210 | Randall Munroe on Imagining What If...?
What this covers
Randall Munroe, the physicist-turned-cartoonist behind xkcd, sits down with Sean Carroll to discuss how he approaches his "What If?" columns and books—a project that answers absurd hypothetical questions by running real physics on impossible premises. The conversation centers on why this kind of inquiry matters: the mathematical tools required to answer ridiculous questions (how fast would a human need to travel to lap a racetrack? what happens when a baseball travels at relativistic speeds?) are identical to those solving practical, important problems. Munroe demonstrates this principle throughout, showing how equations governing choked flow, the rocket equation, and stagnation temperature surface in both contrived scenarios and genuine engineering challenges.
The discussion ranges across physics pedagogy, the nature of explanation, and the act of visualization itself. Munroe explains why showing the process of reaching an answer—not just the answer—is what educates an audience, and how the breakthroughs often come from recognizing that another discipline already solved your problem (consulting a cookbook to determine how a steak cooks while falling through the atmosphere). He explores how spherical cow simplifications in physics are never arbitrary but always chosen to answer a specific question, and reflects on the broader value of curiosity-driven work with no economic incentive—holding up NASA as an institution uniquely tasked with answering questions society decided were worth learning, regardless of profit. Along the way, the conversation touches on survivorship bias in advice-giving, bounding estimation as a tool for inference, and the surprising physics of things like white dwarfs and geysers.
Munroe argues that taking absurd hypothetical questions seriously—pressing 'play' on impossible premises and letting real physics run—is both a powerful pedagogical tool and a genuine driver of curiosity, because the simplifications and equations used to answer ridiculous questions are the same ones that solve practical, important problems.
- The same mathematical tools (choked flow, rocket equation, stagnation temperatures) apply to both absurd and practical problems
- Showing the process of getting to an answer, not just the answer, is what satisfies and educates an audience
- Curiosity-driven inquiry without economic incentive is valuable in itself, as exemplified by NASA
Absurd hypothetical questions use the same mathematics as practical ones and teach genuine scientific tools.
- A baseball pitched at 90% the speed of light would crash into air molecules in front of it, undergo nuclear fusion, and disintegrate into an expanding cloud of radiation that destroys the playing field, releasing significant energy from both kinetic energy and fusion.
“the ball would start crashing into the air in front of it, and might undergo nuclear fusion... you end up with this sort of millisecond or nanosecond by nanosecond picture of the ball disintegrating into an expanding cloud of radiation and destroying the playing field.”
- A ridiculous What If question uses the same mathematics as practical, important questions—the geyser-launch scenario relies on choked flow equations ubiquitous in fluid dynamics and the rocket equation for momentum transfer—which is why answering absurd questions teaches genuinely useful tools.
“one of these questions might be completely ridiculous, but the math that you're using is the same as math that you'd use to answer a very practical and important question... it involves these really cool choked flow equations that come up almost everywhere when you do any kind of fluid dynamic problems... you can just use the rocket equation to figure out what momentum's being transferred”
- Old Faithful erupts superheated, high-pressure water at roughly 50-100 meters per second; though low in density (like a fluffy pillow), the momentum is like being hit by a pillow at highway speed, so a direct hit or standing over it with an umbrella could fling a person high into the air, with the fall likely more dangerous than the burns.
“It's coming out, going on the order of 50-100 meters per second, I think. The amount of momentum carried by the stream, it has pretty low density. It's like one of those cotton candy type fluffy pillow, but it's like getting hit by a pillow going at the speed of a car on the highway.”
- Because a near-light-speed baseball's radiation cloud approaches faster than nerve signals travel down the optic nerve, a batter would literally not see what hit them—the cloud disintegrates and reaches them before visual signals could be processed.
“the batter doesn't have any sign that it's coming until the first wave of radiation hits, and once it does, the speed at which signals move down the optic nerve is slower than the rate at which the actual cloud... of the ball approaches and starts disintegrating, so you would literally not see what hit you.”
Good explanations show their work, bridge domains when stuck, and assume listeners are interested but busy.
- When struggling to answer a problem within one domain's framework (e.g. heat propagation through flesh via physics), the breakthrough can come from recognizing a different domain of expertise already knows the answer—Munroe closed his physics tabs and opened a cookbook to answer how a steak falling through the atmosphere would cook.
“I had a moment where I sat back and I was like, wait a minute, and then I closed all the physics tabs and just opened a cookbook... there's a different regime, a different domain of expertise where they know a lot about what happens if you apply different amounts of heat to a steak”
- An explanation that doesn't show some of how you got there and why it's satisfactory just raises more questions—answering 'why is the speed of light what it is' with the permittivity and permeability of free space leaves two new unexplained terms, so you must show why a step was helpful for the answer to satisfy.
