Key Assumptions of the Transcension Hypothesis: Do Advanced Civilizations Leave Our Universe?

Low-mass X-ray binary (LMXRB) star system. Strange as it seems, Earth’s future may look something like this, with us inside a black hole-like environment of our creation, on a highly accelerated path to merging with other universal civilizations doing the same. If true, our destiny is density, and dematerialization.

This post is a followup to a popular paper of mine on three big topics: the Fermi paradox, accelerating change, and astrosociology (the nature and goals of advanced civilizations). The paper is called The transcension hypothesis: sufficiently advanced civilizations may invariably leave our universe, and implications for METI and SETI. It was published in Acta Astronautica in 2012.

Speculation on the Fermi paradox has grown considerably in the last two decades, as it has become increasingly obvious that we live in a universe that is very likely to be teeming with Earth-like planets, and also with intelligent, curious, and technologically accelerating forms of life. When we extrapolate our own accelerating progress in science, IT, and nantechnologies, we can imagine that any one of these civilizations could easily send out self-replicating nanotech that would spread across our Milky Way galaxy and beam the information that it finds out to the rest of the universe (or alternatively, just back to the originating civilization), creating a Galactic Internet, and making our universe as information-transparent as our planet is becoming today. Our galaxy has a radius of 100,000 light years. Replicating nanotech, traveling at just 5% of light speed, which we can imagine building even today, could reach all corners of our galaxy in 2 million years. So if other Earth-like planets and their intelligent life likely emerged, closer to the center of our galaxy, at least one billion years before ours did, as several astrobiologists have estimated, and any one of them could have easily expanded, why don’t we see any signs of this Galactic Internet today? Or signs of past alien visitation, probes, and megastructure beacons near Earth? Or signs of intelligent structures or civilizations anywhere in the night sky? In other words, Where is Everybody? That’s the Fermi paradox.

In a nutshell, the transcension hypothesis predicts constrained transcension of intelligence from the universe, rather than expansion (colonization) within the universe by intelligence, wherever it arises. If the hypothesis is correct, the reason we don’t see and haven’t heard from advanced civilizations anywhere is that the vast majority leave the visible universe as they develop, and the few that do not are very unlikely to be visible to us, with our presently weak SETI abilities. That’s a very strong claim. Could it be right?

My paper makes a series of assumptions about the nature and future of intelligent life in our universe. Most of these key assumptions may need to to be correct, in some fashion, for the hypothesis itself to be correct. A few colleagues have asked me to summarize these assumptions in one place, so here they are. This list is a good way to get a quick summary of the hypothesis as well.

Here are the key assumptions of the transcension hypothesis, as I presently see them:

