218 | Raphael Bousso on Black Holes and the Holographic Universe
What this covers
Raphael Bousso, a physicist who studied under Steven Hawking at Cambridge, discusses with Sean Carroll how gravity encodes information about quantum states in ways no other classical theory does. The conversation centers on Hawking radiation, black hole entropy, and the black hole information paradox—the puzzle of whether information truly vanishes when a black hole evaporates. Bousso traces how Bekenstein-Hawking entropy demonstrated that gravity can count quantum states without any microscopic theory of what composes the black hole, a feat impossible in classical electromagnetism. The heart of the conversation is the Holographic Principle: information describing a region is bounded by its surface area rather than volume, implying the universe is fundamentally non-local in ways that contradict the intuition of quantum field theory.
The discussion moves through recent theoretical breakthroughs and their limits. Bousso describes how Jeff Pennington, Netta Engelhardt, and collaborators recently derived the Page curve—showing information escapes black holes—using gravity alone, confirming decades of conjecture. He explains the AdS/CFT correspondence, which realizes holography precisely in idealized anti-de Sitter universes, but emphasizes the central obstacle: extending these results to our actual universe with its positive cosmological constant and de Sitter space geometry remains unsolved. The conversation touches the firewall paradox—a sharpened version of the information puzzle that pits the outside observer's view against general relativity's insistence that falling into a black hole feels normal—and Bousso's skepticism toward proposed resolutions. Throughout, he argues for Economical Description over questions of absolute reality, comparing successive theories like those of Ptolemy, Kepler, and Einstein as increasingly efficient explanatory frameworks rather than revelations of truth.
Bousso argues that gravity, uniquely among classical theories, secretly encodes information about quantum states (via black hole entropy and the holographic principle), and that recent gravity-only calculations show black hole information is recovered—but extending this from idealized anti-de Sitter universes to our own positive-cosmological-constant universe remains the central open challenge.
- Bekenstein-Hawking entropy proves gravity can count quantum states without a microscopic theory
- The holographic principle bounds information by surface area, not volume, implying deep non-locality
- The Page curve was recently derived from gravity alone, confirming information escapes black holes
The holographic principle shows gravity encodes quantum mechanics via boundary information, realized in AdS/CFT.
- Whereas quantum field theory is quintessentially local, the holographic principle tells us the world is in some deep way non-local: if all information about a spacetime region can be stored on its boundary, then reading the boundary lets you instantaneously read off what is happening some distance away in the interior—which is literally realized in AdS/CFT.
“Gravity is telling us that the world is in some deep way non-local... the holographic principle tells us that at some level it has to be wrong. If I can store all of the information about some space-time region on its boundary, well, that means that in some sense I should be able to look at this boundary and read off information about what's going on instantaneously some distance away from it in the interior.”
- Anti-de Sitter space has a negative cosmological constant and behaves like a box: anything you throw away comes back oscillating like a pendulum with a frequency independent of how hard you throw it, and this boundary structure is what allows the holographic CFT description to live on its edge—unlike our de Sitter universe where objects accelerate away forever.
“anti-de Sitter space is kinda like a box. You're closed in a box... no matter how hard you throw away your ex-friend, they're gonna come back to you... the time that it takes for them to come back won't depend on how far away I throw them.”
- The holographic principle states that the number of quantum states (information content) in a region is bounded by the area of a surface, not its volume—because if you tried to store a volume's worth of information you would make a black hole first, whose information content is set by its surface area—and Bousso turned this into a precise statement using the largest area on a light cone, giving the observable universe an entropy of no more than about 10^120.
“you would have thought that the amount of information you can have in a region should grow like the volume, not like the area... that's another surprising feature, and that's why it's called the holographic principle. It's a bit like the information is stored on the boundary or on some surfaces.”
- The recent progress in understanding how gravity encodes quantum mechanics could in principle have happened entirely without knowing anything about string theory; while string theory's tools historically enabled the work, the underlying properties of gravity that allow it are logically independent of string theory.
