What this covers

Sean Carroll talks with Vitor Cardoso, director of the Center of Gravity in Copenhagen, during the center's Black Hole Week. The conversation opens with history. Albert Einstein did not accept black holes and tried to show that collapse would stop short of a horizon, and Arthur Eddington also doubted that nature allows eternal collapse. Cardoso then explains why collapse wipes out the details of stellar matter, leaving the Kerr family described by mass and spin. He adds that the dimensionless spin is tiny, so a child's top beats any black hole.

The middle of the discussion is about what observation has done to the subject. Cardoso describes the light ring and the shadow, and says the Event Horizon Telescope images are very blurry, with roughly 90 percent of each reconstructed from simulations. On gravitational waves, he stresses that matched filtering demands percent-level predictions from general relativity. He recalls experts in 2001 forecasting that detectors would see only earthquakes, and notes that LIGO, tuned for neutron stars, found black hole mergers. Among hundreds of mergers, none has had an electromagnetic counterpart, and LISA is meant to reach supermassive sources.

The open questions are where it turns contentious. Cardoso covers the black hole bomb mechanism of superradiant amplification and the unresolved cosmic censorship conjecture, which fails in higher dimensions. He also raises unknown routes to billion-solar-mass black holes, the unsolved final-parsec merger problem, and primordial black holes as dark matter, which he calls a desperate remaining corner. Quantum gravity, he says, offers no concrete predictions yet.

Sharpest takeaway

Cardoso argues that black holes are the rare objects whose simplicity (fixed by mass and spin, hidden behind horizons) makes them uniquely testable, and that gravitational-wave and imaging data have turned black hole physics into a precision, data-driven science whose deepest questions, including formation and quantum gravity near the horizon, remain open.

  • The no-hair theorem reduces vacuum stationary black holes to mass and spin, so observations can test whether real black holes match that simple description.
  • Gravitational-wave matched filtering requires precise general relativity predictions, making the field a precision science in which surprises come from new instruments.
  • Black hole formation at high and intermediate masses, and quantum gravity near horizons, remain unresolved because the relevant predictions and observations are missing.

The argument · threads9 threads · 58 claims
0.82

Observations and GR's precise predictions must test black hole theory, since data repeatedly overturn expert expectations.

9 pointscentrality 5/5
  • Since gravitational waves were detected in 2015 and black hole images and stellar flybys near black holes began appearing, purely theoretical black hole researchers have had to ask whether the phenomena they predict are actually observed, turning theory into an applied, observation-linked enterprise.

    “We started seeing black holes with gravitational waves in 2015. And in the last five, six years, we've seen images of black holes”

  • General relativity is known to break down inside black holes, which justifies checking detail by detail whether its predictions hold near black holes.

    “we also know that the theory itself breaks down inside of black holes.”

  • The no-hair proof requires vacuum and stationarity, but real black holes are embedded in matter and the universe changes in time, so the research job is to break these assumptions and determine how the object and its emission change.

    “vacuum is not a good assumption. And because we are talking, stationarity is also... Things happen in the universe.”

  • Gravitational wave astronomy relies on matched filtering, which requires general relativity predictions accurate to the percent level, so it is a precision science rather than hand-waving, and gravitational wave data is moving toward the richer inference that older optical astronomers did from light curves.

    “We know to the percent level what we're going after. And so this is not a hand-waving kind of science.”

  • Gravitational wave detection requires template banks of millions of precomputed waveforms, one for each combination of masses and spins, because a 10 solar mass merger with a 20 solar mass black hole must be distinguished from neighboring parameter values, which requires solving Einstein's equations for the full merger history.

    “you can imagine the unbelievable effort that it took us to have template banks of millions of waveforms.”

  • Cardoso knows of no alternative theory of gravity that is more elegant or solves general relativity's problems better, so the aim is to test how close data can push to confirming or refuting general relativity rather than to pick a rival theory.

    “I don't think there's any alternative to general relativity that's more elegant or that solves any of the problems that general relativity has in a better way.”

