Before getting into the paper I want to discuss today, I want to mention the new biography of John von Neumann by Ananyo Bhattacharya. I make no comment on The Man from the Future (W. W. Norton & Company, 2022) yet because while I have a copy, I haven’t had time to read it. But be aware that it’s out there – it’s getting good reviews, and given the impact of this remarkable figure on everything from programmable computers to game theory and the interstellar dispersion of civilizations, it’s a book you’ll at least want to stick on your reference list.
I figure anyone who masters calculus by the age of eight, as von Neumann is reputed to have done, is going to turn out to make a substantial contribution somewhere. I’m also interested in how polymaths function, moving with what seems effortless ease through diverse fields of study and somehow leaving their mark on each. What a contrast to our age of micro-specialization, where relentless drilling down into a single topic – and this seems true of most academic disciplines – is the mode of choice.
Image: John von Neumann, shown here with technology that might have been more to his taste, the 18,000 vacuum-tube strong ENIAC. One can only wonder what the sybaritic mathematician would have made of quantum computing. If only he were here to tell us.
It’s a good time for this book to come out, because von Neumann isn’t exactly in the spotlight these days. In a review in Science, Dov Greenbaum and Mark Gerstein note that he seems to have dropped out of public view:
In 2022…von Neumann could be the smartest person most people have never heard of. To wit, Google Trends shows that his online popularity last year was almost an order of magnitude less than that of Alan Turing, a contemporary in computing; Erwin Schrödinger, a predecessor in quantum mechanics; and Stephen Wolfram, a successor in the world of automata.
All fame is fleeting, but it’s also mutable, and the Bhattacharya biography should go some distance in pumping up von Neumann’s recognition. But let’s talk interstellar, where his name comes up today because Greg Matloff has just published a new paper dealing with what we now call ‘von Neumann probes.’ By this we simply mean probes that are self-replicating, a notion that originated with von Neumann and has now gone on to wide-ranging study. Throw self-replication and interstellar probes together and you generate various notions about how long it takes to populate the entire galaxy, as found in the work of, for example, Frank Tipler, Michael Hart and others.
Most of those exploring this space have been what Milan ?irkovi? calls ‘contact pessimists,’ who point out that if von Neumann probes could visit all stars with habitable planets in an entire galaxy, and do this within a small fraction of the galaxy’s age, their existence should be obvious. A more subtle school of thought holds that 1) dispersion need not be uniform and 2) a von Neumann probe may already be in our own Solar System, much less others, for we have only begun to explore deep space.
We can imagine these probes as having the built-in intelligence to make the interstellar crossing, which could be on the order of tens of thousands of years or more given that no biological crews need be involved. Around a target star, such a probe uses local resources – mining a native asteroid system, perhaps – to produce a new probe that, in turn, moves on to the next nearest star, or whatever target it chooses. Robert Freitas has considered self-replication in terms of nanotechnology, in which the size of the probe may be reduced to something as tiny as a needle packed with assemblers.
I come back to the question of biological crews, for without them (or perhaps given probes that carry biological materials that can be activated at destination), the von Neumann probes are free of the massive constraints of species lifespans. Miniaturize a probe to nanotechnological levels and a space-based solar-pumped laser array can push it up to relativistic velocities, possibly using materials like graphene or some kind of future metamaterial at levels of thickness no more than a single atom. But Matloff believes a 20-nm aluminum sail performing an Oberth maneuver (close pass by the Sun followed by a propulsive burn to maximize the gravity slingshot) could reach speeds in the range of 300 kilometers per second. That translates to one light year every 1,000 years.
Either way, we have a method to move human technologies out into the galaxy once our engineering is up to the challenge – the physics behind the project do not preclude this. So let’s imagine that we or some other civilization reach a stage in which we can build von Neumann probes and set them on their journeys. Matloff develops a conservative estimate of the expansion rate of a civilization using such probes.
Because of the vast canvas of time we have to work with given the age of our galaxy, we can afford to be quite conservative in our assumptions. Suppose that to minimize transit times, we say that civilizations doing this send out probes only when another star makes a close approach to the parent probe’s planetary system. Remember, the goal here is the eventual placement of probes galaxy-wide. We give up on all notions of probes reaching destinations within the lifetime of those who build them, even the lifetime of their civilization!
