penrose says the beginning was the end of something else. Hawking says the beginning was a smooth geometric transition from a timeless [music] state. Both dissolve the question of what came before, but neither answers the question that emerges from dissolving it. If time emerged, whether from a previous eon or from imaginary geometry, what determined which way it would point when it appeared? There is one piece of this picture that connects the larger scales to the smallest in a way none of the theoretical frameworks have managed to reconcile. It concerns black holes and what happens to information inside them. And it matters here because of what it implies about whether the past is a definite [music] thing. Hawking showed in the 1970s that black holes emit radiation, now called Hawking radiation, as a consequence of quantum effects near the event horizon. [music] This radiation is thermal. It carries energy but no information about what fell into the black hole. It is random. And as the black hole slowly evaporates, the information about everything that ever fell in, every particle, [music] every quantum state, every bit of physical data, appears to be gone, not hidden, not encoded in correlations too subtle to detect. Gone. This violates unitarity, the requirement in quantum mechanics that quantum evolution is reversible and that information is always conserved. [music] If information can be destroyed, the past cannot be reconstructed from the present, even in principle. The chain of cause and effect that makes the past a well-defined thing becomes undefined for anything that has [music] ever crossed an event horizon. The reason this belongs in a conversation about time is that [music] temporal order and an information are the same thing seen from different angles. To say that the past is definite, that events happened in a specific sequence, is to say that information about those events is preserved somewhere in the present. If black holes destroy information, they destroy the definiteness of the past. They introduce a genuine ambiguity into the timeline of the universe that no experiment from outside can resolve. And there are a lot of black holes, which means the past may be less definite than we assume. And if the past is not fully definite, if the timeline of the universe has genuine gaps in it, then what exactly is the thing we are trying to measure when we point a clock at reality? Here is something that should probably bother you more than it does. The twin paradox, the thought experiment where one twin travels at high speed and returns younger, is not [music] a thought experiment. It has been measured. Astronauts who spend 6 months on the International Space Station return to Earth having aged approximately 0.007 seconds less than people who stayed on the ground. Scott Kelly, who spent nearly a year on the ISS, is measurably younger than his identical twin Mark, who remained on Earth. The difference is 7 milliseconds over 340 [music] days. Real. Measured. Which means two people who were born at the same moment, who started with identical clocks, live different amounts of time during the same historical period. They experienced different [music] quantities of duration. And the part of this that gets overlooked in the standard reporting is not the 7 milliseconds. It is what 7 milliseconds implies about the concept of a shared present. When Scott Kelly returned from the ISS, he and his brother were not at the same moment in time in any absolute sense. [music] They were at the same location in space, but the amount of time each had passed through was different. Their nows the same now. And if that is true for two people separated by 400 km of altitude and orbital velocity, it is true in principle for any two objects anywhere in the universe. Every object carries its own local time. Those local times do not sum to a single shared timeline that the universe is moving through [music] together. There is no universal now. And once you fully accept that, not as an abstract physics result, but as a physical fact about the structure of reality, the question that follows is one I do not think has a comfortable answer. If there is no shared present, what exactly is the universe doing right now? The relativity of simultaneity goes further than the twin result, and it is worth taking seriously, [music] rather than filing away as too abstract a matter. Einstein showed in 1905 that two events simultaneous for one observer are not simultaneous for an observer moving relative to the first. This is not an illusion caused by signal delays or imperfect clocks. It is a genuine feature of the geometry of space-time. [music] The events really do occur at the same time in one reference frame and at different times in another. Neither frame is more correct. Both are equally valid descriptions of reality. What this means is that the concept of now, a single present moment extending across space, is not a feature of the universe. It is a feature of a particular reference frame. Change your velocity, and your now changes. Events in your future become simultaneous with your present or part of your past. For someone moving differently relative to you. The physicist Reed Dyck and the philosopher Putnam argued in the 1960s that this implies a block universe, a four-dimensional structure in which all moments of time exist equally. Past, present, and future all equally real. With the experience of flow being a perspective effect, rather than a fundamental feature of reality. David Albert has pushed back on this, arguing that the block universe interpretation smuggles in assumptions about what real means that the physics [music] does not force. Both positions are still actively debated. What is not debated is the underlying physical fact. The present moment, as you experience it, is a local phenomenon. Real for you. [music] Real for everything in your immediate vicinity. But not a global fact about the state of the universe. [music] And if the present is local, then the flow of time, the movement from one present to the next, is also local. [music] Personal. Not the universe’s. Which means when we ask what time is, [music] we may be asking what we are, not what the universe is. What Rovelli concludes from all of this is not that time is an illusion. It is something more precise and in some ways more troubling. He says time is not a river. A river has banks, a current, a direction that exists independently [music] of the water. Time is more like heat. Heat does not exist in a single molecule. It is not a property of any individual particle. It emerges from the collective behavior of enormous numbers of particles [music] interacting. And it flows from hot to cold, not because any law demands it, but because the statistics of large numbers overwhelmingly favor it. Remove the large numbers, go down to individual interactions, and heat loses its meaning entirely. Time, Rovelli argues, [music] works the same way. At the level of fundamental physics, in the equations describing individual quantum events and their relations, there