  1. Intelligent life, on Earth and elsewhere in our universe, is not only evolving (diversifying, experimenting), but also developing (converging toward a particular set of future destinations, in form and function), in a manner in some ways similar to biological development. In other words, all civilizations in our universe are “evo-devo” both evolutionary and developmental. The phenomenon of convergent evolution tells us a lot about the way development may work on planetary scales. A kind of cosmic convergent evolution (universal development) must also exist at universal scales.
  2. The leading edge of intelligence always migrates its brains and bodies into increasingly dense, productive, miniaturized, accelerated, and efficient scales of Space, Time, Energy, and Matter (what I call STEM compression), because this is the best strategy to become the niche-dominant local intelligence (and for modern humans, Earth’s biosphere is one precious and indivisible niche), and because the special physics of our universe allows this continual migration into “nanospace“. Human brains with their thoughts, emotions, morality, and self- and social-consciousness, are the most STEM-compressed higher computational systems on Earth at present. But our biological brains are just now starting to get beat at the production of intelligence by deep learning computers, which are even more profoundly STEM-compressed in certain kinds of computation than neurons (for example, electrical interneuron communication in an artificial neural network is seven million times faster than chemical action potentials between biological neurons). Once today’s weakly bio-inspired machine intelligence becomes fully self-improving, it seem likely to continue growing and improving at rates that make biological intelligence appear rooted in spacetime by comparison, in the same way that Earth’s plant life appears rooted in spacetime by comparison to self-aware animal life. Fortunately, accelerating STEM-compression of both human civilization and of our leading computational technologies is stepwise measurable and testable, as argued in my paper. Our academics need better funding and training to do so, however. Measuring the STEM-efficiency growth of new computational platforms, like quantum computing, is today far more art than science.
  3. The acceleration of STEM compression must eventually stop, at structures analogous to black holes, which in current theories appear to be the most computationally accelerated and computationally efficient entities in the known  universe, an insight Seth Lloyd made in 2000 which remains widely underappreciated by most information, computation, and complexity theorists today. Fortunately, this idea of a developmental “black hole destiny” for civilization seems quite testable observationally via search for extraterrestrial intelligence (SETI), as argued in my paper.
  4. A civilization whose intelligence structures are compressed to scales far below the nanoscale may well be capable of creating or entering black-hole-like environments without their informational nature being destroyed. There are 25 orders of magnitude in size between atoms and the Planck scale. This is almost as large a size range as the 30 orders of magnitude presently inhabited by life on Earth. We simply don’t know yet whether intelligence can exist at those small scales. My bet is that it can, and that STEM compression drives leading universal intelligence there, as the fastest way to generate further intelligence, with the least need for local resources.
  5. Due to general relativity, extreme gravitational time dilation occurs very near the surface (event horizon) of black holes. Thus black holes, wherever they exist, can act as forward time travel devices, for any highly STEM compressed civilization that can arbitrarily closely approach their surface without destroying itself. Black-hole-like conditions are thus gateways to instantaneous meeting and merger with other unique civilizations in our universe within any gravity well. Our gravity well includes the Milky Way and Andromeda galaxies, each of which is destined to merge all its black holes, and each of which may contain millions of intelligent civilizations. The rest of our universe is accelerating away from us, due to dark energy. Perhaps the vast majority of black holes in these galaxies (billions?) are unintelligent collapsed stars. But if the transcension hypothesis holds, some smaller number (millions?) may also be a product of intelligent civilizations.
  6. As local acceleration of STEM compression stops, the more black-hole-like we become, local learning will saturate. Local intelligence will be running as fast as it can in this universe, yet it will be both resource and speed constrained. It’s local conditions, in other words, will be increasingly boring and predictable. It will be, from its own reference frame, at “The End of Science”, the end of what it can easily learn and know, to use the title of the elegant and profound book The End of Science (1996/2015) by science writer John Horgan. In those interesting conditions, it may irresistable to slow down local time via black hole entry, and thus simultaneously accelerate nonlocal time, making meeting and merger with other civilizations near-instantaneous. In “normal” universal time, galactic black holes are predicted to merge some tens to hundreds of billions of years from now, as our universe dies. But from each black hole’s reference frame,  this merger is near-instantaneous We can think of black holes as shortcuts through spacetime, just like quantum computers are shortcuts through spacetime. Indeed, quantum physics and black holes (relativity) must eventually be both evo (chaotically) and devo (causally) connected, both physically and informationally, in any future theory of quantum gravity. If some type of hyperspace, extradimensionality, or wormhole-like physics is possible, there might also be ways of future humanity instantaneously meeting civilizations beyond our two local galaxies. But such exotic physics is not necessary for our local gravity well, and for all other civilizations in their own galactic gravity wells. Standard relativity predicts that if we can survive in black-hole-like densities, and if our galaxies are life and intelligence-fecund, we will meet and merge with potentially millions of civilizations as soon as we approach the surface of any black hole, from our reference frame. In other words, our universe appears to have both “transcension physics” and massive parallelism of intelligence experiments built into its relativistic topology and large scale structure.