“the kind of progress that we're making right now could have, in principle happened entirely without knowing anything about string theory”
- String theory has provided the only consistent, complete candidate theories of quantum gravity, most importantly through the AdS/CFT correspondence, which gives a complete quantum theory of gravity for a class of universes—one you could put on a computer and run—even though that class does not include our own universe.
“String theory, I think, has given us from my perspective, as far as I can tell, the only consistent candidates for theories of quantum gravity that are in any sense complete. In particular, it has led to the discovery of the AdS/CFT correspondence.”
- The frequency of a pendulum does not depend on its amplitude, which is why Galileo used pendulums as clocks—though for large displacements this breaks down, whereas in anti-de Sitter space the analogous return time is exactly independent of how far an object is thrown.
“the frequency of a pendulum doesn't depend on the amplitude, which is why Galileo used them as clocks... If you move a pendulum too far out, then actually you start seeing that it's no longer true. But in anti-de Sitter space, this is just exactly true.”
An unresolved conflict exists between general relativity's smooth horizons and quantum mechanics' information preservation.
- General relativity is absolutely insistent that crossing the horizon of a large black hole is like any other empty space—there is no sign, no slap, nothing special—so we should be able to cross the horizon of a very large black hole right now without knowing it; this smoothness is what conflicts with the membrane/firewall picture.
“Einstein's theory, general relativity... is absolutely insistent that crossing the horizon of black hole is like empty space everywhere... We should be able to cross the horizon of a very large black hole right now without knowing about it.”
- Bousso argues that Susskind's complementarity-style approach concedes that black holes can in fact have firewalls (just trying to avoid them for freshly-formed, not insanely old, black holes), and to Bousso this concession is unacceptable because once firewalls can exist at all it is a terrible affront to general relativity regardless of when they appear, and we have no idea what the dynamics of firewall formation would be.
“I think, for example, Lenny would concede that black holes, in fact, can have firewalls. He's just trying to save us from having firewalls when the black hole has maybe freshly formed... once you can have firewalls, it's just a terrible affront to general relativity.”
- When two deep physical principles conflict, this is a great opportunity for progress, but a resolution should come down on one side—breaking quantum mechanics or gravity decisively—rather than breaking both 'a little bit'; Bousso argues approaches that allow firewalls even in old black holes are an unacceptable affront to general relativity, and the worst approaches violate both.
“I think, for example, Lenny would concede that black holes, in fact, can have firewalls. He's just trying to save us from having firewalls when the black hole has maybe freshly formed... once you can have firewalls, it's just a terrible affront to general relativity.”
- There is a genuine conflict—the firewall paradox, a sharpened black hole information paradox—between the outside observer's view (information comes out, so it must be stored on the horizon like a membrane) and general relativity's insistence that an infalling observer crosses the horizon and finds nothing special; physicists care about both quantum mechanics and gravity, and there is currently no resolution that fully preserves both principles.
“those two principles are in a serious conflict with one another... It's called the... In its old form, it's called the black hole information paradox... And the modern incarnation of that is called the firewall paradox.”
- There are two reasonable possibilities about quantum gravity: a pessimist view in which every new theory only poses new questions leading to the next frontier forever, or a view that successfully unifying gravity and quantum mechanics might genuinely be 'it'—the end of fundamental physics—though consequences and other sciences would remain, and we cannot describe a zebra's behavior from fundamental physics.
“if you figure out how to put together gravity with quantum mechanics, then at least on the... front of trying to understand nature fundamentally, you would be done... But it might be the end.”
Physics progresses by adopting simpler, more economical theories, not by discovering absolute truth.
- Bousso argues that the debate about what is 'real' is meaningless; physics instead asks what are the most efficient, economical ways of describing nature and explaining experiments—just as the geocentric Ptolemaic model was not wrong but merely less economical than heliocentrism, and Kepler's ellipses were superseded by Einstein, each successive theory explaining more data with fewer ingredients without the prior one being simply 'wrong.'
