Sorry for bothering, but I seem to miss the point entirely.
We could do a lot of things but don't do it, why should it be conclusive or inevitable that every form of life must build fancy replicating space probes just because it's possible?
Cheers
Tschachim
It's not, by any means, that "
every form of life must build fancy replicating space probes" -- but rather that
any do. The problem is this: Once started, the process carries on by itself, and grows at a geometric rate. As Hanson demonstrated, all one needs to know is 1) what the doubling time for a von Neumann probe is, 2) what the average distance between the stars in the galaxy is and 3) what the transit speed is to calculate how quickly every star system in the galaxy gets at least visited. As it turns out, even very conservative values for 1 and 3 yield very, very short time frames for such a result, something Fermi figured out in a few minutes in a less rigorous way.
Back in 1945, when the little exchange I described above took place, there were still quite a few unknown values for the variables in the Drake Equation. Heck, the Drake Equation wouldn't even be thought of for another 15 years. As Bostrom points out in his article, one crucial variable has been determined, i.e. how common planetary systems are. Applying the "cosmic principle" -- the point of Scott's post above, i.e. that we aren't unique -- allows us to make a damned good guess as to the next variable in the Drake Equation, i.e. how many stars with planets have Earth-like planets. It seems highly probable now that our galaxy should be teeming with Earth-like planets. There should be thousands, if not tens or hundreds of thousands, at least.
Again, to quote Chairman Mao, "it only takes one spark to start a prairie fire." And consider that its not just that one intelligent species that has to do it, but that one individual or small group of individuals have to do it. Assume that we DO develop the ability to build von Neumann probes within the next 100 years. What would you say the chances are that not one single person decides to send just one out over the next, say 1,000 years?
I think you the "post of the year" award for the forum, Greg, that's a good story. You should package that up and submit it to the forum's newletter, Delta-V for reprint as an article, along with a reference to this thread and the article cited in the OP.
Now, much as I'd like to agree with you, and as ignorant as I am of the math, I have to keep my skeptic hat on. Everything von Neumann and others in that sphere are projecting is so very theoretical that it's hard to buy it until the eveidence becomes real, in particular until the von Neumann devices and a means of transporting them to other systems become feasible.
I imagine that such nanomachines would, at first anyway, be very limited in capability, so that they cannot be fed just anything, but only a specific range of matter, say, hydrogen or carbon. In addition, what would they do besides just reproducing themselves? So these most simple machines would be interesting but useless. Why would I expend resources to send these things to Alpha Centauri? What's to gain? But I can believe that more advanced models could be more versatile and useful, which leads to the next question:
How would we transport them? You need a transportation system that can operate autonomously in space for decades at a minimum, but realistically it has to work for centuries, and your von Neumann machines would have to be able to maintain or replace components of it as needed, from scratch, from random matter found in the destination system, and it would have to be able to self-navigate to other neighboring systems to continue seeding.
All if this is sounds like it's possible, but it's still way out there, even for us humans, who have already thought of it and believe we are around the corner from doing it.
I keep coming back to my thought that intelligent, space-faring species are very, very rare, even where life flourishes. Here on Earth we have seen exactly one. There might have been others, but they got "filtered out", case in point might be neanderthals, a sentient, tool-using species which actually shared turf with humans but didn't make the cut. We won't make the cut, either, if the wrong catastrophe befalls us, be it natural or artificial. The circumstances that led to modern human civilization just seem so lucky at so many points along the last 4.5 billion years, so many near misses, and we have more near misses to get past before we start seeding the galaxy, that I find it hard to expect that we would see widespread evidence of "them". Add to that what Tschachim said above: not every sentient species, what few there are, may feel compelled to act as we do.
But seriously, Delta-V needs some "gravitas".
I certainly wouldn't say my post has the quality you ascribe to it. But the point of "chapter two" -- the bit about the work of Drexler et al. -- is that we now know that it actually isn't nearly as hard to build a von Neumann probe as might have been thought even 30 years ago.
Please don't picture some gigantic, clunky machine when you're picturing a von Neumann probe. Doing so is comforting, but a mistake. Comforting, because such an image carries with it all the issues of mass and the rocket equation and the kind of breakable complexity that would make a robust self-replicating interstellar probe unlikely. Instead, imagine something that's small and outwardly very simple. Very small -- perhaps as small as a few grams. Outwardly simple, because to the naked eye, a von Neumann probe in "transit phase" would appear to be an inert lump -- mainly shielding from the cosmic ray bombardment it would have to endure.
Upon approach to a star system, much of that mass is transformed into a solar sail, which is used for braking and tacking into an orbit around the star. The sail itself becomes a solar power collector, and a reconfigurable antenna and lens for gathering information about the system it is entering. Once it locates a likely source of resources, it tacks toward that target. Up arrival, it begins to reconfigure itself into a factory, using the power of doubling to quickly grow into the systems it needs to begin building a "hatchery." The middle phase of its life cycle is pretty obvious -- it makes copies of itself and launching systems. The launching systems could be something as simple as a kinetically-pumped "slinger" to things as complex as magnetic catapults and lasers to power sails outward.
So how much mass would such a device have to have in its transit phase? Maybe as little as a few grams -- no more than tens of kilograms. Its most valuable payload is information, stored redundantly and securely in its core in molecular-scale memory. It's been calculated that a mass that small could hold the information-equivalent of many, many human minds. Dozens or even hundreds.
The probe's intelligence during the transit phase would not have to be very great; far less than a human equivalent. Most of its "smarts" would be in cold storage, locked deep within a core of shielding mass to protect it from the sluice of cosmic rays during transit. And how reliable would it have to be? Not THAT reliable, especially if its "hatchery" were quite fertile. Even a fairly low success rate for reproduction would be no great barrier once the nest begins to really crank out its offspring.
You ask what these probes would DO. Benignly, they would merely sense, record and report back eventually through copies. Ultimately, with a reasonable fertility rate, one gets complete saturation of the galaxy fairly quickly in astronomical terms. Through a kind of cosmic "Brownian motion" the probes would collect and disperse information about their travels quite effectively on that time scale.
Less passively, hatcheries could do more than simply reproduce themselves. I leave what else they might do as an exercise for the reader ....