This gets intriguing, based on current data. The second data release of the Gaia space observatory tells us that a star like the Sun will pass within one light year of the Sun every half million years or so. This is, Matloff notes, a pretty conservative figure, for Gaia underestimates the number of low-mass red dwarfs that might also serve. Working the math, we come up with an estimated rate of expansion, granting that some stellar systems will not be suitable. After 500,000 years, we have but two occupied stellar systems. After 18 million years, we have 68.7 billion systems. Says Matloff:
This approach is only an approximation; not all stellar systems will be suitable for occupation by von Neumann probes, and some close stellar encounters will be repeated. But it does indicate that not many long-lived space-faring civilizations that deploy von Neumann probes are required to occupy the galaxy. Even if the slowest interstellar propulsion technique presented above — unpowered giant planet gravity assists — is the one selected by ET, the required galactic occupation time is not substantially increased.
Ah, the joys of exponential growth. I’m reminded of George Gamow’s treatment of such growth in his delightful One Two Three… Infinity, first published in 1947. With probes generating new probes and continuing to push outward, it becomes clear that it would not take a great number of spacefaring civilizations to occupy the entire galaxy even using nothing more than sundiver maneuvers or even gravity assists around gas giant planets to serve as the propulsion technique. Obviously, the process quickens if we reach relativistic speeds with nanotech probes that can exploit the resources they find. The process is fast enough that it’s inevitable to ask where such probes might be located if they are already here.
But first, why would a civilization choose to mount a campaign to spread through the galaxy using such probes? In the next post, we’ll consider a range of possible motivations.
The paper is Matloff, “Von Neumann probes: rationale, propulsion, interstellar transfer timing,” International Journal of Astrobiology, published online by Cambridge University Press 28 February 2022 (abstract).
The presumption is that there were other technological civilizations in the galaxy’s past. What if that is false, and we humans are the first? Then this whole edifice collapses until we are advanced enough to build these machines and start the process. Sadly we will not likely be around to see the fruits of our technology. So the Fermi Question still holds.
Suppose it turns out that the only viable self-replicating machine is life itself, and that [litho]panspermia is the only reliable way to disperse such life?
Previous estimates such low probabilities of such dispersion of life. Our technology may hugely increase this probability by designing machines to support directed panspermia. However, as the machines cannot self-replicate, we can only send such machines out from the star systems we have colonized, which will tend to be a small bubble, possibly even only Sol itself.
So it doesn’t really matter how the math works, it requires at least one early technological civilization to emerge, be able to develop such self-replicating probes, and send them on their way. Whether they reach our system depends on how long before us this civilization started sending out the probes.
The other assumption is that the self-replicating probe remains in the star system it visits. Suppose that is not true either. It may visit, offload its payload [life?], manufacture new probes and all then leave for new star systems 9the parent probe may even die by being cannibalized or destroyed]. A descendant probe may only return if it detects no biosphere successfully seeded and maturing/matured. In this scenario, we will not find any probes in our system no matter how long they have been sent because our system has been successively seeded with life and therefore no visit is required and the seeder probe has long since disappeared.
I look forward to the next post on motives for launching such probes.
Well, unless you look thru long thin tubes at distant objects, there is plenty of evidence they (AI) probes are already here. The problem is that they are not stupid like the naked apes on this planet.
https://i.redd.it/5kilixusiur71.jpg
So give us some unambiguous examples. ;)
Well, you’re a good example of a naked apes…
Of the (AI) probes; ;). Are you implying humans, or all terrestrial life, is the alien probes? In this case, aren’t we the evolved payload, not the probe? How would we test this theory against natural panspermia or local abiogenesis?
“Suppose it turns out that the only viable self-replicating machine is life itself, and that [litho]panspermia is the only reliable way to disperse such life?”
Shouldn’t we be able to determine at least the second part of this pretty quicky with our current or near future technology base? Suitable detectors around the Solar System should detect space borne microbes and see variations from different sources in different directions to compare with life on Earth. Wickramasinghe has been hypothesizing this for decades.