  7. If we live in not only a developmental universe, but an evolutionary one, each local universal civilization can never be God-like, but must instead be computationally incomplete, an evolutionary “experiment” with its own own unique discoveries and views on the meaning and purpose of life. Thus each civilization, no matter how advanced, would be expected to have useful computational differences, and be able to learn useful things, from every other civilization. In such a universe, we would greatly value communication, assuming that we could trust the other advanced civilizations that we might communicate with. Computational incompleteness would also make us increasing value simulation over physical experimentation, the more complex intelligence becomes. The better faster, better, and more resource (STEM) efficient models of our physical world get, the more we choose virtual rather than physical experiments to address perennial incompleteness in our intelligence.
  8. If not only intelligence, but also immunity (stability, antifragility) and morality grow in leading intelligences in our universe, in rough proportion to their complexity, in other words, if these three life-critical systems are each not only evolutionary, but also developmental, and thus their emergent form and function is at least partly encoded in the “genes” (initial conditions, laws, and environmental constraints) of the system itself, then we can predict that more advanced intelligences, including our coming deep learning computers, will be not only more intelligent, but also more immune and moral than we are today. This idea is called developmental immunity and developmental morality, and I explore it in my paper, Evo-Devo Universe? (2008). If these developmental processes exist, they tell us something about the nature of postbiological life. Such life is going to be a whole lot more collaboration-oriented, intelligence-oriented, immune, and moral than we are today. Social morality, for its part, pushes complex intelligences toward a more ethical impact on the world and each of its sentiences. Decreasing violence has been a mild trend in human societies in recent centuries, as documented in Pinker’s The Better Angels of Our Nature, 2011. But I expect it to be a much stronger trend in postbiological intelligences. Physicists Stephen Dick and Seth Shostak have stressed the importance of thinking hard about the norms and morality of postbiological culture. It’s a big assumption that surviving human and machine collectives must on average become increasingly intelligent, immune, and moral in proportion to their cognitive complexity, under natural processes of evolutionary selection and development, but this is where all the evidence seems to be leading, in my view.
  9. In a universe with developmental immunity and morality, a moral prime directive must emerge, a directive to keep each local civilization evolving in a way that maximizes its intelligence, uniqueness and adaptiveness prior to transcension. That means one-way messaging (powerful METI beacons), self-replicating probes able to interact with less advanced civilizations, and any other kind of galactic colonization would both be ethically prohibited by postbiological life, due to the great reduction in evolutionary diversity that would occur. Wherever it happened, we would meet informational clones of ourselves after transcension, a most undesirable outcome. In biology, evolution keeps clonality a very rare outcome, due to the diversity and adaptiveness cost that it levies on the progeny. In such an environment, any future biological humans that wanted to continue to colonize the stars would be prevented from doing so, by much more ethical and universe-oriented postbiological intelligences. That is assuming biological organisms even continue to be around after postbiological life emerges. Due to STEM compression, their status as biologicals would likely be vanishing short, once they invent technology capable of colonization. It seems much more likely that biology develops into postbiology, relatively soon (just a few centuries perhaps) after digital computers emerge, everywhere in the universe. This outcome also seems likely to be testable via future information theory and SETI, as I argue in my paper.
  10. Some physicists, most notably Lee Smolin in his hypothesis of cosmological natural selection, propose that black holes may be “seeds” or “replicators” for new universes. That gives us a clue to what we might do after we meet up with other cosmic intelligences. We would likely compare and contrast what we’ve learned, and then seek to make a better and more adaptive universe (or universes) in the next replication. Current physics and computation theory suggest that our universe, though vast, is both finite and computationally incomplete. It may have gained its current amazing levels of internal complexity in the same way life on Earth got its amazing living complexity, via evolutionary and developmental (“evo-devo“) self-organization, through many past replications, in some kind of selection environment, a “multiverse” or “hyperverse.”
  11. If all of this is roughly correct, our future isn’t outer space, it’s “inner space.” Both the inner space of black-hole like domains, and the inner space of increasingly virtual and computational domains. The lure of our continually improving inner space is why 21st century folks spend so much time (too much time!) interacting with our still-dumb mobile devices today. It is why the growth of virtual and augmented reality heralds far more than just better entertainment experiences. Combined with the growth of machine learning, virtual/augmented reality will increasingly become the thinking, imagination, and simulation space for eventual postbiological life. Virtual space is where intelligent machines will figure out what they want to do in physical space, just as our own simulating brains are biology’s virtual reality. And just like humans have have done as our civilization has developed, future machines will do more and more internalization, or thinking in virtual space, and less and less external acting, in physical space, the more intelligent they get. This internalization process has a name. It’s called dematerialization (both economic dematerialization and product and process dematerialization), the substitution of information and computation for physical products, processes, and behaviors. The futurist Buckminster Fuller called this process ephemeralization. But ephemeralization of intelligence is only half the story. It describes dematerialization, not densification. If the transcension hypothesis is true, the developmental destiny of all complex life is both accelerating “densification” (eventually to a black hole-like state) and “dematerialization” (becoming increasingly informational and virtual, over time). See my online book, The Foresight Guide, for more on these planetary megatrends, densification and dematerialization (“D&D”) and how they appear to drive universal accelerating change. 

