« Visiting Hamilton | Main | More Travel »
May 1, 2008
Fermi's Paradox
Nick Bostrom writes an interesting essay on anthropic observations, as usual. I agree with him that the answer to Fermi's "paradox" --- "why hasn't spacefaring life colonized our entire universe in ways we can detect" --- is that there isn't any.
But I think he misses a powerful argument that the "Great Barrier" to spacefaring life is the difficulties in our past, not the dangers in our future. It seems that if current progress continues, we're at most hundreds of years away from developing self-sustaining, space-capable artificial life --- AI, if you like --- even if we take the brute-force approach of brain simulation. It's hard to think of inevitable catastrophes that could wipe out all of a multitude of space-based intelligent machines with reproductive capability, even they were all confined to our solar system. Even if we don't make it to that point --- and I will not be surprised if we (or God) write an end to our history --- since we made it this far, given enough other civilizations one of them would be luckier and make it all the way, and go on to colonize the universe in observable ways. But apparently they haven't.
Posted by roc at May 1, 2008 12:11 PM
Comments
The universe is pretty big (estimates from Wikipedia figures suggest that our galaxy has a diameter of around 0.003% of the entire universe).
I don't know what percentage of the universe is observable to us, but I'm going to go out on a limb here and suggest that it's somewhere between "limited" and "basically nothing".
Drawing any conclusion based on what we can (or rather can't) currently observe seems somewhat bold. I can't detect the existence of any god (or gods), but I'm not going to dismiss the possibility based on what I can't see.
Posted by: Ben Basson at May 1, 2008 2:59 PM
Well, I messed up that percentage with a typo in my calculation. Turns out it's more like 0.0001%.
Posted by: Ben Basson at May 1, 2008 3:01 PM
The thing is, if spacefaring, colonizing life emerges frequently in the universe, some of it should have emerged close enough to us, and long enough ago, to have expanded to be all around us, even right here in this solar system. Bostrom points out that 20M years is much more than enough time for a civilization emerging in our galaxy to occupy the whole thing. Give them another 200M years and we're in range of a lot of nearby galaxies.
It's true our observations are consistent with all civilizations preferring not to expand and colonize --- just sitting in some hard-to-observe corner of the universe --- but that seems unrealistic. Bostrom covers that in his essay.
Posted by: Robert O'Callahan at May 1, 2008 6:22 PM
Well, I'm not quite as optimistic as you about the future... we have a few hurdles to overcome there too, resource issues and population pressures being the biggest (and these are in many ways closely related).
That's not to say that one of the biggest hurdles we had to overcome was simply existing at all (I suspect technologically capable intelligence is a rather rare environmental adaptation), but that we're not "out of the woods" yet as it were.
Still, an interesting article.
Posted by: D Milne at May 2, 2008 12:57 AM
Since we can essentially observe next to nothing, it's awfully difficult to draw valid conclusions based on what we can observe. In fact, it's hard enough even to decide whether we can draw valid conclusions.
The problem is that the universe is big. Really big. Communication is next to impossible. Colonization is completely speculative, or more realistically, in the realm of science fiction. This requires that you suspend disbelief: the notion of colonization requires amazing technology that somehow overcomes immense times and distances or somehow circumvents the laws of nature as we understand them. Could happen, I suppose. Or maybe not.
You can postulate that somehow a few microbes get blasted into space on a meteor that wanders the universe and comes to rest on a hospitable planet--and that those microbes evolved into intelligent life. Could happen, I suppose, but maybe it didn't. Well, maybe organic molecules just form in interstellar space and evolve into intelligent life. Or maybe not. Maybe they'll find some intelligent moss under some icy rock on Mars. Or maybe not.
I was sort of hoping they would find Sasquatch and the Yeti too, but so far, nothing. I just don't see how we'll ever know if there's anybody out there.
Posted by: VanillaMozilla at May 2, 2008 5:37 AM
"Bostrom points out that 20M years is much more than enough time for a civilization emerging in our galaxy to occupy the whole thing."
This remains to be proven. I doubt whether 20 M years is even enough time to occupy the nearest solar system. I'm not sure it's sufficient time even to travel to it.
Posted by: VanillaMozilla at May 2, 2008 6:07 AM
Since you are a convinced Christian, don't you think that you are actually already in contact with extraterrestrial intelligent life? Granted, it's not exactly what Nick Bostrom is talking about, but that's not surprising considering how hard it is at any given point in time to predict the exact path a civilization will take from the current situation it is in.
