July 18, 2007

Centauri Dreams: Odds on the human future

Centauri Dreams has posted a review of Princeton astrophysicist Richard Gott's take on the Fermi Paradox and the future of human civilization. Gott uses the Copernican Principle to suggest that humanity likely represents a typical civilization and that humans better get going on colonization efforts before it's too late.

This conclusion is very similar to the one I'll be presenting at TransVision 2007 a week from today in Chicago. Specifically, I will be speaking about the implications of our failing search for extraterrestrial civilizations. While my presentation does touch upon the threat of human extinction, I will also be offering some other non-catastrophic solutions to the Great Silence problem (namely localized digital existence).

As for Gott's argument, I whole heartedly agree that the Copernican Principle can be applied to the Fermi question. We should self-sample ourselves and subsequently not assume that a) we're somehow different than other civs, and b) there are more radically advanced civs in the Galaxy than pre-Singularity civs.

Yes, it's an upsetting conclusion, but that's where the data is pointing.

At TV07 I will also be arguing that the Copernican Principle trumps Occam's Razor in this matter. It's been said (by Kurzweil and others) that the most simplest explanation to the Fermi problem is that we're the first intelligence to emerge in the Galaxy.

To tackle this, I spend the first third of my presentation taking apart the Rare Earth Hypothesis (and other related notions) in an attempt to show that this suggestion is not only grossly improbable, but a bi-product of human arrogance and anthropocentrism.

It's Copernicus all over again.


steven said...

I have to say I find the anthropic selection issues confusing. I simply can't believe, though, that if you ran a million civilizations similar to Earth-2007 into the future, none of them would start a colonization wave. "Localized digital existence" seems to me to run into all the standard problems: it takes only one, expansionists will tend to out-evolve others unless totalitarianly restricted, probes are cheap to an advanced civilization, almost all decision systems will have some sort of preference as to the makeup of other star systems, etc etc. A civilization that can do uploading should have no problem avoiding extinction and getting into space.

Marc_Geddes said...


Gott's argument is exactly the same as the Doomday argument that has been extensively analyzed by super-genius Nick Bostrom - Nick says the argument hasn't been proven. Better check with Nick methinks ;)

Yeah I agree that there's probably no one home as regards ETI, but Gott's argument is not a good reason for the conclusion.

Roko said...

"Gott's argument is exactly the same as the Doomday argument that has been extensively analyzed by super-genius Nick Bostrom - Nick says the argument hasn't been proven."

I love that - appeal to personal reputation to refute an argument. Dude, if Nick Bostrom said that 1+1=42 it wouldn't make it true. If you think that there's an argument which refutes Gott's Ideas, or the doomsday argument, then why not tell us?

Anyway, great post George. you've got me thinking about this stuff again, I might go do some more reading around the subject, which is always good. Hope your transvision talk goes well!

Jeffrey said...

Not to defend Copernicus, but…

What if life is abundant in the universe but just can't get very far? I've read a lot lately about just how radically different from life on earth other life in the galaxy could be. Life could function in extreme environments throughout the solar system. Lakes of methane or ammonia, frozen over oceans, hydrothermal vents, the radioactively heated cores of moons are all places life could thrive. None of these however are really conducive to the rise of a technologically advanced civilization. If life evolves on every ice moon in the galaxy but is comprised mostly of fish-like beings that live in lakes of sub zero ammonia how could they discover fire, let alone fusion or reach the 'singularity'?

Life like we know it on earth could be extremely rare in the rest of the universe. There are a vast number of variables (distance from the sun, composition of the planet, large outer planets to eat up potential impacting bodies, random geological changes that altered weather patterns to give rise to the savanna we evolved on, and countless others we haven’t considered) that have affected the development of life on earth. Maybe life like it is here is vanishingly rare and the rest of the universe belongs to the ammonia fish. It is not a bit of an anthropocentric thought to think maybe our environment is one of the few in the universe that allows an intelligent being to develop technologically if we allow for life itself to exist in the rest of the Universe. Those ice moon fish might be smart or they might be fish, but how are they supposed to build computers when they live in a lake? Those who argue for fine tuning like to say that the Universe is designed for life but looking around it appears to be designed for extremophils, not humans.