As Fermi paradox scholar Stephen Webb says at his blog, this is quite a lot of “ifs!” Disproving any of these assumptions would be a good way to start knocking aspects of the transcension hypothesis out of contention. We would learn a lot about ourselves and the universe in the process, so I really hope that each of these gets challenged in coming years, as the hypothesis gets further exposure and critique.

Webb is the author of Where is Everybody?2015, a book that offers seventy-five possible solutions to the Fermi Paradox. Webb did a great job condensing the transcension hypothesis into just three pages in his book. His 2002 edition didn’t include it, as I published my first paper on the hypothesis in mid-2002. At Webb’s blog, he charitably says the transcension hypothesis is “one of the most intriguing” possible solutions that he has seen. He also observes that “Unlike so many “solutions” to the Fermi paradox, this one offers avenues for further research.” It certainly does, which is why I hope it continues to gain scrutiny and critique.

A few scholars are now citing the transcension hypothesis in their academic papers on the Fermi paradox and accelerating change, including Sandberg 2010, Flores Martinez 2014, and Conway Morris 2016. I am hoping that trend continues. The more attention it gets, the more critique it will get.

Perhaps the strangest and hardest-to-believe part of the transcension hypothesis, for many, is the idea of universal development. It is particularly relevant to the first, seventh, eighth, and ninth assumptions above. The most amazing and odds-defying thing I’ve come across in my own study of the natural world so far is the process of biological development. Most people don’t think about both how wonderful and how improbable, on its face, is the process of organismic development.

Think about it. Development is guided by a small handful of genes in our genome. It’s incredible that it works, yet it does, and it made you! Development is a good candidate for the most incredible process in the known universe. In many ways, development is even more surprising than evolution, which I define as the set of biological genes and mechanisms that create unpredictable experiment and variety, as opposed to that small subset of chaos-reducing biological genes and mechanisms that statistically guarantee a hierarchical set of future-specific forms and functions. Standard evolutionary theory requires development as an organismic process, yet it also treats development as subservient to the variety-generating processes in natural selection. That model is only half-correct, in my view.

Fortunately, a growing contingent of evo-devo biologists argue that development’s long-range role in constraining the possibilities of evolutionary change may be equally important to evolution’s long-range impact on development. Both processes seem fundamental to mature theory of adaptation. Ecologists have published good work on the way ecosystem development limits the future of evolutionary processes. For example, think of ecological succession, in which increasing senescence of the ecosystem limits short-term evolutionary variety, while also making the oldest parts of the system increasingly vulnerable to death (and renewal). Think also of niche construction, which tells us how growing intelligence, which we use to fashion comfortable niches, limits the future selection placed upon us by our environment. Scholars of convergent evolution also describe apparently universal processes of morphological and functional development that will constrain evolutionary possibilities on all Earth-like planets. Cosmologists who take fine-tuned universe arguments seriously also talk about both local variety and processes of universal development, though they don’t often use that clarifying phrase, when they describe physical and chemical constraints on the possibilities of evolutionary change. All these are important clues toward a meta-Darwinian, evo-devo universe paradigm of universal change.

In short, if our universe actually replicates, as seems plausible in several cosmology theories, and if it exists in some kind of larger selection environment, as also seems plausible, then not only evolution, but development (“convergent evolution”) must also occur not just in species forms and ecosystems, but for our increasingly intelligent planet, as a developing life-human-machine “Global Superorganism”, and for our entire universe itself, as a replicator in the multiverse. Certain aspects of the future of complex systems must be statistically highly biased to converge on particular destinations, and today’s evolution-centric science still has a lot of growing up still to do in order to see these destinations. It needs to become “evo-devo”, seeing the contributions of both evolution and development to the future of universal complexity.

