Holiday Message: Reflections on Immortality
What this covers
Sean Carroll reflects on whether immortality is possible in principle, drawing on physics, philosophy, and literary thought to argue that the laws of cosmology rule it out entirely. The episode, a year-end holiday message, opens the question of whether living forever is desirable, then moves through the thermodynamic constraints that make it impossible. Carroll anchors the discussion in the idea that Life as Non-Equilibrium Process — all thinking, metabolism, and action — consumes the universe's Low Entropy as Finite Resource, and that this budget is capped by the observable universe's maximum entropy of roughly 10^122. He calculates that this implies an upper limit of about 10^93 human life equivalents could ever exist, vastly more than have lived so far but strictly finite.
The episode ranges across distinct territory. Carroll examines Freeman Dyson's 1979 proposal for indefinite life through slowing thought in an ever-cooling cosmos, then shows why the cosmological constant invalidates it by establishing a finite De Sitter space endpoint instead. He addresses Quantum Immortality and the persistence of microscopic information through death, arguing that neither constitutes meaningful survival since Macroscopic Emergent Identity — personal continuity — becomes inaccessible once the organism dies. The discussion also brings in Jorge Luis Borges and Julian Barnes on the psychological and social costs of endless existence, and considers Limit Concepts as a philosophical tool for understanding mortality. A closing thought introduces the Ceremony of Forgetting as a way to honor finitude, alongside a note on the Carroll-Chen scenario — a speculative cosmology in which life continues to emerge across infinite future baby universes.
Carroll argues that while longevity and even cosmically-vast lifespans may be physically and biologically possible, true immortality is impossible because life is a non-equilibrium process that necessarily increases entropy, and the universe has a finite entropy budget and tends toward a dead thermal-equilibrium state.
- Life and thinking are non-equilibrium processes that consume the universe's low-entropy resource
- The observable universe has a finite upper entropy bound (~10^122), capping total possible life
- Quantum immortality and persistent microscopic information do not constitute meaningful immortality
The universe's finite entropy budget fundamentally limits total life and thought across all future time.
- No amount of cleverness or resource-hoarding can let a living being last beyond the entropy bound; as long as you are still existing as a human being having thoughts you keep increasing entropy, so the only way to avoid hitting the upper bound is to stop existing entirely.
“you can't do it unless you die, unless you literally don't exist. As long as you are still existing, like a human being still having thoughts, still increasing the entropy of the universe, there is an upper bound to doing that”
- Living organisms are quasi-homeostatic non-equilibrium systems embedded in an entropy gradient; life on Earth is possible only because the sun is a hot spot in a cold sky, reflecting that we are far from thermal equilibrium and must consume free (low-entropy) energy to metabolize, act, and think.
“Life on Earth, in particular, is possible because the sun is a hot spot in an otherwise cold sky.”
- Roughly every bit of information processed increases the entropy of the universe by one unit, and although reversible computation is a caveat, the kinds of computation relevant to thinking, remembering, and predicting do generate entropy in the surrounding universe.
“Roughly speaking, every bit of information that is processed increases the entropy of the universe by one... the kind of computations that are relevant to thinking, remembering, predicting the future, stuff like that. Do generate entropy.”
- The initial low entropy of the universe is a finite resource that life uses up: by metabolizing, acting, and thinking we consume free energy and increase the universe's entropy toward its final empty equilibrium, so there is only a limited supply available for life.
“The initial low entropy of the universe is a resource that we are using up. We are increasing the entropy of the universe, approaching that final empty thermal equilibrium state, and we have a finite amount of that resource to use.”
- Because a universe with a cosmological constant approaches De Sitter space with a finite maximum entropy for the observable universe (about 10^122), and because thinking and living necessarily generate entropy (roughly one unit per bit processed), there is an upper limit on how much life can ever exist in the future, refuting Dyson's energy-based argument which ignored the entropy constraint.
“if thinking, and therefore if living and thinking, increase the entropy of the universe, and there's an upper limit to the entropy of the universe, then guess what, there's an upper limit on how much life there can be in the future history of our universe.”
- Scaling from ~10^29 bits per human life against the universe's ~10^122 entropy bound implies an upper limit of about 10^93 human life equivalents that could ever be experienced; vastly larger than the ~10^11 lived so far or the ~10^34 possible in the Milky Way, but still finite, which is the final lesson that true immortality is impossible.