“an explanation that doesn't show some amount of how you got there and why you think that's satisfactory, just raises more questions... it's because it's this over the permittivity of free space times the permeability of free space. You're like, 'Okay, well, wait, now I have two new questions.'”
- The way to explain a complex answer well, without condescending, is to imagine going back in time to tell yourself the CliffNotes summary that would have saved the hours of research—identifying the minimum needed to understand the answer plus the cool things found along the way—while assuming the listener is interested but busy.
“if I went back in time, and I was gonna tell myself, 'Hey, I figured it out, I'm gonna save you all that time,' what would my CliffNotes summary be? What would my, 'Okay, here's the stuff that you need to understand to understand the answer. Here's the answer. Here's cool stuff that I found on the way that you're not gonna wanna miss.'”
Communicating well requires humility: you cannot control how people feel, only adjust your words if misunderstood.
- Munroe doesn't deliberately balance jokes against content; he goes with what's cool to him, because over-analyzing humor is like thinking too hard about tying your shoes—you stop being able to do it. The goal is simply to make it worth the reader's time, with either a good punchline or something so interesting they don't need one.
“I think that you can only go so far analyzing this stuff before you kind of... If you think about tying your shoes too hard, you stop being able to do it.”
- Communicating with people is hard, and the only measure of doing it right is whether others get the message you intended; this requires humility—accepting you can't control how people feel, but if you're not getting the response you wanted, you should try different words rather than blame the audience, otherwise you grow more angry and lonely.
“fundamentally the only measure of whether you're doing it right, is if other people are getting the message that you want them to... if I use this word and people are understanding me wrong... I can just be angry all the time or I can be like, 'Okay, what words would get this across?'”
- Debates over whether something counts as art, a comic, or whether the creator is an author are empty semantics that don't matter; what matters is whether it's a cool way to get a point across using these tools, so one should let the thing be what it is rather than worry about labels.
“People are like, oh, is this art or not? Are these comics or not? Are you an artist?... it's all semantics.”
Physics simplifications are always purposeful choices guided by what question you're trying to answer.
- Simplifying away complications in physics—like the 'spherical cow in a vacuum'—is never arbitrary but always done in service of answering a specific question; the simplifications you choose to make are guided by what you're trying to find out, so all physics is choosing which complications to focus on.
“the simplifications you're making are guided by that question you're trying to answer... It's all complications. Just choosing which ones you're gonna focus on.”
- Because Google won't disclose its data center size, you can estimate the order of magnitude by setting bounds from multiple directions—they can't occupy a majority of Earth's surface, can't buy all the hard drives Earth manufactures, and their power consumption from public municipal records constrains how many servers and drives they can run—until the bounds converge on a probable range.
“how many hard drives does earth manufacture? They can't be buying all of them, so like, that puts a bound on it... how much power are they using? In some places, they have to contract with municipalities and have public records of what power they're consuming... you just kind of keep coming at it from different directions until you find kind of something that hems in the answer”
Visualization and perspective drawing surface non-obvious consequences that pure calculation misses.
- A room temperature white dwarf would not emit light, so a spaceship visiting it would need headlights to see the star—a counterintuitive realization that only emerged when Munroe tried to draw the scene, illustrating how the act of visualization surfaces non-obvious consequences.
“now I've gotta draw the star, and normally when you draw a star, you're like, well, this is bright. But then I realize that it's not. It's not emitting light. So the spaceship will need headlights.”
- Anyone who does physics should take a class in perspective drawing, because so much physics instruction depends on sketching three-dimensional scenarios on a board—planes, particles, forces, surfaces—and being able to figure out roughly where the lines go is genuinely useful for communicating science.
“I feel like anyone who does physics should have to take a class in perspective drawing just for the part on where you're drawing on the board... It's really helpful to be able to take a picture and just figure out like, 'Where do these lines roughly go?'”
NASA uniquely serves society's pure curiosity-driven inquiry with no economic incentive or customer, which is valuable.
- NASA is unique among large technical institutions because it is tasked with doing things for no reason except that society collectively decided it's worth learning—solving problems like whether there was life on Mars where there is no economic incentive or paying customer—acting as society's representatives in pure curiosity-driven inquiry, which is a good thing to do.
“they are tasked with doing things for no reason except that we think it's a good idea, like collectively as a society to do this to learn more... they are working on these problems because they want to know and they are acting as our representatives... solve problems that there's not necessarily an economic incentive to solve.”
- Tires get smaller as they wear, but roads don't thicken with built-up rubber because the lost rubber goes everywhere into the environment—which may be a huge pollution problem that no one currently knows how to solve.