The paper’s second key assumption, STEM compression is more palatable to most people, in my experience, and may turn out to be the most enduring contribution of the paper, even if the rest of the hypothesis is eventually invalidated. If you’ve heard of nanotechnology, you know that life’s leading edge today, humanity, is doing everything it can to move our complexity and computation down the smallest scales we can. We have been very successful at this shrinking over the last several hundred years, and our ability to miniaturize and control processes at both atomic and subatomic scales is growing exponentially. In fact, human brains themselves are already vastly denser, more efficient, and more miniaturized computational devices than any living thing that has gone before them. But they are positively gargantuan compared to the intelligent computing devices that are coming next.

Fortunately I think each of the key assumptions outlined above are testable, though some are obviously more testable than others in today’s early stages of astrophysical theory, SETI ability, information, complexity, and evo-devo theory, and simulation capacity. If anyone is doing work that might shed light on any of these assumptions, I would love to hear of it.

You can find my paper here: The Transcension Hypothesis, 2012. See also this fun 2 minute YouTube video of the hypothesis, by the inspiring futurist Jason Silva and Kathleen Lakey, which has raised its visibility in recent years.

You can find an overview of the evo-devo (evolutionary and developmental) universe hypothesis in my chapter-length article, Evo-Devo Universe? A Framework for Speculations on Cosmic Culture, 2008.

Comments? Critiques? Feedback is always appreciated, thanks.


  1. While I’ve seen the “go inward” hypothesis before (and it appears in different forms in the fiction of writers like Ken Macleod and Greg Egan, it faces the same problem many answers to Fermi have — they posit something that must always, always happen to every intelligence that ever rises on any planet in any galaxy for the past 14 billion years. Just one exception, one expansionist, human-like culture in a galaxy and soon it’s everywhere.

    So while it is possible that some physics we don’t know directs them all to dive inward, it’s an extraordinary claim and needs extraordinary evidence. Of course for now we have none, other than the silence, which can be explained in far simpler ways.

    The proposal that all the civilizations would wish to wait to meet up with all the others without knowing anything about these civilizations is hard to accept. One would not know if the other civs were non-violent or intellectually compatible and it would be a big risk to wait 20 billion years in an eyeblink just to find out you don’t like the neighbours you waited for. As such I would posit that in fact any civ that did that would only do it if it knew there were going to be cool friends to meet, because that civ had in fact explored the galaxy and found remnants saying, “Hi, we went into the black hole. Here’s our library, you can see we’re cool. When you get bored, come join us.” In other words, instead of being immoral, communication between the civs is a must. Even so, by the way, it takes the risk that some later civ that comes along and joins the club is nasty and more powerful.

    Do planets need to vanish for a civ to take the inward dive? I am not sure we can predict that, though as you say if we find deserted planets that will provide evidence that everybody went somewhere.

    I have always believed that the proposed inner space must be truly infinite, that a group of beings can satisfy all their needs with no scarcity of resources. If there is any scarcity, then an exponentially growing species (like humans) will wish to expand. We don’t know there isn’t such a physics of non-scarce resources on the small scale, but as yet we have no evidence that there is. So I’ve always put this one in the “interesting Fermi speculations” class.

    One thing we now know is this: There is either no intelligent civilization anywhere in our region of the galaxy that is interested in communicating with people at our level of technology, or somebody is blocking such communications to us. Kepler has taught us planets are everywhere, and that even people of our technological level can find them, even search for the products of life like free oxygen on them. Everybody in the region with slightly more tech than ours has known there is life here for a billion years. Anybody with tech slightly ahead of us could be blinking a laser at us that you could see with the naked eye, if they wished to. That has left us with 3 options: They are not there, they do not wish to, or somebody is blocking it. The go-inward hypothesis is an example of them not wishing to — we’re too boring to bother signaling.

    • Thanks for the comment Brad. I don’t posit in the paper that all civilizations must enter inner space. What I say is that if transcension is both a developmental and accelerative process, failures should be very hard to find. If our telescopes are good enough, and the timescales work out right, I argue we should find evidence of a few areas of a few galaxies, and perhaps even a few entire galaxies, where we see expansionist activities. Any developmental process only holds statistically, there are always failures.