“that's 10 to the 93 human life equivalent that we can fit into the universe. That's a very, very large number, 10 to the 93... it is finite. That's the final lesson here.”
Cosmological fate shows complex structures dissolve and the universe approaches empty equilibrium eventually.
- We are past the era of peak star formation: star formation in the observable universe peaked around 4-5 billion years after the Big Bang and rates have been declining for almost 10 billion years, indicating that this particular kind of cosmic complexity is no longer increasing.
“we are past the era of peak star formation in our universe. Star formation peaked around 4 or 5 billion years after the Big Bang, and now we're 14 billion years after the Big Bang. Star formation rates have been declining for almost 10 billion years.”
- The Carroll-Chen scenario proposes that most of spacetime is empty De Sitter equilibrium, but rare quantum fluctuations spawn baby universes with their own low-entropy initial conditions and arrows of time; given infinite time, infinitely many such universes form, so the universe as a whole lasts forever and life keeps re-emerging, even though no information passes between universes.
“if we add to that, the possibility that there can be a quantum fluctuation that creates a little baby universe in some local region of space... You have infinity years to wait. So eventually you'll get lots of baby universes”
- The arrow of time arises because the early universe had very low entropy (a hot, dense, smooth, rapidly expanding state corresponding to very few micro-states), and the universe continually approaches thermal equilibrium as it expands, cools, forms black holes, and ultimately empties out.
“the arrow of time... comes about because our universe used to have very low entropy, entropy thought of as a way of counting how many micro-states look a certain way in the macroscopic configuration of the universe.”
- Carroll and Aidan Chatwin-Davies proved a version of the Cosmic No-Hair theorem without using Einstein's equations, by instead using a generalized entropy definition (from Bousso and Engelhardt) covering both matter and gravitational entropy, showing that if entropy rises to a constant finite value the universe approaches empty De Sitter space.
“We wrote a very fun paper that said that you can actually prove what is called the Cosmic No-Hair theorem... what Aidan and I did is not use general relativity... we used a definition of entropy that is applicable to a situation where you have both gravity... but also, you have matter”
- All the churning complexity of the present universe will eventually disappear: the last stars will burn out around 10^15 years, black holes will evaporate, and the accelerating expansion driven by vacuum energy will stretch the remaining photons into effective non-existence, leaving empty De Sitter space as the high-entropy final state.
“The last star will probably peter out around 10 to the 15 years from now... eventually, those stars are gonna die, they'll fall into black holes, the black holes will evaporate, and the evaporation products will scatter across the universe.”
- In the baby-universe scenario there will always be future beings in the universe's future experiencing thought, love, caring, meaningfulness and purpose, at least temporarily, in regions of spacetime that will exist forward in time — a hopeful conclusion despite no information passing between universes.
“there will always be future beings in our future... there will be thought, there will be love, there will be caring and meaningfulness and purpose, at least temporarily, in regions of space and time that will exist in the future of our universe, that's the good news.”
Human desires and imagination are finite while eternity is infinite, causing eventual boredom and loss of meaning.
- Carroll personally would choose to live 10,000 years (given sustained health) because of the many things he wants to do, while agreeing he would not want to live literally forever, drawing a sharp distinction between large longevity and true infinity.
“I don't wanna live forever, but I personally do want to live for 10,000 years.”
- Because human desires, interests, and imagination are finite while time and possibility may be infinite, an immortal person would eventually run out of things to be interested in and lose motivation, as illustrated by Borges's troglodytes and Julian Barnes's heaven where everyone eventually chooses to stop existing.
“time is infinite, possibility is infinite, but our imaginations, or maybe to be more fair, our desires or our interests are finite. We human beings are finite.”
- Ted Chiang argued that immortality would be bad communally rather than individually: society would stagnate and fall into a rut because successful people keep doing what worked, and there is an evolutionary (social as well as biological) reason people must die so that new ideas from younger people can emerge, since effective immortality precludes having new young people.
“he was saying that society would stagnate. We would fall into a rut. There is something to be said. There is an evolutionary reason why, and not just biological evolution, but social evolution, why people need to die.”
Aging is solvable as a biology problem since physics permits indefinite organism repair.