is no time, no flow, no direction, no before and after. These emerge only when you zoom out, when you have enough interacting parts, enough entanglement, enough thermodynamic complexity that statistical behavior kicks in and entropy starts to increase. Time is what the universe looks like from the inside, when it is big enough and complex enough to have a thermodynamic arrow. So, is time real? I think that is the wrong question, or at least an underspecified one. Temperature is real. It is measurable. It has genuine physical consequences. But it does not exist at the level of individual particles. Asking whether temperature is real is asking whether collective emergent phenomena count as real. [music] And the honest answer is that physics does not have a clean way to draw that line. What it has [music] is a growing collection of results from 1895 through 2026 [music] that keep pointing in the same direction. Time is not what we thought it was. And the more precisely we measure it, the more the question opens up rather than closing. Lee Smolin looks at the same data as Rovelli and arrives at almost the opposite conclusion. And the fact that two serious physicists [music] can sit with identical experimental results and reach incompatible positions is, I think, the most honest summary of where the field actually stands. Smolin’s argument, developed [music] in his 2013 book and subsequent papers, starts with a question that Rovelli’s framework leaves unresolved. If the laws of physics are what they are because of what they are, if the constants and equations are fixed, [music] given from outside, not derived from anything deeper, then the universe requires an explanation that physics cannot provide. [music] Where do the laws come from? Why these values and not others? The standard answer is that we measure them rather than derive them. Smolin finds this unsatisfying. His proposal is that the laws evolved was that the universe has a history, not just of events, but of physical law itself. That the parameters of physics have changed over cosmic time. Selected by something like natural selection acting on universes. Each black hole that forms might spawn a new universe with slightly [music] different constants. Universes that produce more black holes reproduce more. [music] The laws we live under are the laws that survived. If the laws evolve, then time is prior to the laws. You cannot derive time from the laws if the laws are themselves products of time. Time has to be the one thing that does not emerge from anything else, because everything else, including the rules, emerges from it. This puts Smolin in direct opposition to Rovelli. One says time is emergent. [music] The other says time is fundamental. Both are using the same experiments as their starting point. I do not know who is right. I’m not sure anyone does. And I find that fact, the genuine unresolved nature of this disagreement between people who have spent their careers on exactly this question, more interesting than any confident answer either of them could to this picture is quieter than Rovelli’s or Smolin’s, but cuts just as deep. His argument, laid out in his 2000 book, identifies what he calls the past hypothesis. The single assumption that does all the work in our experience [music] of time. The assumption is this. The universe began in a state of extremely low entropy. Extraordinarily, improbably, almost unimaginably low entropy. And from that starting point, [music] entropy has been increasing ever since. Not because any law requires it, as Boltzmann [music] showed, but because there are overwhelmingly more high entropy states than low entropy ones, and random processes move toward the more probable. That is the entire explanation for why you remember the past and not [music] the future. Why causes precede effects. Why broken things do not spontaneously reassemble. Why time feels like it has a direction. All of it is downstream of one fact. The beginning was very ordered. Remove that one fact, keep everything else identical, and the arrow of time disappears. Past and future become indistinguishable. Memory becomes meaningless. Causality collapses. Everything that makes time feel like time depends on that single initial condition. Albert calls it the past hypothesis because [music] it is not derived from any deeper principle. It is an assumption we make because the evidence is consistent with it, not because any theory predicts it. And I find myself sitting with the strangeness of that for longer than is probably comfortable. Because what it means is that every experience you have ever had, every memory, every plan, every sense that there was a yesterday and there will be a tomorrow is not grounded in a law of nature. It is grounded in a statistical accident at the beginning of everything. The directionality of your life, the feeling that time is carrying you somewhere, the entire structure of before and after that makes human experience coherent. All of it rests on one unexplained initial condition that nobody chose and nothing required. And then the question Albert leaves open is the one that nobody has answered. Why was the beginning so ordered? What selected that initial condition out of the vast space of possible initial conditions? Penrose offers an answer through cyclic cosmology. The ordered beginning was a mathematical consequence of the disordered end of a previous eon. Hawking offers one through imaginary time. The beginning was a smooth geometric transition from a state where time and space were equivalent. Both answers are serious. Both involve real mathematics. But both answers assume something about time in order to explain time. Penrose needs eons to have a sequence. Hawking needs imaginary time to transition into real time. Neither one escapes the circle. They push it back one step, which is not nothing, but it is not a solution either. What I keep returning to is a simpler and more uncomfortable version of the question. If the low entropy beginning is not explained by any law, and if the arrow of time is entirely the product of that beginning, then the direction of time is not a discovery about the universe. It is a statement about where we happen to be in a process that could, in principle, have started anywhere. We experience time as moving forward because we are downstream of an unusual starting point. But unusual from whose perspective? From the perspective of all possible initial conditions, our beginning was extraordinarily improbable, which raises a question I genuinely do not know how to frame properly. Is the past real because it happened? Or [music] does it feel real because we are built to remember in one direction only? And if the latter, if our sense of the past is a cognitive consequence of thermodynamics rather than a direct perception of something that actually exists, [music] then what exactly are we accessing when we remember anything at all? John Wheeler spent decades pushing a single idea that most of his colleagues found too strange to engage with seriously, and that the experiments