Another point to consider is that just because you can get smart does not mean that you will. We developed brains as a response to our environment. Smart is our survival mechanism. But there are plenty of things on Earth that just have big claws or good eyes. What if the ammonia fish don't need to get smart?

Nato said...

The fact that life was single-celled for more than three billion years suggests that multicellular life might be far rarer than life in general; it might be the truly rare event. Jeffrey points out other considerations that could also make the difference, though it seems to me that Earth-like planets would be too plentiful for that to explain things.

What I *do* suspect is that only systems in our generation are likely to have the proper mineral mix, meaning that George's estimates for how long life has been around might be somewhat irrelevant. If multicellular life requires a certain element mix (and cosmic ray level?) to occur with any frequency, then it becomes more plausible that the first advanced species really have arisen in the last few hundred million years. Maybe we are even the first, as unlikely as it seems.

Of course, I think that localized digital civilizations is also still a good explanation, but I have to wonder how universal an answer that can be for every civilization.

Jeffrey said...

If post singularity life was really abundant in this old Universe might we not be here? In the 600 million years that earth has supported multi-cellular life shouldn't something have popped up, hit the singularity and then torn this place down to build a Jupiter Brain?

I am of the opinion that life is more than likely very abundant in the Universe, but most of it is not like us. Look at it this way: single cell life can be very abundant (in terms of population size) in a small space with very limited resources and can survive in varied and extremely harsh environments and survive cataclysmic disasters. But as soon as that life gets more advanced, it becomes more fragile. The environments and resources to support it are more narrowed and less available. Without technology humans could only survive on a very limited portion of our own planet and in very small numbers. From a standpoint of simple numbers Earth isn't even our planet. It belongs to the single cells, which far outnumber us and can kill us in many disgusting ways. From a purely evolutionary standpoint single celled life has been phenomenally successful. It can live for billions of years just about anywhere without a second thought about digital encoding (or a first for that matter). Only the intelligent worry about intelligence.

George said...

Thanks for all the great comments. Just a couple of notes:

Given the context of 400 billion stars in the Milky way *at this current time* (and not including the billions that have preceded this point in time), the idea of 'rare' has to tempered against these sorts of available resources.

The Copernican Principle can also help us dismiss the Rare Earth Hypothesis; life doesn't appear to be something special here on Earth. It appeared *almost immediately* after rocks formed more than 3 billion years ago. The Copernican Principle would imply that the emergence of life on suitable planets is likely a common event and we're probably making too much of the different variables involved.

As for my answer about localized digital existence, I'll admit that it's hard to prove non-exclusivity. My initial thought is that virtual existence is adaptive for some reason -- that it's the only possible trajectory for an intelligence to follow. Megascale computers may be this adaptation, as the possibilities involved greatly trump the desire to do anything else. I suspect that civs naturally evolve into this state of being.

Jeffrey: Sounds like you're talking about Hanson's Great Filter Hypothesis, and that you've placed the bottleneck in our past -- that we've passed some sort of rare milestone for life. Again, using the mediocrity principle and the weakness of the Rare Earth Hypothesis, I don't think we can make this assumption. The threat of human extinction is very real and the Great Filter most likely awaits us.

Nato: You may be right that the conditions for complex life to exist have only recently been established in the universe. This is what's referred to as a phase transition. What we need to figure out is what sterilization mechanism has been responsible for this. The laws of the Universe seem fairly uniform over time, so it can't be that. Common consensus is that it must be catastrophism of some sort, namely hyper novii -- the kind that can literally wipe out 1/4 of a Galaxy (can you imagine!?). These have almost certainly decreased over time -- but when exactly did they decrease such that life could start to really develop. I would have a very hard time believing it's been in the last hundred million years or so, more like billions. Somebody should really look into this.


steven said...

Any civilization that could upload minds could eventually make self-replicating probes at negligible cost. Going virtual and colonizing the universe is not an either-or proposition.

If there were multiple big bottlenecks in creating a technological civilization, we should expect them to be more or less evenly spaced. I think Hanson talks about this in his paper. This makes the "life popped up early" argument a lot weaker. There's also the possibility life came from Mars.