- Radically extending human lifespans is realistic because nothing in the laws of physics forbids repairing a biological organism indefinitely; aging is a biology problem, not a physics problem, and over long enough timescales lifespans of tens of thousands or millions of years may be achievable.
“this is a biology problem, not a physics problem. There's no problem, as far as the laws of physics are concerned, with repairing yourself as a biological organism.”
Mortality can be honored through acceptance and ceremony rather than through denial or escape.
- Youthful energy is socially important because young people have the drive and desire to change the world, and even though that sometimes leads them to extremes, society needs to encourage that energy and allow them to make mistakes rather than suppress it.
“we still need to let them make those mistakes. We need to encourage that youthful energy and that willingness to change the world.”
- Pushing yourself to the limit of what you are comfortable contemplating — a 'limit concept' like immortality — is valuable even if the concept is unrealistic, because doing so teaches you about yourself and clarifies your more realistic concerns in the world.
“Push yourself to the limit of what you're comfortable thinking about or contemplating, and look at what you learn. Look at what you bring back.”
- Rather than struggling against the inevitable passing of ourselves and our memories, we can honor mortality through a 'ceremony of forgetting' — deliberately letting go and moving on — because life continues for those still around, an idea Carroll finds touching and aligned with taking mortality seriously.
“If you take mortality seriously, if you don't struggle against the inevitable passing of ourselves and our memories, then rather than doing that, let's honor it.”
- The legitimate reason to be upset about death is not that you will suffer once dead (you won't care then), but that the prospect of dying in the future is upsetting now; valuing continued life is a present-tense preference even in a single classical universe.
“The reason to be upset about the prospect of death, which is a perfectly legitimate reaction to have, is not because once you're dead you will be sad. That was never the idea. The prospect of dying in the future is upsetting to me now.”
- There are two kinds of death: physical death when you stop breathing, and the social death the last time anyone thinks of you or says your name; the finitude of life means there is a last book you read, last meal, last person you love, and eventually a last time anyone remembers you.
“everyone dies twice; when they stop breathing, and the last time anyone ever thinks about them, or says their name.”
- A human increases surrounding entropy by about 10^23 bits per second (mostly infrared radiation), versus roughly 10^20 bits per second of brain computation; over a century this yields about 10^29 bits per human-life-equivalent, with ~10^11 human life equivalents lived to date.
“human beings increase the entropy of the universe around them by about 10 to the 23 bits per second... The number of computations in a human brain is roughly 10 to the 20 bits per second... let's get 10 to the 29 as the number of bits of entropy generated in a typical human lifetime.”
Personal identity persists through time as emergent macroscopic continuity despite constant change.
- Defining a human lifespan only makes sense at the macroscopic emergent level where a person retains some but not all information (psychological continuity that is neither absolute nor absent); we change over time but do not change utterly, keeping enough memory and common characteristics to remain identifiable.
“as long as we have this macroscopic perspective, and we admit that even though we change over time, we don't change utterly over time, we keep some information, some memory, some common characteristics of who we are from moment to moment, I think it is perfectly well defined to say that people have a lifespan”
- Although modern physics holds that the microscopic information contained in a person persists through time (as in Laplace's demon or the black hole information paradox), this does not constitute immortality, because personal identity is a macroscopic, emergent phenomenon and the accessible information that defines you becomes inaccessible at death.
“modern physics thinks that the information contained in you persists through time. This doesn't count as immortality. That is my first message.”
Quantum immortality and Dyson's infinite-thought scenarios fail because they ignore entropy constraints or quantum weighting.
- Even granting the many-worlds interpretation, the branches of the wave function on which you live forever count infinitesimally little (per the Born rule weighting by amplitude squared), and those surviving versions are merely your descendants rather than you, so quantum immortality provides no meaningful comfort against death.
“these branches on which you live forever count enormously tiny amounts, like infinitesimally, almost immeasurably tiny amounts. It's not something that should affect your everyday life in any way”
- Freeman Dyson's 1979 argument held that in an open, forever-expanding universe one could live forever by having thoughts that consume less and less energy but take more and more time, allowing an infinite number of thoughts to be fit into an infinitely old, ever-cooling universe.
“you could live forever in an open universe, because if you think that it takes a certain amount of energy to have a thought, well, what you can do is just have thoughts that take less and less energy, but take more and more time.”