It has a lot to do with what man expects the future to be, which is directly influenced by the state of development he is in, another consequence of the anthropic principle. In other words, there could be another filter at work here: the one that makes it impossible for us to even think about something that's too far outside of what we have achieved so far.
Posted by: laszlo at May 2, 2008 8:14 AM
D Milne: like I said, I'm actually not all that optimistic about *our* future.
VanillaMozilla: there's no reason interstellar travel should be hard for intelligent machines. Bostrom discusses that a bit. Reaching speeds of 0.01c is not hard. Travelling 100 light years --- a 10,000 year voyage --- is not a big deal, the machine can just sleep for as long as it takes.
laszlo: interesting point.
Posted by: Robert O'Callahan at May 2, 2008 11:31 AM
I dunno. Taking an extremely conservative approach to the whole colonizing business, assuming 0.01 C as the average speed possible (thus 500 years average between stars), and being pessimistic about how many stars have colonizable/terraformable environments (set to 1%, meaning typically 10 LY between colonizable stars), a (live, not robotic) civilization that only moves on to colonizing a new world after an existing world is terraformed and stable (large supportive natural population that is reasonable to economize building another 1000-year colony ship), I can see it taking up to 4 trillion years just to colonize a single galaxy like ours.
[Note: YMMV on the numbers; taking several very rough personal estimates.]
Being slightly more aggressive (.1C travel, expansionist rather than colonist policies) could still take a few billion years.
Of course that all assumes that there are never any issues which might set the civilization back (civil wars, change in political philosophies, etc), nor intrinsic limitations on perpetual growth (communication between the original world and the outliers becomes more and more difficult the larger the empire grows). Simply getting a civilization to invest in a project that won't even reach its starting point for 1000 years, never mind becoming usefully independent and able to provide a return on the investment to the original world, would be quite a challenge.
If your only goal is to plant a flag on every single world in the galaxy to say "Kilroy was here!", sure, you could do that in a few tens of millions of years. But to equate that with the time it would take to grow a stable civilization seems a bit presumptuous.
Posted by: David at May 2, 2008 11:31 AM
David, that assumes each colony only sends out one new colony so overall growth is only linear.
What would really happen is that each colony can send out many new colonists --- expanding in all directions at once. Then the total time to colonize the entire galaxy depends mainly on the diameter of the galaxy. So even if it took 10,000 years to spread 10LY (that's very slow, 1000 years of 0.1c travel followed by a 9000 year holiday), we're talking less than 100M years to colonize the Milky Way.
Posted by: Robert O'Callahan at May 2, 2008 11:38 AM
It does indeed assume a linear growth. I considered a more exponential pattern, but eventually thought better of that. Other than the initial foray from the home planet to sort of get things jump-started, what value is there in continual colonization? What exactly do you get out of it?
It's going to take a number of years to build each colony ship (something that's expected to last a minimum of 100 years of space travel), and building such a ship requires a large number of resources, including coordinated expertise, raw materials, and appropriate processing capability. Once a colony is established, it needs to: initiate terraforming for habitability; establish living conditions; explore and develop resources; and allow for sufficient population growth and political stabilization to be relatively self-sustaining before it even considers building colony ships of its own (assuming it's intended as an actual colony and not just a way-point).
How long will that take? How many people can you reasonably send on a colony ship, and how many generations would it take to grow into a fully viable/independant world?
So what about established worlds? Why not have them continue to send out more colony ships? Well, at some point it becomes self-defeating. With a limited distribution of viable colony-stars, eventually it's going to take longer for the ship to reach those destinations than it will take for the initial colonies to grow into healthy communities that can send their own ships out that are going to be closer to the destinations anyway.
Unless, as I mentioned, you're merely intent on reaching every possible colonizable location as quickly as possible, in which case you can spray out colony ships by the dozen. However every ship built is a collective resource sink, and a sink that won't realize any value for hundreds, perhaps thousands of years. If you're a mature enough species to be able to think -that- long term, more power to you. However having only the example of humans to work with (eg: the ISS), getting enough agreement to put together more than a handful of such coordinated efforts that may start providing us returns within a few hundred years seems a bit of a stretch.