What people call the Rare Earth hypothesis does *not* say the probability of life forming on a suitable planet is low (though I would argue that this is very possibly true), which is the most that the "life popped up early" argument proves. It says suitable planets are rare for various reasons.

Not all kinds of human extinction would be an actual Great Filter; an AI singularity gone wrong or something like a vacuum transition from physics experiments would probably spread to other star systems.

I had a previous comment making some similar points, but it hasn't appeared; did I say something wrong?

steven said...

In the third paragraph it should read "disproves" rather than "proves", sorry.

George said...

sorry for the delay moderating comments -- I didn't realize I had a bunch waiting in the queue.

George said...


Cirkovic and Bradbury argue that energy constraints, heat issues, and the desire to maximize computationability will forgo peripheral desires to colonize or police the Galaxy. They also speculate that a meta-ethical imperative to leave other civs alone may be in place to maximize the amount of original information present in the Galaxy. Moreover, the clock-speed of these civs would be *trillions* of times faster than the analogue world. For this reason alone participation in the Galactic community may be not worth the effort.

I know what you're going to say, and that these are sociological and potentially non-exclusive explanations, but hey, we're grasping for straws trying to find a solution to the Fermi Paradox that doesn't refer to human extinction. We'll just have to ignore Occam's Razor while we try...


steven said...

I might expand on this in a blog post later, but I just read the Bradbury/Cirkovic paper and have some big problems with it:

1. Posthuman civs don't migrate, they *expand*. Given self-replication, the cost of having some Matrioshka brains at the edge of the galaxy is for all practical purposes the same as the cost of having those same Matrioshka brains *and* colonizing all the stars.

2. They claim posthumans leave us alone in a sort of zoo scenario just in case we can come up with a more optimal way of using computing power. I find that extremely implausible. If I needed creative out-of-the-box ways to compute better, I could think of far better uses to put an entirely galaxy to than leaving it alone so a small part could be used to grow some self-destructive apes that I could outsource my research to.

George said...

Steven wrote: "Posthuman civs don't migrate, they *expand*."

Really? Then why haven't we seen evidence of this? The Fermi Paradox is a paradox because one (or more) of our assumptions is wrong. The exponential colonization model is nice on paper, but it has never been executed in the 4.5Gyr window of possibility.

The point of discussing the Fermi Paradox is to solve it, which means rejecting some of the premises.

Thus, we must conclude that either:
a) colonization is possible, but it's not done for reasons x, y, and z.
b) colonization is not possible

I think most of us agree that the answer is buried somewhere in a). I'm currently leaning towards the Great Filter as an answer. Unsatisfied with this, however, I have looked for non-catastrophic scenarios given the inadequacy of Rare Earth scenarios. Cirkovic's and Bradbury's solution is one of the best non-colonization solutions I've yet encountered.

steven said...

Scratch that and make it, "Posthuman civs, if they existed, wouldn't migrate; they would *expand*". You're right the problem needs a solution, but I believe any solution that conceives of posthumans as migrating is a bad one, because to them, having a presence everywhere in the galaxy is as cheap as having a presence in only one region.

My favored solution is the standard one that civilizations simply don't exist anywhere in this galaxy or nearby ones. I don't believe it to be the case that we know enough about all the many different Drake factors to say with any confidence that civs should emerge with non-negligible probability.

Nato said...

For the sake of argument, let's say that hardware humanity, as its lives run a trillion times faster, continues to have the same general kinds of social goals as it always has, amongst them the desire to raise young humans. That would imply to me that the human population would grow up to a trillion times faster. 1% "annual" population growth becomes a population that doubles every two milliseconds or so. One expects that resource scarcity would, after a fashion, set in quickly.

Jeffrey said...

What if we are already finding indications of local machine intelligences? Most of the solar systems we have discovered are composed of so called 'super jupiters' so what if these are jupiter brains? Just saying.

George said...

Jeffrey: The gas giants that are being discovered have an atmospheric composition comparable to our own; a Jupiter Brain, which would consist of computronium matter, would not give off these signatures.