“why aren't our roads getting thicker as a layer of rubber builds up? Where is all that rubber going? And that, it turns out it's going everywhere, and it may be a huge problem and no one knows how to solve it.”
- The Sun will eventually collapse into a white dwarf that cools thermally over billions of years; all current white dwarfs are still hotter than the Sun's surface (tens of thousands of degrees) because the universe isn't old enough for them to have cooled, but room temperature stars will begin appearing somewhat more than the current age of the universe from now.
“the white dwarf stars in our universe are cooling down over time, all of them are still hotter than the surface of the sun is now. They're all in the tens of thousands of degrees, I think, because the universe is not old enough for them to have cooled... it's like in that order of magnitude is when we'll start getting room temperature stars.”
- Munroe indexes all the comics he's ever done to avoid telling the same joke twice, comparing it to needing a table of which friends you've told a particular story to.
“Definitely. I mean, but I have that problem in real life. Like, I'll tell a story, I'll have a really good little story... it's like you have to keep a table of which of your friends you've told the cool microwave story to.”
- Asking a successful person for advice is sampling from a highly selected pool, much like only interviewing lottery winners who say 'buy a lot of tickets'—so a successful person knows what worked for them but cannot know how likely it was to work, making advice as a genre unreliable.
“if you get someone who's successful at something... It's like, you're really sampling from a pool of people that's selected here... if you only interview lottery winners and you're like, 'What's the secret?' And they're just like, 'You gotta buy a lot of tickets.'”
- The upper limit on skyscraper height is not engineering but economics—structures can be built much higher into the atmosphere than has been done, but cost rises rapidly with little economic incentive, so the tallest skyscrapers are built largely for bragging rights.
“The upper limit on skyscrapers isn't really engineering, it's more money. You can build a skyscraper up into the... Much further up into the atmosphere than we've successfully built them. It just gets rapidly more expensive, and there's just not a lot of economic incentive to do that. Most of the really tall skyscrapers are build for bragging rights as much as anything.”
- The greatest danger near Yellowstone geysers is not the eruption itself but the surrounding boiling mineral pools covered by a thin mineral crust, where a person walking on what looks like rock can plunge through into boiling water—which is why the Park Service railings exist.
“there are all these boiling mineral pools that are just barely on the edge of boiling water with a mineral crust over the surface. And so, if you are walking around there, you could step on what looks like rock and just plunge through into boiling water.”
- Finding Wikipedia article titles that can be sung to the Teenage Mutant Ninja Turtles theme isn't just about syllable count—the words must have the right patterns of stress, which is genuinely tricky, requiring a natural language processing library to extract syllable and stress patterns and filter candidates.
“it's not just any word with the right number of... Syllables, any eight syllable... You have to have the right patterns of stress and that's actually tricky. So I got a language processing library, figured out how to use it, figured out how to get... Extract syllable patterns from words and stress patterns”
- Approaching a room temperature white dwarf, you would reach non-survivable speeds measured in fractions of the speed of light as you plunge toward it; coming within the distance of the Sun's current surface is the point of no return—not from black-hole physics but because the acceleration required to escape would flatten you to the floor of your spaceship.
“If you approach the star to within the distance of the sun's current surface, that's around the point at which you're now too late to turn around and escape not because of black hole style... but that's the point at which the acceleration it's gonna take is gonna flatten you to the floor of your spaceship.”
- A billion-storey building would extend far enough out that Earth's rotation flings its upper portions outward by centrifugal force harder than gravity pulls down, turning the structural problem from compression to keeping it from being flung apart—creating a space elevator situation where the end could swat against the moon, with catastrophic energy released when fragments fall back through the atmosphere.
“you'd end up with a skyscraper that is far enough out that as the earth is rotating, it's being flung outward by centrifugal force harder than gravity is pulling it down. So you end up with a sort of space elevator situation where suddenly your problem is not making it strong against compression, it's keeping it from being flung apart.”
Scientists are eager to discuss their work with genuine questioners regardless of how silly the question seems.
- Reaching out to scientists is far more rewarding than expected because, even if unfamiliar with your work, people are genuinely excited to talk about what they work on—so questions that seem too silly to bother an expert often yield enthusiastic, helpful answers.
“reaching out to scientists, is like, they're really nice. And even if they aren't familiar with my books or comics or whatever, it's like people are just excited to talk about the things they work on.”
- Little kids ask some of the best What If questions because they aren't trying to come up with something weird or deliberately destructive—they're simply asking concrete, straightforward questions they're genuinely curious about, which often yield richer answers than contrived scenarios.
“this is where I found that little kids asked some of the best questions. Because little kids aren't trying to come up with something weird, they're just asking an actual question that they're curious about.”