      But the timescales may be wrong for us to see whole galaxies that are developmental failures, if they exist, and the signs of intelligence engineering too subtle for us to see the few partial failures where intelligence expanded regionally within our own galaxy before they too transcended. We’d love to communicate with neighboring civilizations, but if they are all spaced too far apart, on average, and the process too accelerative, meeting them via transcension will I think be the overwhelming norm, if indeed that occurs.

      Again, if it’s a developmental process, we can expect all kinds of emergent constraints on our expansionist urges, constraints that we can dimly imagine today. Few would have predicted, looking at early humanity, just how civilized we would become via social and ethical development these last few centuries. Pinker’s Better Angels of Our Nature is great for data on this, and if this process continues, postsingularity societies will be even more ethically restrained than most of us would imagine today.

      I suspect if the transcension hypothesis is right, those who try to send one-way communications or probes would be considered highly unethical, due to the loss of evolutionary diversity that would likely ensue, and somehow ignorant of the way the whole universal system is guiding them toward inner space at an accelerating rate. See my post, for more thoughts on this accelerating “inward dive” we seem to be engaged in. In either case, I think the fraction of spacetime that gets expanded into, rather than transcended out of, by higher intelligence may be quite small in our universe, and perhaps very hard to detect by us at this stage.

      There is yet another assumption here, which requires yet more extraordinary evidence, as you remind us, that as leading edge computing systems get more complex, they get naturally more self-policing and ethical, again on a statistical basis. You may not share this assumption. Fortunately, I think this issue will be settled long before we have the option to transcend or not transcend. I suspect the technological singularity will make a lot of these issues a lot clearer to postsingularity minds. But it’s fun to speculate on them now, and consider early tests of the theory, like the Missing Planets prediction I make in the paper. Again, thanks for your insights.

  2. I remember when these ideas were hesitant conjecture, John. You have nurtured and refined your early insights into an evermore robust and compelling hypothesis. Your work is important, I’ll continue to follow your progress and cheer your success.

    • Thank you John. Adam Ford just notified me that there’s a good discussion of the hypothesis going on these last few weeks at Extropy Chat. This is the oldest transhumanist email chatlist. Some of our readers might want to join it.
      Their archives are a great browseable source of interesting topics.

      I find one of the hardest to accept implicit assumptions of the transcension hypothesis for many people is the idea that intelligence may get massively more self-policing, ethical, and integrated (organism-like, rather than independent network like) as it becomes postbiological, in any civilization. In other words, it might be only an ethical failure that would lead to interstellar expansion or messaging after AI arrives, and such failures could be extremely rare statistically, once we’ve got AI ethics anywhere.

      This idea goes against our “the future always gives us more freedom” instinct, but historically we see great swings between evolutionary diversification (new freedoms) and developmental integration (rise of multicellarity, moral codes, city states, laws, etc.). The dominant trend for our planet’s next phase of development may be a massive new integration (a devo over evo phase), emergence of a “collective mind” if you will, with all the new diversity coming in the sub-minds, the way we each have mindsets that argue inside our own head but which remain part of one highly integrated organism. These future minds may have massive new freedoms in what they can construct and experiment with in “inner space” (very small physical scales, and virtual space), but the price of that may be their loss of freedom in outer and physical space, at the scale of the collective and beyond. The whole planet looks, to outside observers, like one integrated organism at first, and a computronium “crystal”/black-hole-like entity later. All the new freedoms might happen inside, in a space outside observers can’t easily see. And perhaps if they weren’t deeply and rigorously integrated, any one of these minds might possibly, with future exponentially powerful tools, threaten the very existence of the civilization.

      Given the trends in violence we’ve seen over both the last millennium and the last century, I personally think there is already reasonable evidence for this idea that our social collectives must become more integrated and self-policing and moral as a function of their complexity. Of course, others draw entirely different conclusions from the same history. It may be too early to know, but I don’t think it’s to early to start guessing and trying to improve our hypotheses. As I mentioned earlier, I think Steven Pinker’s new book, Better Angels of our Nature, is an excellent summary of all these issues, and very evidence-based. It’s a great place to get a long range perspective on integrating trends in social collectives, and what they may mean for the future if they continue.