I have suggested this in the past. The problem is that any microbes are likely to be very sparse. We need to provide a large “net” to hope to capture free-floating microbial spores traveling at high velocity. The aerogel approach used by the Genesis probe to capture comet particles might be the way to go, but with a much larger area. Ideally, we want to ensure that we can capture both incoming microbes from other star systems and outgoing to get some idea of our systems production. Realistically this might prove a bit of an expensive snipe hunt for incoming particles.
I was looking into the question of von Neumann probes, and came to the conclusion that they are *not* an effective way to explore the galactic neighbourhood. Unfortunately I ran into a difficulty with part of the paper I was writing in order to deal with this, and then got distracted by another project, so that paper is currently part of my backlog of things to finish. So sorry, I can’t be any more specific at the moment.
Let us know more as it becomes available. Thanks.
?Oumuamua – while the natural explanations are good, I still think it’s the best current example of potential extraterrestrial tech.
The probes could stop off in the Oort cloud or Kuiper Belt if they want to stay in the shallows of the Sun’s gravity well, then drop towards the Sun when they need a boost to a new target.
There could be hundreds or thousands of probes the size of our current interplanetary probes zipping around the solar system right now and we would classify them as small asteroids – if we see them at all.
Again, as shown by ?Oumuamua they (or their solar sail-like propulsion) can be of significant size and we would have difficulty detecting them.
As always we should remember what the correct formulation of the Fermi “paradox” is:
Hypothesis: Under certain assumptions, we should see evidence of extra-terrestrial intelligence in our Solar System.
Observation: We do not see such evidence.
Conclusion: One or more of the assumptions are incorrect.
Note this is not incompatible with “rare-earth” or whatever flavor of contact pessimism you prefer. But I see too many statements on line blatantly conflating absence of evidence with evidence of absence.
That is true. However, it also implies that anything not seen might exist somewhere. Flying teapots around Mars, for instance. At some point, we need to bolster any argument with orthogonal information, e.g. the impossibility of something existing, or its probability of existence. So not “seeing bacteria” (without the right tools) doesn’t mean they are absent, and data about infections, growth of colonies on media, etc. implies their existence.
At this point, we just have no data on ET civilizations, extant or extinct, or the possibility of probes, self-replicating or not. We can conjure up as many models as we like, but they mean little without hard data. Today, we are a like a species that lives in a house that looks out the window at the wider world, but cannot yet do anything beyond the doorstep. Until we can explore the garden, the surrounding countryside, and then the world, we will have no real information about what exists outdoors. I hope that our attempts to find unambiguous biosignatures gives us some data on the frequency of life outside the solar system, and what that might imply for complex life and intelligence.
Is it physically possible for a probe, whether biological or technological, to passively keep for thousands of years uncorrupted information necessary to replicate itself and to transmit meaningful observations to home? I don’t think so, it’s impossible to stop all kinds of particles to interfere with the data. Both bits in memory and DNA must be constantly checked and repaired, this means the probe needs supply of energy and thus would emit detectable waste heat.
Bacteria have been revived from spores in ice after being frozen for 8 million years. So clearly biological life can survive for very long periods with data integrity retained. For life to survive long journeys across interstellar space, being surrounded by an ice shield to protect from galactic cosmic rays and other forms of radiation is the best way to avoid degradation.
If artificial probes can shut down in a similar way and remain viable for long periods, then a similar approach should work. Life is the proof of concept, technology must be able to do something similar.
One question is, will the von Neumann probe be up to it in the fabrication of microchips and seniconductors on an asteroid or exoplanet. Seems hard enough to do it here on Earth.
You are assuming probe technology comparable to our current state of technology. That may be as problematic as the Victorians assuming computation akin to Babbage machine, or von Neumann assuming computers need vacuum tubes (which may be made from asteroids). Self-replicating probes will not require the sort of technologies we use but are more likely to mimic life in some way. They may even be living, although with a very different biology than terrestrial life.
“Suppose that to minimize transit times, we say that civilizations doing this send out probes only when another star makes a close approach to the parent probe’s planetary system. ”
This would require an extremely odd psychology for a biological race.
Consider that if you forgot about such an approach, and just launched whenever, you could launch today, or a reasonable approximation of such, and expect your probe to arrive at its destination within, reasonably, a few hundred thousand years.
