Imagine a power plant that extracts 100% of the energy stored in matter itself — not from chemical bonds, not from splitting atoms, but from the rest mass that Einstein’s most famous equation told us was there all along. Every kilogram of fuel would yield 9 × 10¹⁶ joules, roughly the energy consumption of the entire United States for two weeks. No such machine exists on Earth, but the physics that permits it has been understood for over a century. The machine is a black hole, and the physics is general relativity.
According to Adam Brown, a physicist at Google DeepMind and former Stanford professor who studies black holes and the early universe, the path from Newton’s apple to that conclusion is remarkably short — provided you follow the right clue. Speaking on the Dwarkesh Podcast, Brown laid out a derivation of general relativity from first principles, showing that the theory’s seemingly exotic predictions all flow from a single observation that was sitting in plain view since 1687.
The Coincidence Newton Never Questioned
Every high school student learns Newton’s second law: F = ma. Push a shopping cart, and its acceleration depends on its inertial mass — the mass that resists changes in motion. Every student also learns Newton’s law of gravity: the attraction between two objects depends on their gravitational mass. For two and a half centuries, nobody asked why these two masses should be the same.
They didn’t have to be. In electromagnetism, the charge of a particle and its mass are completely independent. “You can have heavy objects that have no charge, like the neutron,” Brown noted. “You can have light objects, like the electron, that have high charge. There is no necessary relation between the charge of a particle and its mass.”
Gravity alone breaks this pattern. “There is another difference between the gravitational force law and the electrostatics, and that is this object that plays the role of the charge, the analog of the charge in electrostatics and gravity, and it’s the fact that it’s the mass sitting here,” Brown said. The gravitational “charge” is the inertial mass itself. This identity has now been verified to one part in 10¹⁵ — a precision equivalent to measuring the distance from New York to Los Angeles to within the width of a human hair.
Einstein called this realization his “happiest thought.” If inertia and gravity are fundamentally the same thing, then gravity might be what physicists call a fictitious force — not a genuine interaction between objects, but an artifact of observing from the wrong reference frame. The centrifugal force you feel in a turning car is fictitious in exactly this sense: it is proportional to your inertial mass because it is your inertia resisting the curved path. Gravity shares the same signature.
The implication is radical. If gravity is a fictitious force, then the people normally considered to be accelerating — standing on Earth’s surface, pressed into chairs — are actually the ones deviating from straight-line motion. The freely falling astronaut, weightless and drifting, is moving in the truest straight line that curved spacetime permits. Matter curves spacetime; curved spacetime tells matter how to move. That is the entire theory.
Three Formulas That Explain Everything
In 1916, while serving as a Prussian artillery officer on the Eastern Front, Karl Schwarzschild solved Einstein’s equations for the simplest case: the spacetime around a single spherical mass. His solution produced a family of formulas, each governed by the same square-root factor, each describing a different physical phenomenon — and each collapsing at a specific radius where the denominator goes to zero. That radius is the event horizon of a black hole.
FormulaBehavior Far from the MassBehavior at the Event HorizonGravitational acceleration for a stationary observerReproduces Newton’s inverse-square law: GM/r²Goes to infinity — infinite force required to stay putGravitational time dilationNegligible; factor ≈ 1 with only a 10⁻¹⁰ correction at Earth’s surfaceFactor goes to zero — time appears to stop for a distant watcherRedshifted energy of a mass lowered toward the surfaceRecovers Newtonian potential energy: mc² × (1 − GM/rc²)Goes to zero — the mass’s entire rest energy becomes accessible
For Earth, these effects are tiny. “You’ve extracted a fraction 7 times 10 to the minus 10 of the original rest mass energy of the brick,” Brown said, describing how much energy you could recover by lowering a brick from infinity to Earth’s surface. But Earth’s gravitational binding energy and the chemical energy in rocket fuel happen to be almost identical. The best oxygen-hydrogen combustion releases about 1.5 × 10⁻¹⁰ of its rest mass as useful energy. “These two small numbers are almost exactly equal to each other, which is why we can use chemical rockets to get to space, but it’s hard,” Brown explained. Rockets are barely adequate for escaping Earth because the energy density of their fuel and the gravitational debt they must pay are nearly the same number. A planet only slightly more massive would be unreachable with chemical propulsion.