  • LIGO and Virgo were designed with a sensitivity sweet spot for neutron star masses, but they instead detected black hole merger after merger, showing that observations repeatedly surprise physicists.

    “we thought we were building it to see neutron stars, actually... we're just catching black holes after black holes after black holes.”

  • At a school in 2001 leading experts predicted that the gravitational wave detectors would only ever see earthquakes and never gravitational waves, yet fifteen years later they detected gravitational waves, showing how expert forecasts can fail.

    “I think in the end these detectors will only see earthquakes. They'll never see gravitational waves.”

  • No quantum gravity theory currently makes concrete predictions that observers could measure to discover departures from general relativity, so physics must rely on data to reveal inconsistencies with the paradigm before the theory can be refined.

    “if we had predictions from quantum gravity of what would happen and that would tell us, 'You go out there, you measure this, you're gonna find this,' but those predictions don't exist.”

0.68

Black hole surroundings, matter, formation, and growth remain poorly constrained and poorly understood.

9 pointscentrality 4/5
  • Growing heavy black holes by repeatedly merging smaller ones is difficult because black holes are very sparse and tiny relative to their mass, so they rarely encounter one another.

    “The problem with this is that the black hole is not such a populated place. So it's hard to merge two black holes.”

  • The light observed around black holes likely comes from the remnants of stars that were tidally disrupted and formed an accretion disk, where friction heats the material until it glows, and both the Milky Way and M87 show transient flaring episodes.

    “the actual light that we're seeing comes from possibly remains of stars that were tidally disrupted and formed what we call an accretion disk around black holes, and friction heats up the material.”

  • Observations of the early universe show objects that appear to be black holes far more massive than expected and formed much earlier than expected, possibly through collapse of dust clouds, though this remains an ongoing debate.

    “if we look at the beginning of the universe, we are also seeing objects that seem to be black holes way more massive than we thought they could be, way earlier than we thought they would form.”

  • Across the two or three hundred stellar-mass black hole mergers seen, no electromagnetic counterpart has been observed, implying that stellar-mass black holes have little surrounding matter because few mechanisms produce a substantial accretion disk around them.

    “we've seen two or three hundred of these guys merging, and we have never seen an electromagnetic counterpart”

  • It is unclear whether dark matter near a black hole behaves like ordinary dust that piles up and orbits, or like a field or long-wavelength radiation that hovers around it, and modeling it is a significant challenge that is being actively addressed.

    “Is dark matter like normal matter, like dust, that we can somehow model... Or is it something a bit different, like a field”

  • How supermassive black holes grow to a billion solar masses is not fully known, and there is no robust mechanism for bringing two supermassive black holes close enough to merge across the last parsec, though galaxies are known to pile black holes at their centers.

    “We don't have yet a very robust way of getting them across the last parsec, actually, of distance.”

  • Stars above roughly 80 solar masses become unstable, so there is no elegant way to produce a stable star that burns its fuel and then collapses into a black hole across the mass gap, and current theory suggests a large cloud of that mass would fragment into several stars.

    “if I try to form one of these stars above 80 solar masses, for example, the star becomes unstable.”

  • Gravitational wave observations have revealed black holes across a wide mass range up to about 120 solar masses, including masses in a predicted mass gap, and their formation is still not well understood.

    “We see black holes exist all the way up to 120 solar masses. We don't really know how they form.”

  • A black hole's rotation is inherited from its progenitor star's angular momentum as it collapses, but black holes tend to spin slowly because general relativity caps their rotation, while typical stars carry much more angular momentum than a black hole of the same mass can hold, so stars must shed angular momentum in winds during collapse.

    “rotation is inherited from the progenitor... black holes spin slowly. There's also a result in general relativity that says that the rotation of a black hole is limited by an upper number.”

0.57

Current and upcoming instruments will probe black holes across scales, frequencies, horizons, and early-universe physics.

10 pointscentrality 3/5
  • The GRAVITY instrument, focused on the Milky Way's center, works in the infrared, where electric field oscillations are too fast to record directly, so it combines beams from four telescopes on the spot in one go instead of recording and combining later.