Or you could wait ten million years to cut the trip time in half.
What biological species would chose the latter course? For living beings, to not do things within your own lifetime is about the same as not doing them at all.
Actually, what species, period, would do this? Wait ten million years to save 50,000 years on the trip time? Doesn’t make sense even if you’re immortal.
You make a good point.
One reason I could imagine is if there’s a great cost to transit time. Say risk of something going wrong, off course, etc.
But that’s just devil’s-advocating. I agree w/ you.
The notion of von Neumann machines paving the galaxy is kind of an acid test. It suggests how fast the galaxy could be paved over.
Why?…
Well, imagine this: Getting a notification that our planet has been selected for impact by a von Neumann probe for establishing its replicating devices…
A moment of feeling how honored and then mulling the implications. If you object, where do you register your complaint? Up the line to where the replicating devices were already established?
And when the devices are all done setting up, what is the environment that they are engineering? Why, it’s another staging point for a trans-galactic network. Perhaps disseminating themselves like pingpong balls in a room full of spring loaded mouse-traps, a metaphor passed around in the community worried about the eventual consequences of orbital debris. Perhaps early evidence for a period of influence of the von Neumann machines in the solar system’s history was the 3.9 giga-year ago era of bombardment. Under construction?….
Issues that von Neumann devices should address: Have we been here before? What is our end- product when we are done locally?
Sure, it seems like a great way to have a mechanical amanuensis to pave the way for civilization across the galaxy. But the description sounds so single-minded that it could be just as likely be the work of a child playing with matches, the Sorcerer’s Apprentice, or the Alexa version of the Vogons from Hitch Hiker’s Guide to the Galaxy, minus their reading of poems while you are on board.
Absence of evidence of such an exercise might argue as well for an intelligence or intellect that does not need it or intervenes to prevent it.
I still maintain that the Kardeshev scaling of civilizations mainly satisfies our hope to be able to observe a civilization across the depths of space, based on such exponential growth concepts – not that exponential growth is a great benefit to a civilization. Here on Earth we have some that were vast and some that were tiny. Scaling them up and down did not necessarily mean that the quality of life remained the same. In fact, continuity might be more valuable than the capability to inflate. And power scale inflations are bound to end disastrously.
But then we could re-examine that initial premise too. Von Neumann and Fermi were contemporaries and (not having read the recent biographies ), perhaps they might have discussed “Where is everybody” and machine terraformation over lunch. But that still does not rule out
possible galactic travel by less observable means. As argued by others above. stellar proximities practically compete with world ships in solving the interstellar transit problem. And biological units seem quite akin in function to von Neuman mechanisms in their function, if Earth was tied into a cosmic system. And then consciousness itself still might have means to transmit itself in less heralded ways than sitting on the cold side of a hot plate blasted by a beam worthy of a Kardeshev Roman Numeral…
My hope is for a more subtle form of physics that will allow transit to other stars and their planets. Quantum mechanical entanglement is likely not the solution, but it is suggestive.
Another answer to Fermi is that we were seeded by a probe that gave us the Cambrian explosion…a probe long since rusted away or subducted and awaiting excavation. Want to find ET? Look down.
That hypothesis has lots of evidence to the contrary. Firstly, there is no indication that any terrestrial life has evolved from some other life. IOW, we all come from a last universal common ancestor (LUCA). Any tweaking of the genomes in the Cambrian to result in the “Cambrian explosion” is subtle. Secondly, there is [controversial] evidence that there was prior radiation of complex life. Would this be a failed prior probe visit? Thomas Gold once suggested that all life was the result of visiting extraterrestrials leaving behind their “garbage”. It could be deliberate seeding by a probe. If so, it is a variation on a supernatural force creating terrestrial life – a common creation myth of many religions – but with no supporting evidence and currently, no way to test.
If one wants life to have been seeded, why not panspermia from a nearby, early inhabitable Venus, or Mars? This would be a preferable hypothesis to ET probes.
This is a fascinating discussion. Avi Loeb is now leading the Galileo Project which is actively searching for possible observable robot probes already in our Solar System. (Maybe Oumuamua?)
https://projects.iq.harvard.edu/galileo