of the last decade have made increasingly [music] difficult to ignore. He called it it from bit. The idea is that information [music] is not a description of physical reality. It is what physical reality is made [music] of. That every particle, every field, every physical quantity derives its existence from answers to yes or no questions, from binary choices, from bits. The universe is not a collection of things that can be described with information. It is a collection of information that we interpret as [music] things. I want to stay with that for a moment because it is easy to let it slide past as an interesting theoretical position without fully registering what it would mean if it were correct. If information is fundamental, if the universe is, at its deepest level, built from answers to questions, then the question of who or what is doing the asking is not a side issue. It is the central issue. A bit is not a bit in the absence of something that distinguishes between the two states. A yes or no question requires something that can ask it, which means Wheeler’s framework does not just describe physics. It 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 [music] showed, if the direction of time is a product of the act of measurement, of the creation of records, of the thermodynamic cost of looking, then the observer is not just a witness to time. The observer is part of what produces it. This is not mysticism. It follows from the mathematics of quantum mechanics and from the experimental results of 2013 and 2025 taken together. The Marivor experiment showed that time exists only for internal observers. The Oxford experiment showed that the arrow of time emerges from the act of observation. Wheeler’s framework gives these results a coherent interpretation. The universe, at its deepest level, is timeless. Time is what it looks like from inside when something is doing the looking. But here is where I find myself genuinely uncertain in a way that I do not think is just a gap in my understanding. Wheeler’s picture raises a question about sequence that I cannot resolve. If observers produce time, if the arrow of time requires something that measures and records, then observers had to exist before [music] time had a direction. But observers are physical systems. Physical systems exist in time. Something that exists in time cannot be prior to time. The logic seems to circle back on itself at the exact moment it becomes most interesting. And I do not think this is a failure of the framework. I think it might be pointing at something about the relationship between time and observation that our current concepts are not equipped to describe cleanly. Wheeler himself asked the question in its sharpest form near the end of his life. Did the universe require observers in order to have a past? Not just a future. [music] A past. His delayed choice experiment suggested that the act of observation can, in a precise quantum mechanical sense, reach backward [music] and determine how a particle behaved at an earlier moment. The past in those [music] 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. [music] And if that is true, then asking what time is maybe inseparable from asking what memory is, from asking what it means to know that something happened, [music] from asking whether the difference between the past and the future is a feature of reality or a feature of minds embedded in reality. Is memory not just a record of time? Or is memory the thing that makes time real in the first place? Here is where we actually are, not where the textbooks say we are. 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 run 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 independent measurements of the expansion rate of the universe disagreeing by five sigma. We have a theoretical picture in which time emerges from thermodynamic complexity, [music] and a competing picture in which time is the only fundamental thing, and both are consistent with the available data. And we have an unsolved problem at the center of all of it. Why did the universe begin in such an ordered state that none of the frameworks can answer. [music] What we do not have is a clear, consistent, experimentally confirmed account of what time is. And I want to sit with that for a moment rather than move past it too quickly because I think there is a way of hearing that sentence that makes it sound like a temporary situation. A gap in knowledge that will be closed by the next generation of experiments, the next theoretical breakthrough, the [music] next decade of data from ACs or DESY or the nuclear clocks. And maybe it will be. But I’m not sure that is the right way to read [music] where we are because the deeper the experiments go, the more the question ramifies rather than converging. Every result that was supposed to clarify the nature of time has produced a new version of the puzzle. The more precisely we measure, [music] the more precisely we can see that we do not understand what we are measuring. What strikes me about all of this, and I keep coming back to this, is that the question of what time is has turned out to be the same question as what an observer is, what information is, what the past is, what it means for something to exist rather than not exist. These are not separate questions that happen to be related. They seem to be the same question approached from different directions. And if that is right, then the answer to what time is will not come from a better clock or a more precise measurement alone. It will require a new way of thinking about what physical reality is at the level where time, information, and observation meet. We do not have that yet. We have pieces of it scattered across a hundred years of experiment and theory pointing in roughly the same direction without quite [music] connecting. I do not know what that theory will look like when it arrives. I do not know whether it will dissolve the question of time the way Hawking and Penrose dissolved the question of what came before the Big Bang by showing that the question was malformed, or whether it will answer it directly. I do not know whether time will turn out to be fundamental or emergent, [music] real or relational, a river or heat or something we do not yet have a metaphor for. What I do know is that every experiment running right now, ACES on [music] the space station, the nuclear clocks in Braunschweig and Boulder, the quantum network protocols being built in laboratories across three continents, is probing the same boundary from a different angle. The answer, if there is one, is somewhere in the data that is being collected right now. Or is somewhere in the data that will force us to ask a question we have not thought to ask yet. Either way, the clocks are running. The numbers still do not add up. And somewhere in the gap between what the measurements show and what the theories predict, the actual nature of time is waiting. Whether we are close to finding it or just beginning to understand how deep the search goes, that I genuinely do not know.