    “So oscillations in the infrared, oscillations of the electric field, are just too rapid for us to collect the electric field. So they need to combine on the spot.”

  • The Event Horizon Telescope targets the black holes in M87 and the center of our galaxy because they are the largest black holes on the sky, with M87's black hole about a thousand times larger but also about a thousand times farther, so their apparent sizes are similar.

    “It's a thousand times larger, but the galaxy is also a thousand times farther. So roughly, the position on the sky is similar.”

  • LISA will use mirror separations on the order of a million kilometers, compared with LIGO's four kilometers, increasing sensitivity and granting access to low-frequency gravitational waves from supermassive sources because a detector's size roughly matches the size of the source it targets.

    “So LISA is gonna fly into space with a length between the mirrors of the order of a million kilometers, which we should contrast with LIGO's four kilometers.”

  • LISA might detect non-homogeneous processes from the early universe that were redshifted to lower frequencies as the universe expanded, along with whether supermassive black holes interact and merge.

    “one of them relates to the birth of the universe... non-homogeneous things that get redshifted, that get pushed to lower frequencies as the universe expands.”

  • Although the event horizon is impossible to see, the name Event Horizon Telescope is apt because the goal is to probe physics close to the horizon, where new physics is expected to appear.

    “in the end, it is, I think, a good description of what we're after. And we're after physics close to the event horizon of a black hole”

  • The most interesting open challenge is quantifying how well gravitational wave measurements, which only track how mirrors move, can establish that an object is a black hole and how deep into the gravitational well observations can probe.

    “can we quantify the evidence that the things we're looking at are black holes? How do we do that? The only thing we're doing is measuring how mirrors move.”

  • The Einstein Telescope is a ground-based interferometer like LIGO with longer arms and improved technology, to be sited in Europe, possibly Germany or Italy, and it is expected to probe different scales and possibly reveal unexpected physics.

    “In parallel, we're thinking about the Einstein Telescope, which is really a better version of LIGO. The arms are gonna be slightly longer”

  • New gravitational wave detectors include KAGRA in Japan already operating, a detector in India around 2030 built from US-origin hardware, the planned European Einstein Telescope, and the LISA space mission that Europe and the US are about to launch.

    “Some of them are online, like KAGRA in Japan is working. In late, so around 2030, India is gonna have one detector”

  • The Event Horizon Telescope is a global array of radio telescopes that combines the electric field measured at each site to produce the equivalent of a telescope as large as the whole Earth, replacing a single giant light collector.

    “we can replace a mega telescope with just an array of different telescopes spread throughout the globe and combine the data.”

  • Light passing near a black hole can orbit it on the light ring; photons inside this ring are captured and fall into the horizon, while photons directed outward from it reach observers, so the light ring defines the black hole shadow that observations image.

    “So this light ring is actually what defines the things we can see when we look at a black hole. The shadow of the black hole is governed by this light ring.”

0.56

Horizons hide black hole interiors, so quantum gravity must modify physics near horizons.

5 pointscentrality 4/5
  • A black hole is like a sealed box of precious contents that no one can open, hiding all secrets about quantum gravity, strong-field quantum effects, and the fate of collapsing stars from observers.

    “All the secrets that we think there are about quantum gravity or quantum effects in strong gravitational fields, the fate of the star that collapses, all of that is hidden from us.”

  • By definition, the event horizon is a surface beyond which no information can reach observers outside, so what happens inside a black hole is causally disconnected from us and no exterior experiment can reveal it.

    “By definition, the horizon is this surface beyond which we have no access. So it's causally disconnected from us.”

  • Einstein did not accept black holes; in a famous work he attempted to show they do not form by modeling collapse and finding that matter would reach the speed of light outside the horizon, concluding that nature must avoid such collapse, while Eddington also doubted that nature would allow eternal gravitational collapse.

    “Einstein was aware of something like a coordinate singularity or a singularity at the horizon. He didn't like it either.”