The Sun, by contrast, has a much deeper gravitational well. Light climbing out of it loses a fraction of about 2 × 10⁻⁶ of its energy — two thousand times larger than Earth’s figure. This is measurable: light from the Sun’s surface is detectably redder than identical light produced in a terrestrial laboratory, a prediction known as gravitational redshift.
What You See and What You Experience
Schwarzschild’s square-root singularity creates a dramatic split between two perspectives on a black hole, a distinction that confused even Einstein, who published incorrect papers claiming that infalling objects would bounce off the event horizon.
From the outside, watching a friend fall toward a black hole, you never see them cross. Gravitational time dilation stretches their emitted light to longer, redder wavelengths as they approach the horizon. The last photon they emit arrives after infinite coordinate time. They fade from red to black, asymptotically frozen at the edge of the abyss. For you, the event horizon is a feature of infinite delay.
From the inside, crossing the horizon is uneventful. For a black hole of sufficient size — galactic mass or larger — tidal forces at the horizon are negligible. Your local clocks tick normally. You cross at finite proper time, feeling nothing. You are doomed in the sense that every future path now leads to the singularity, but for a supermassive black hole hundreds of light-years across, you could live a full human lifespan and raise descendants before the crushing tidal forces become fatal near the center.
This duality — coordinate time versus proper time — was the conceptual gap that early relativists failed to bridge. It was not until Roger Penrose and Stephen Hawking proved singularity theorems in the 1960s that black hole formation was understood to be generic, not a mathematical curiosity requiring unrealistic symmetry. Black holes are not exceptions to physics; they are inevitable predictions of it.
Brown invoked the Schwarzschild energy formula to resolve a paradox that Newtonian physics cannot address. If you lower a brick on a rope toward a compact mass, classical reasoning suggests you could extract more than 100% of the brick’s rest-mass energy — an apparent violation of energy conservation. General relativity closes the loophole. “The force gets so strong when you try and get within this radius that in fact you cannot slowly lower the brick down towards the surface because you’ve formed a black hole,” Brown said. “The gravitational force becomes infinite a finite distance away.” The rope breaks, or the brick is ripped free, before the paradox becomes real. The maximum energy you can extract by lowering a mass to just above the horizon is exactly 100% of its rest energy — no more, and with ideal efficiency, no less.
A Century of Verification
General relativity has progressed from a controversial theory confirmed by a single eclipse expedition to a routinely used engineering tool. The arc of its verification is one of the longest and most thorough in scientific history.
The 1919 Eddington expedition remains the most famous single experiment. During a total solar eclipse, Arthur Eddington measured the apparent position of stars near the Sun’s limb. Newtonian physics, treating light as massive particles, predicts one deflection angle. General relativity predicts exactly double that value — because space itself is curved, not just the trajectory through space. Eddington’s plates confirmed the GR prediction, making Einstein an international celebrity overnight and giving the result political weight as a rare moment of Anglo-German scientific collaboration in the aftermath of World War I.
But the Mercury anomaly was actually the first correct prediction. For decades before 1915, astronomers had observed that Mercury’s perihelion — its point of closest approach to the Sun — shifted by 43 arcseconds per century more than Newtonian mechanics could explain. GR matched that number exactly. Einstein later said that when he saw the calculation work, he had heart palpitations.
The modern era has extended verification to regimes Einstein could barely have imagined. LIGO’s first gravitational-wave detection, GW150914, recorded on September 14, 2015, captured the final inspiral and merger of two black holes of 36 and 29 solar masses located 1.3 billion light-years away. The signal had been traveling toward Earth since before complex multicellular life evolved. LIGO has now detected thousands of such events. Gravitational-wave astronomy is no longer a novel experiment; it is an operational observational science.
The Event Horizon Telescope — a planet-scale interferometer linking radio dishes across four continents — produced the first direct image of a black hole in 2019, showing the shadow of the 6.5-billion-solar-mass black hole at the center of the galaxy M87. In 2022, it imaged Sagittarius A*, the 4-million-solar-mass black hole at the center of our own Milky Way. Both images show the photon ring — light bent by gravity into a bright circle around the dark silhouette of the event horizon — matching GR predictions at roughly the 10% level.