“this debate about what's real, I mean, to me anyway, is completely meaningless. What we can ask about is, what are efficient ways of describing nature? What are... economical ways of explaining the data that we have”
- Although the geocentric model is not wrong, thinking about phenomena in a simple rather than convoluted way is vital for progress: Newton could not have derived his force law from Ptolemy's epicycles, but once Kepler placed the sun at the center with elliptical orbits, it became much easier to reverse-engineer the force producing planetary motion.
“It's vital for progress that you think about it in a simple way and not in a convoluted way... I don't think Newton could have gone from Ptolemy to his force law that tells you why the planets move like that around the sun. But once you had Kepler... It's much easier to figure out like, okay, what do I have to put in to reverse engineer the force”
- Quantum mechanics, unlike relativity, could never have been discovered by pure abstract thought about contradictions between theories the way Einstein discovered relativity; it is so far from everyday experience that experimental data—atomic spectra, blackbody radiation—were needed to drag physicists 'kicking and screaming' to its counterintuitive conclusions like superposition.
“I don't think quantum mechanics could have ever been discovered the way that Einstein did discover his stuff, by just kind of thinking abstractly... Quantum mechanics is so crazy that it had to be discovered by experiments... you had to have experimental data dragging you kicking and screaming to this crazy theory”
Black holes have finite quantum states; recent gravity calculations confirm information escapes, resolving the information paradox.
- Hawking initially calculated that a black hole formed from a definite state radiates into a mixed state of many possibilities, meaning the information about the initial state is lost—you always get the same garbage out no matter what you put in—which is the core of the nearly 50-year black hole information loss debate.
“Hawking initially calculated it radiates into radiation, which is in a mixed state, lots of different possibilities and so you've lost the information about what that initial state was. You always get the same garbage out, no matter what you put in.”
- Bekenstein and Hawking calculated the coarse-grained entropy—the number of quantum states—of a black hole knowing only coarse features like its mass and spin, without knowing the analogue of atoms or being able to experiment on one, which is analogous to counting the quantum states of a block of metal without knowing what metal it is or having a sample.
“Bekenstein and Hawking were able to calculate how many quantum states it has. And so that's really analogous to, being able to figure out how many quantum states a block of metal has without me telling you which metal it is, what the atoms are, and without me giving one to you so you can experiment on it”
- Jeff Pennington, Netta Engelhardt and collaborators recently and independently discovered how to derive the Page curve using just gravity—by sewing together multiple copies of the spacetime—showing the entropy of Hawking radiation rises and then falls back to one, providing the first direct gravity calculation confirming that information is not lost in black holes.
“a group of four other physicists, simultaneously and independently discovered a way to derive this Page curve using just gravity. So that was a huge breakthrough, because among other things, it tells you that the entropy or the number of possible states at the end, when everything is just radiation, the black hole is gone, is one.”
- Gravity, unlike any other classical theory, knows something about its own quantum states and the quantum states of all matter fields—for example Bekenstein-Hawking calculated how many quantum states a black hole has without any microscopic theory of what black holes are made of, a feat impossible for classical electromagnetism.
“gravity, Einstein's theory, bizarrely, knows something about its own quantum states, and actually the quantum states of all the matter fields and forces that exist in nature”
The string landscape with infinite-dimensional Hilbert space is preferable to finite-dimensional quantum gravity for our universe.
- The Banks-Fischler proposal that quantum gravity for our universe has only a finite number of states (a finite-dimensional Hilbert space matched to our cosmological constant) was compelling enough to pursue, but the Dyson-Kleban-Susskind argument that such a theory is inconsistent with observation led Bousso to reject it and embrace the string theory landscape; he now guesses the fundamental Hilbert space is infinite-dimensional, while cautioning that whether quantum gravity even uses standard quantum mechanics is unsettled.
“it was only this argument of Dyson, Kleban and Susskind that that theory is necessarily inconsistent... with observation. And at that point, I guess I embraced the landscape and moved on.”
- String theory suggests the particles and forces we see are no more unique than which atoms or molecules exist—they may be assembled from more fundamental string-theory ingredients that could be put together differently in far-away parts of the universe, helping explain puzzling features of our universe such as how it got to be so old and large.