  • If an object with too much angular momentum collapsed, such as the Earth, it could not form a black hole and the equations suggest there would be no horizon, exposing the interior, which current theory cannot describe; this is why cosmic censorship is posited as an open conjecture.

    “it could not be a black hole because it's spinning too fast. And what the equations would tell us is there would be no horizon.”

  • The most serious problem in gravitational physics is the existence of singularities, which a quantum gravity theory should resolve; since singularities are hidden behind horizons, quantum gravity should also modify physics near the horizon, and hints of quantum mechanical problems there motivate searching for changes at the horizon.

    “the most serious problem we face in gravitational physics... is the existence of gravitational singularities.”

0.43

He is unmoved by confirmations of general relativity but would be excited by contradictions of it.

5 pointscentrality 3/5
  • Cardoso says he is not excited by confirmations such as equal gravitational and light speeds, because his career aims at the unexpected, and he hopes to see results that contradict general relativity or the standard picture of black holes.

    “maybe we're gonna see in our lifetime something that's totally against the lore, something that's gonna say general relativity is wrong”

  • Black Hole Week in Copenhagen is held every two years, began with the 50th anniversary of Hawking's 1974 discovery, and in 2026 runs from 22 to 29 August with events such as an opera, Science and Cocktails talks, and an immersive performance.

    “22nd to 26th, well, to the 29th of August. So the last week and a half of this month.”

  • Studies of gravitational physics in spacetime dimensions higher than four have shown cosmic censorship failing, so the conjecture is not universally valid, though its status in four dimensions remains unresolved.

    “we have seen cosmic censorship failing. We have seen all of this goes to... Yeah, doesn't work.”

  • The only electromagnetic counterpart event was a neutron star binary merger, where light and gravitational waves arrived about 1.4 seconds apart from the same event, so the gravitational wave and light speeds agree to extraordinary precision.

    “we used that beautiful event to understand that gravitational waves travel at basically the same speed as light, 'cause the interval between the arrival of a photon and of a gravitational wave was basically zero.”

  • After two black holes merge, the single resulting black hole relaxes to a quiet final state, which was the expected behavior; for the first time in history, humans observe this relaxation in gravitational waves, which happens at light speed and takes a fraction of a millisecond for a black hole the size of Copenhagen.

    “We are for the first time in the history of humankind seeing two black holes relaxing in the gravitational wave channel.”

0.29

Black holes reduce to few parameters, so surrounding details, charge, and spin matter little.

5 pointscentrality 2/5
  • A mathematical result in general relativity says that stationary vacuum black holes belong to the Kerr family, fully specified by just two parameters, mass and spin, so all black holes in the universe would be described by these two numbers.

    “black holes need to belong to a certain family. We call it the Kerr family. And the Kerr family basically is specified entirely by two parameters”

  • Electric charge is left out of the simple black hole description because the universe appears electrically neutral to a very good extent, so any charge tends to be neutralized by other charges.

    “we think that the universe doesn't really like charge very much.”

  • Stellar structure depends on details such as the equation of state, but gravitational collapse is so efficient that it removes all other details, leaving black holes as simple curled-up vacuum regardless of the matter around them.

    “because gravitational collapse acts so efficiently, it just cleans them of any detail, of any other detail.”

  • Light of sufficiently low frequency shone on a rotating black hole can come back with higher amplitude, extracting rotational energy from the black hole by slowing its spin, and if the light had mass it would fall back and be amplified repeatedly, which is called the black hole bomb mechanism.

    “if light had a weight, it would fall back, it would be amplified, and so on and so forth. We call this the black hole bomb mechanism.”

  • The dimensionless spin, angular momentum divided by the square of the mass, is very small for black holes, so a child's spinning top has orders of magnitude more of this quantity than any black hole could carry.

    “the angular momentum that this toy has is orders of magnitude above that which a black hole could carry.”

0.29

Black-hole dark matter is heavily constrained, leaving only narrow, speculative possibilities.