So now you've established your alien empire in a tiny little corner of the galaxy -- say 10,000 colony worlds out of 1,000,000 stars (a tiny fraction of 1% of the galaxy). You still (as far as we know) have no faster-than-light communication, and it's 400+ light years from one end of the empire to the other. That's like a 20 generation separation between different parts of the civilization. How do you coordinate ideas and laws across that distance? For that matter, who decides who gets to build more colony ships at that point? There's a huge amount of potential redundancy and wasted effort in a network like that.
Essentially, at that point, each individual colony world has to independently evaluate the benefits it could gain from sending another colony ship out. Each one also knows that all the other colonies are doing the same (or already did so a few hundred years ago), so of all the potential colony-star targets, any that are closer to other existing colonies are likely to have already been targeted by those other colonies, which means there's much less reason for -this- colony to consider any but the closest available targets. As the outer surface of the empire grows, those available targets will continue to be mostly linear in growth.
So why continue to grow at all? With 10,000 planets, surely you have enough potential resources at your disposal to do nearly anything you want? Well, eventually the size of the empire is such that it is in fact easier to create a new colony to gather the resources you need than it would be to request it from another part of the empire. Consider a request sent from opposite sides of this empire: 400 years for the message to get from A to B, and 4000 years for the requested materials to get back to A (assuming immediate responses). Or you could send a colony ship (100 years), let an entire new planet grow up (1000 years) and send stuff back to you (100 more years) three different times in the same period.
So, the larger the empire grows, the more difficult it is to make use of the value of all those existing planets, and the easier it is to make use of (a limited number of) new planets. That, however, assumes a consistant approach by all parties. Why would it not be easier (greater value for a given expense) for those disconnected outer planets to "recolonize" planets that have already had all the hard work done for them? IE: Why not build a war machine and take the planets closest to you that have what you need? This is actually most likely to occur within the older, innermost planets that see less and less value from any additional colonized planets.
In order to get around some of these problems, you need to start making assumptions about the aliens: they've achieved hugely long lifespans that can afford to take the long view on planning; they're less prone to self-destructive behavior; etc. However the only thing we are given at the start is that they are intelligent, and are able to assess the value of a long-term proposition like space colonization. I have in fact added a third assumption, tied to the second, in figuring that a certain degree of selfishness is likely to be inherent in self-aware beings, and that thus typical issues of game theory are likely to come into play.
So as a long, roundabout way of getting to the point, anything greater than linear growth would appear to me to be exceedingly optimistic. I'd actually be interested in seeing the opinions of someone who's highly experienced at networking (traditional hardware networking, neural networking, or social networking) who could analyze the probable growth patterns.
Posted by: David at May 3, 2008 12:12 PM
David, you seem to have missed the point about the colonizers being intelligent machines, not organic life.
Machines don't need terraformed planets. They don't need planets at all, they can "live" just fine in space, collect solar power and get raw materials from asteroids, moons, etc. There are no obvious limits to their lifespan. They probably don't need to form large societies in order to be self-sustaining; it seems completely plausible that a single machine would be able to reproduce itself an unlimited number of times. They can power-off for long interstellar voyages, so these "colony ships" are not vast endeavours, just a machine mind and its robot agents attached to some propulsion device, plus some shielding and an alarm clock to wake it up on arrival.
I agree that colonization of the galaxy by organic life as we know it is highly implausible, but that's not what we're talking about. (It's possible the machines might seed organic life as they go for some reason, e.g. if they were programmed to, but that's a different proposition.)
Posted by: Robert O'Callahan at May 3, 2008 7:04 PM
Err.. Sort of. The original article did not make any such distinction between organic and inorganic form (as you pointed out in your blog post), yet still assumed the 20 mil year colonization period. That's mostly where I launched from.
As for the inorganic form progression, that radically changes the motivations and requirements of the beings in question, to the point where I cannot comfortably make any particular assumptions about their actions (but will of course have to make a few for the purposes of explaining myself). Taken to its extreme, though, such a civilization doesn't need to create colony ships or colonies themselves at all (what does a self-contained, effectively immortal, space-faring artificial being need with a planet-bound habitat?), which itself destroys the argument of finding alien-colonized worlds all around us by this point in time.
It also makes their population density vs area of explored galaxy effectively undeterminable. How many of them do you actually need? If a few billion (and that seems an overly generous number) can explore every corner of the galaxy within 20 million years then you'd still be lucky to find even one at any given star system. Further, once they're done exploring this galaxy, why not launch themselves to another one? Why any expectation that such beings would still be around at all?