  3. I’m a little late to the party but I have a question that has been nagging me.

    If the future of hyper advanced civilizations are black hole like environments and in such an environment we assume that these beings possess higher levels of consciousness or awareness that just happens to coincide with density of the black hole like environment. Does this somehow equate to consciousness or more particularly highly organized consciousness being a primal driving force in the creation of universal structures that enable the development of said consciousness in a positive feedback mechanism? Could gravity = primal consciousness?

    • Let me clarify. IF this inward race is the inevitable destination of sufficiently advanced civilizations and the identifying characteristic is black hole like conditions, then would that not dictate that gravity and intelligent consciousness are one in the same forces? Could complexity be how kaku describes when he talks about ants not realizing the interstate system IS evidence of higher intelligence because they do not possess the complexity necessary to understand it as such? How does your theory diffrentiate natural blackholes from intelligently created blackholes? Is it possible that they are one in the same? If so what does such a speculation hold for the true nature of stars?

      • Hi Steve,

        I share your intuition that gravity may be one of the driving forces, or at least a critical enabler, of complexification within the universe. I talk about that possibility in my 2008 paper.

        Click to access SmartEvoDevoUniv2008.pdf

        Others have proposed that the informational properties of thermodynamics (entropy increase) may instead be the most fundamental driver, an alternative approach also mentioned in my paper. We don’t have good theories in this area yet, though many are publishing and thinking about it.

        To address your other comments, if we live in an evo devo universe, both stars and black holes, if they are both replicating evo devo systems, may each be doing their best to encode and express increasing evolutionary and developmental complexity in each cycle. Stars would be very simple intelligence encoding (simulating) and replicating evo devo systems, but they’d still be information-encoders, in a future theory of astrophysics that includes a more mature information theory, to address your last question. As Smolin himself has speculated, the vast majority of black holes in our universe may be analogous to bacteria in our environment–very simple information encoders and replicators.

        Bacteria, and most of the “unintelligent” black holes in our universe, may each express significantly less complexity than, for example, humans do as they complete their lifecycle. But we may need to *start* from these simpler replicating systems as a base on which far more complex replicating universes may emerge. It is a parsimonious theory, and it proposes a lot of self-similarity between replicating complex adaptive systems both within and outside the universe.

        For another model that proposes self-organizing replicative stability as the central feature of all complex systems, let me recommend Addy Pross, What is Life?: How Chemistry Becomes Biology, 2012.
        His theory of dynamic kinetic stability (DKS) as a replicative (evolutionary and developmental) perspective on the emergence of adaptive complexity fits very with the models described above. We shall see, as they say.

  4. Robbie Shapiro says:

    Hello, i find your theories very interesting, but i have a question, you say that every new physical complex entity is smaller yet mpre complex tan the pne before it, but i cannot see how this is true for, lets say a Galaxy, which even if it is more complex tan a star is far bigger, and another example of my question would be, an animal which is more complex tan a cell, but it is bigger too.
    I hope you have time to answer my questions.

    • Hello Robbie,

      Good thinking! I can see your concerns. Let’s consider each of your questions.

      1. Stars emerged after and within galaxies. And they progressed from Population III to I to I-type stars over billions of years, over a small subset of galactic space and time. This is an example of both increasing hierarchy and locality of complexity emergence.

      2. Single eukaryotic cells did indeed go to multicellular organisms. From the perspective of eukaryotic organisms, this is indeed a small but real decrease in locality. But keep in mind that the first living organism was likely bacteria, and bacteria are best understood as one interconnected global superorganism, like an ant, bee, or termite colony. They share genes between each other all the time (bacterial conjugation). Their various subtypes go miles deep into Earth’s crust, and likely traveled on meteorites, as spores, to neighboring planets (Mars, Venus) via meteorites. Thus the range of the bacterial superorganism is, to date, far greater than any eukaryotic organism or species, including humanity (though a few of our instruments have of course gone farther, but they aren’t autonomous). Locality, the drive toward physical and virtual inner space, really does seem to be the dominant life trend, at least as far as I can tell.

      See my online papers, Evo Devo Universe? and The Transcension Hypothesis if you want to explore these ideas more.

      Thanks for the good questions!