5 pointscentrality 2/5
  • Microlensing surveys have ruled out most of the black hole mass range that could make up dark matter, because lensing would cause visible brightness changes in background stars that have not been seen, leaving only a tiny corner of parameter space, which the speaker finds desperate rather than plausible.

    “So most of the black hole range that could explain dark matter has been ruled out by microlensing.”

  • A detected merger of objects below one solar mass would be extraordinary, and the least radical explanation would be two primordial black holes formed in the early universe from a quantum process.

    “Then I think the least radical explanation would be this should be two black holes that have to have been formed in the beginning of the universe out of some quantum process.”

  • If dark matter is axions, which have a small mass and are bound by the black hole's amplification process, they can condense into clouds around spinning black holes, forming atom-like systems where the black hole is the nucleus and the axion cloud is the electron cloud, with thousands of papers predicting their spectra.

    “there might be systems there that look like atoms: a nucleus, which is a spinning black hole, surrounded by a cloud, like the hydrogen atom, a cloud of axions.”

  • Very light black holes that could otherwise constitute dark matter would evaporate under Hawking radiation before today, so they cannot be present now, while heavy ones would be detected by microlensing, so the remaining possibility is narrowly constrained.

    “If the black holes were just too small, so very light black holes, they would evaporate under Hawking radiation.”

  • Black holes much less massive than the Sun would have to form from quantum fluctuations in the early universe, and claimed detections of such objects have a signal-to-noise ratio too low to be convincing, but continued observation should make real evidence grow over time if they exist.

    “There's claims flying around that maybe we have seen a couple of them, but the signal-to-noise ratio is really just too small”

0.23

Black hole imaging and waveform analysis depend heavily on simulations and theoretical templates.

3 pointscentrality 2/5
  • Computer simulations produce templates only for the inputs they are given, so analytic theorists are needed to point out missing physics such as plasma and dark matter, and theory and simulation must work together.

    “But you only get what you feed the machine. Then you need other people doing their job, which is to say, 'Look, that's all fine, but I'm afraid you're only doing vacuum black holes.'”

  • The real Event Horizon Telescope images are very blurry and roughly 90 percent of each image is reconstructed from thousands of simulations of matter around black holes, so prior theoretical knowledge strongly shapes the pictures, and future instruments should see without that dependence.

    “they're reconstructed from, I would say, roughly 10%. So 90% of the image is reconstructed”

  • Adding realistic astrophysics such as accretion disks and dark matter halos makes template banks infeasible, so searches must become hierarchical: a template search flags an event, then the event is rescanned for environmental effects.

    “When you start adding dirty astrophysics, 'Now I need an accretion disk. Now I need a dark matter halo,' suddenly building banks of templates for this, it's not feasible.”

0.12

Theorists pursue black hole physics for its intellectual challenge and public appeal.

4 pointscentrality 1/5
  • For observers a black hole is a point-like, very massive, dark object that controls nearby gravitational interaction, whereas for theorists it is an object curving spacetime so that time stops at an event horizon, making it a special geometry whose small-scale structure is the object of study.

    “for somebody doing observations, a black hole is a point-like object which is very massive and dark, and that's the end of it.”

  • Cardoso says it is mysterious that calculations done with pen, paper, and supercomputers about black holes describe objects millions of light years away, suggesting the universe is written in a mathematical language that does not deviate from its rules.

    “the universe is indeed written in some mathematical language and it doesn't go astray from those rules”

  • Cardoso takes the probability of axions existing as effectively zero because none have been observed, but he and others work on axion-black hole physics because there is a nonzero chance they exist in the relevant mass range and because solving the equations is a compelling challenge.

    “So I work, I think we most do, we work because of the theoretical challenge.”

  • Black holes capture the public imagination because they hint at incomplete physics, since they are a key ingredient in the story of unifying quantum theory and gravity, and outreach events like Black Hole Week reach tens of thousands of people and draw volunteers.

    “Black holes still capture the imagination, I think because there's things we don't know. Physics has to be completed, and we know black holes are a key ingredient in the story.”