If they are colonizing worlds in lieu of original organic colonists, but at a far greater pace, then that just exacerbates the issue seen on the organic side; they're creating resource 'source' points far in excess of the civilization's capacity to make use of those resources, further and further from the point where those resources can in fact be used, to the point where it's essentially impractical to even think of them as resources due to the distances involved. And if they're not trying to create a new resource 'source' then what exactly are they doing? Planting 'Kilroy' signs.
I suppose one potentiality is that the AIs would still colonize and create their own machine worlds, though that begs the question of "Why?". At best, perhaps they seed worlds with organic life as a guarantee of the continuation of life within the galaxy, and in that case we return to the original question of why we haven't encountered aliens of some form. But that itself leads to the possibility that earth itself is a seeded world, and as such brings into question the validity of our data, or the possibility that we're just in the middle of an organic>inorganic>explore>seed>leave cycle that doesn't have much in the way of inorganic presence (or that such presence is too diffuse to pinpoint with our technology). In no way does changing the nature of the entities suggest or deny the possibility of an alien presence merely because we aren't detecting any at the moment.
I could make further speculations along the lines of 'detectability', of course. Radio transmissions that were detected in (local) 1950's just a couple stars out (10 LY away) would still require (by earlier assumptions) 100 years for the entity who detected it to reach here. For an entity that apparently plans in terms of millions of years, 100 years would be a trivial amount, and it may consider it easier to discuss things in person than try to explain over a one-way 'phone'. That would still put it 50 years away from reaching us, and that's for something that, galacticly speaking, is right in our backyard.
Ultimately it still devolves back to the original question of "Why?". What motivations would said aliens have for creating AI ships (I'll grant that there several valid reasons already outlined)? What benefits do they get from doing so? And what implications do the answers to those questions have on expected behavior patterns/entity distribution/colonization attempts/etc (this, however, remains unanswered)? Ultimately, why is it assumed that they must necessarily act in such a way as to be (easily) detectable to us?
If anything, I'd consider an AI expansion to be -less- detectable than an organic expansion. The earlier posit about a 10,000 colony empire in a small section of the galaxy would seem to be something far easier to detect (via tradional methods of radio frequency scanning, considering that they're at most 100,000 LY from us and such an empire should have taken a fair bit longer than that to develop) than a billion AI units dispersed throughout the stars. If, as you posit, an inorganic expansion is the more likely of the two then the argument that aliens don't exist because we haven't detected them yet seems even weaker.
And on the sci-fi front semi-related to browsers: data security and rival factions. Must we assume a single unified, fully trusted source civilization? How much of your information would you trust to open broadcast, even encrypted, against advanced AI units that have the luxury of a few thousand years to brute-force the transmission (not even considering the possibility of quantum decryption) and, via self-replication, are capable of building data centers of whatever size necessary for the task?
[Alt example for the organic side: consider all the empires in human history, what allowed them to grow, and how large, and what caused them to collapse. They were all to a large extent limited by the speed of travel and communication. The numbers change in space, but the overall idea still applies.]
In other words, it's all well and good to say the galaxy should be swarming with evidence by now, but it's a lot more difficult when you start considering what real obstacles such entities may have to overcome, rather than giving them an idealized environment.
Posted by: David at May 3, 2008 10:17 PM
Bostrom covers these arguments "maybe they don't like to expand", "maybe they don't like to build" in his essay, to some extent anyway.
He points out that if just one out of all the AI civilizations in the galaxy is expansionist, it will expand throughout the galaxy. If just one of those likes to send signals of the type we could detect, our galaxy will be filled with such signals. If just one likes to build stellar-sized structures we could detect, we should be able to detect them.
The alternative is that something makes all advanced civilizations non-expansionist and quiet, which would seem surprising, since life as we know it is ubiquitously neither. Even for a super-advanced civilization, internal and external competition would seem to favour the supremacy of expansionists.
I suppose one remaining hypothesis is that the galaxy is dominated by a single super-civilization which chooses to remain quiet and hidden and suppresses all nascent competitor civilizations. For that to work, it would want to have probes in every solar system capable of developing life, ready to blow up that life before it switches to machine intelligence and becomes much harder to stop. Seems unlikely.
Posted by: Robert O'Callahan at May 3, 2008 10:52 PM
I've been thinking about Fermi too, and here's what I came up with:
http://michaelgr.com/2008/05/09/virtual-reality-could-explain-the-fermi-paradox/
I think it's a slightly different angle form what we usually see on the subject.
Posted by: Michael G.R. at May 10, 2008 7:52 AM