“string theory suggests to us that the so called fundamental particles that we see might actually really be things that were put together from even more fundamental ingredients that exist in string theory, and that could have been put together in different ways, and probably in some far away parts of the universe that we can't see yet are put together in different ways”
- Bousso believes it is extremely unlikely our universe fundamentally has a positive cosmological constant forever, for two converging reasons: the Bousso-Polchinski landscape proposal (the only viable explanation of the small observed value) implies many possible values our universe can eventually decay through, and the Dyson-Kleban-Susskind argument shows an eternal de Sitter universe is a warm box that would, over infinite time, produce everything infinitely often, making our observations have unbelievably small probability—grounds to reject such a theory.
“it's extremely unlikely that our universe is fundamentally of the type where you have a positive cosmological constant forever”
Extending holography from AdS to realistic de Sitter universes requires new rules and remains fundamentally harder.
- Bousso and Jeff Pennington recently published a rule, logically independent of AdS/CFT though motivated by it, specifying how far away you can create something (like an elephant) via a non-local operation if you control a certain region of spacetime—aiming to extend holographic reasoning to realistic universes like our own; the rule reproduces known AdS results as a consistency check.
“we were able to give a rule for not exactly how that should work, but for how far away you can make an elephant if you have control over a certain region of your universe. And hopefully, by studying that pattern, we can learn more about how quantum gravity actually works in more realistic situations like our own universe.”
- Doing for our universe what AdS/CFT does for anti-de Sitter is hard not only because we lack a quantum gravity theory but because it is unclear what such a theory should compute: AdS has a boundary where idealized exact observables and arbitrarily precise measurements can be defined, but our de Sitter-like universe has no such boundary, so it is not clear that exact idealized observables even exist.
“The problem is worse because it's not totally clear what we would want that theory to compute for us... in our universe, it's not clear whether such a thing exists and what it would be... anti-de Sitter space is like a box. It has a boundary and the boundary is a good place to go if you want to make things exact”
- Acting with a boundary quantum operator in AdS/CFT can instantaneously change the interior state from empty to containing an elephant, but the correct interpretation is not that an elephant suddenly appeared in violation of relativity—rather, there is no consistent classical spacetime picture for a while; in effect the universe has been destroyed and recreated with the elephant in it.
“There is in fact no consistent way of thinking about this as space and time and the way we like to classically think about it for a while. What is really happening is that, we've destroyed the universe and we've recreated it with an elephant in it.”
- The path-integral gravity approximation is far more powerful than realized—it can answer specific single-number questions like the coarse-grained entropy or the entropy of Hawking radiation versus time—but it is likely not powerful enough to answer the vastly more complex question of exactly how information gets out or precisely which quantum state the radiation is in, meaning a full quantum gravity theory is still needed.
“the difficulty of explaining how the information gets out and in which precise state the radiation actually is, that looks like it's a significantly more complex question. It's not clear that we can ever make gravity by itself smart enough to answer that question. It's not like we're done. We still need to discover a quantum gravity theory.”
- If we actually lived in an asymptotically anti-de Sitter universe—our galaxy in a boxy AdS world without the surrounding matter—human evolution and everything would have happened exactly the same way; we would discover classical gravity first, then quantum mechanics, possibly AdS/CFT, and only upon building telescopes discover we got lucky living in the very universe for which we already have a complete quantum gravity theory.
“you could put our galaxy there and just don't put any of the other stuff. And it would have made absolutely no difference for the evolution of humans... we look at our universe and we discover, oh, it's anti-de Sitter space. Bingo. We got super lucky.”
Physical crises from conflicting principles drive progress when resolved by breaking one side decisively.
- In physics, crises—where two deep, otherwise-successful principles are found to be in conflict—are welcomed because they sharpen the problem and tell us exactly where to dig, as exemplified by Einstein resolving the conflict between Newtonian mechanics and Maxwell's electromagnetism by identifying a guiding principle and constructing a bigger, better theory.
“In physics, we like crises because it sharpens things and then tells us, okay, here's where we have to dig.”