Closer to daily experience, the Pound-Rebka experiment of 1959 confirmed gravitational time dilation using two atomic clocks separated by 22.5 meters in a Harvard University building. This effect is now corrected for in every GPS satellite, which must adjust for the fact that its onboard clock runs faster than ground clocks by about 38 microseconds per day due to the combined effects of special and general relativity. Without this correction, GPS would accumulate errors of roughly 10 kilometers per day — rendering it useless for anything beyond crude navigation.
The Ultimate Power Plant
The energy extraction efficiency of a black hole, derived from the same Schwarzschild formulas, exposes an enormous qualitative gap between all terrestrial energy sources and what gravity can achieve. Efficiency here is defined as the fraction of the fuel’s rest-mass energy — its mc² — that can be converted to useful work.
Energy SourceFraction of mc² ExtractedMechanismChemical (combustion, batteries)~10⁻¹⁰Rearranging electron bonds; nuclear rest mass untouchedNuclear fission (uranium)~10⁻³Splitting heavy nuclei; ~0.1% of rest mass releasedNuclear fusion (hydrogen → helium)~10⁻²Combining light nuclei; ~0.7% of rest mass releasedBlack hole (lowering mass to just above horizon)Approaches 1.0 (100%)Extracting gravitational potential energy from curved spacetime
The jump is not incremental; it is a factor of 10² from fission, 10⁸ from chemical. All terrestrial energy sources — combustion, batteries, fission, fusion — only harvest the energy stored in electromagnetic or strong nuclear bonds between particles. They never touch the roughly 99% of energy locked in the rest masses of protons and neutrons themselves. Black hole extraction does touch that reservoir, because in general relativity the mass-energy of the infalling matter is transferred to the gravitational field — and can be extracted before the matter crosses the event horizon.
Brown noted a further implication: if you lower a brick made of equal parts matter and antimatter to just above the horizon, annihilating it would normally produce a flash of gamma rays carrying 100% of the rest energy. But the rope-and-pulley mechanism extracts that same 100% as mechanical work, leaving the protons, neutrons, and antiprotons to fall silently into the hole. The black hole eventually radiates that energy away through Hawking evaporation — and in doing so, violates baryon number conservation, the rule that the count of protons minus antiprotons never changes. This is one of the strongest hints from fundamental physics that no global symmetry is sacred to quantum gravity.
Can Pure Thought Do It Again?
Brown dedicated a substantial portion of his analysis to a methodological question that has become urgent as AI systems begin generating physical theories. General relativity is the strongest case in history for what might be called the “single mind in a cave” approach: Einstein needed only the finite speed of light and the equivalence principle — an empirical fact, but one known to exquisite precision without building any new instrument — to narrow the space of possible relativistic gravity theories to a very small number, possibly one.
But Brown was careful to add that this precedent misleads more than it instructs. Einstein’s later career — his failed attempts at a unified field theory, his wrong papers on black holes — demonstrates that even the greatest theoretical physicist could be led astray without experimental feedback. The history of physics is mostly the opposite of the 1915 story: most theories require constant, iterative confrontation with data.
This is especially relevant for string theory, which Brown described as “going all in” on the pure-thought approach. String theorists assume there is exactly one consistent quantum theory of gravity and that it can be discovered by mathematical consistency alone, without experimental input at the Planck scale — the energy of 10¹⁹ billion electronvolts, unreachable by any conceivable particle collider. Whether that assumption is correct is the central open question in fundamental physics. If there are many consistent theories of quantum gravity, pure thought cannot choose among them, and the field will need experimental guidance it may never get.
Brown suggested that AI systems capable of exhaustively exploring the space of consistent theories — and producing human-comprehensible explanations of their findings — may represent the closest approximation to another Einstein. The recent solution of the Erdős discrepancy problem by an AI that communicated its proof in human-readable mathematics is an early signal. But the core tension remains: without experiment, even a perfect theoretical exploration might discover many possible universes rather than singling out the one we inhabit.
The arc from 1907 — when Einstein first grasped that a falling man feels no gravity — to 2026, with LIGO’s thousandth detection and the Event Horizon Telescope’s direct images of black hole shadows, spans 119 years. A single person, working with pen and paper, wrote down a theory that has been confirmed across twenty orders of magnitude in scale, from the orbit of Mercury to the collision of black holes a billion light-years away. Whether another such theory could be discovered in the same way today is the open question that haunts the field. General relativity is either a permanent monument to what the unaided human mind can achieve, or the last of its kind — and no one yet knows which.