      • Robbie Shapiro says:

        Thank you very much for your answer, but i still have i few questions that i hope you may answer, one is about our future, i now understand that the bacterial superorganism is the biggest life organic form in our solar system, because of that, and going along with the theory of increasing locality, humanity will never live in othe planets, not even the moon, that means that there won’t be any moon base at any time in the future, a i right? if that’s so? what do you think will happen to the projects of the european unión, russia or china, i recently Heard they were trying to créate a moon base.
        On the other hand i now understand that the galaxy came before the stars, but i cannot see how is the galaxy less complex than a star, to me it seems more complex, how is that you measure how complex is any given physical system ? and how do you know which physical system comes after which ? i mean, you say stars are more complex than galaxies, but why stars, why not the solar systems ?
        Thank you for your time and attention.

      • Hi Robbie,

        While I think there is inspirational value in sending another handful of biological humans to the moon again, and a few to Mars, I am also certain that any of the folks planning and thinking about doing either of those at scale are engaged in wishful but deeply flawed thinking. They are not seeing the importance of the real accelerating digital trends here on Earth. A permanent moon or Mars base today, without super advanced AI and robotics, would mostly be a massive waste of resources better spent here today. Most humans aren’t yet willing to recognize that the time for biological beings like us exploring space has nearly ended, and that our real frontier today is transitioning ourselves into postbiology (creating a technological singularity, and becoming technological ourselves). Intelligent machines are the entities that will first go to other planets and (near) space in large numbers, not biohumans.

        Due to the particular physics of our universe, our accelerating transition to postbiology, which involves understanding the key processes and algorithms that life uses and implementing them in our much faster and more durable machines, seems clearly written in the cards. Chapter 2 (Acceleration) of my new online book, The Foresight Guide, goes into this in more detail. I am hopeful that this understanding will continue to grow, as fast as our technology keeps growing. If it does, we’ll soon realize we’re headed for inner space, not outer space, and our science, strategies, and self understanding will all improve as a result.

        To address your other questions, early galaxies condensed gravitationally out of previously even more nonlocal and less complex gases. Increasingly complex populations of stars emerged later, in special, much more local zones within developing galaxies. When stars form, their solar systems form at the same time, along with the star. Complex stars like ours support the formation of iron-rich, rocky core planets like Earth, and life very likely forms inevitably on such planets, perhaps first around undersea hydrothermal vents, almost as soon as these planet get cool enough to support oceans (see Nick Lane’s The Vital Question, 2015 for a great recent set of educated guesses on how that happens). The way we “know” about timing or any these things is through science, which is never perfect, but is self-questioning and always improving. We can be the same way, so keep up your reading and questioning!

        Warm Regards, JS


  1. […] a way to reach us very quickly, before they turned postbiological (they apparently don’t, and postbiologicals apparently have other interests) they would they look roughly like us, as the astrophysicist Frank Drake, author of the Drake […]

  2. […] based in the trending prevalence of electromagnetic frequencies. See additional thoughts by Smart here via blog […]

  3. […] and the potential attraction of universal intelligence to black hole-like environments (the Developmental Singularity or Transcension Hypothesis) have asked me for my take on Christopher Nolan’s latest film. Here it is, along with […]

  4. […] As another example ignored by the show, several evolutionary developmentalists have independently proposed that our very-easy-to-create-yet-general-purpose  “humanoid form”, a bilaterally symmetric bipedal tetrapod with two eyes and two opposable thumbs, is a very likely outcome for all biological intelligences that first achieve our level of sophistication on all Earth-like planets. If we saw such “early” alien intelligences from across a dimly lit room, they would they look roughly like us, as the astrophysicist Frank Drake, author of the Drake equation, has argued. But while we can use science and simulation to argue their existence, in my view we have virtually no chance to meet other biological organisms in the flesh. Why? Because the universe does a very good job of separating all the evolutionary experiments by vast distances, uncrossable by biological beings. Our universe may have self organized to have its current structure in order to maximize the diversity of intelligences created within it, as I argue in my 2008 paper. No intelligence is ever Godlike, so diversity is our best strategy for improving our lot. Another reason we’ll likely never meet biological beings from other Earthlike planets is because the leading edge of life on Earth is now rapidly on the way to becoming postbiological, and postbiologicals are likely to have very different interests than traveling long distances through slow and boring space, when there are much better options apparently available, as I argue in my 2012 paper. […]

Leave a Reply to Brad Templeton Cancel reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: