Projections about the near term trajectory of future technologies suggest a revisit of the Zoo Hypothesis (Ball, 1973) for the so called Fermi Paradox. This article recasts the hypothesis in an updated context with an eye toward machine intelligence, technological singularity, and information transfer as a means of interstellar travel (Scheffer, 1994).
Since we don't know exactly what to look for, the search for extraterrestrial intelligence necessarily involves a good deal of conjecture about the nature of ET. If we ever do discover an ET civilization, it will almost certainly be millions of years more developed than ours. This search, therefore, is necessarily informed by far future projections of our own technological progress. While such long range projections are clearly beyond our reach, much can be gleaned from near term predictions by futurists. Indeed, in a historical context, we find ourselves at the knee of a geometric growth ladder that casts the steps behind us as quaintly short, the ones ahead as dizzying fast, and the present ever harder to anchor. As we learn our future, so too must we adjust our search for ET.
Information Transfer As Means of Travel
In Machine Intelligence, the Cost of Interstellar Travel, and Fermi's Paradox (1994) Scheffer argued it is way cheaper to beam the information (bits) necessary to print an interstellar probe at the destination than it would be to physically propel the probe there. By now, this idea is a familiar theme with companies vying to mine the asteroid belt: it is generally understood that it would be far more cost effective to build the mining equipment on location than to ship them from Earth. And a good deal of this on-site manufacturing will involve printing 3D objects which may then be assembled into larger, useful objects. The blueprints for these manufactured objects of course originate from Earth, and we'd soon be able to transmit improvements to these blueprints at the speed of light.
The Printer As Computer
If the on-site manufacturing of asteroid mining equipment does not fully capture the idea of a general purpose printing technology, we can still contemplate it in the abstract (since we're considering technologically advanced civilizations). So first with a provisional definition..
General Purpose Printer (GPP). A printer that can print both simpler (less capable) and slightly more advanced versions of itself.
It's provisional because ideally one would strive to define it with the same rigor as, say, in asserting that a general purpose computer must be Turing Complete.
Perhaps the idea is better captured in the following Tombstone diagram (borrowed from compiler-speak).
|Bootstrapping the printer.|
Here the bottom "T" represents the printer. Given a blueprint (B), it operates on material and energy inputs (M/E) and outputs similar objects. The upper "T" (written entirely in the "blueprint" language) bootstraps the lower one to produce a more capable printer.
The evolving general purpose printer. From a small kernel of capabilities ever more complex designs can be instantiated. (The kernel here presumably needs a small arm to start off.)
Note we don't necessarily have to pin down the exact technology that enables this fuzzily defined GPP. Kurzweil, for example, suggests it must be nano-technology based (The Singularity is Near, (2005)), which seems reasonable when you consider you also need to print computing hardware in order to implement intelligence. Regardless, a technologically advanced civilization soon learns to manufacture things at arms length.
The Printer As Portal
A GPP parked suitably close to material/energy resources functions much like a destination portal. It's an evolving portal, and it evolves in possibly three ways. One, from time to time, the portal receives code (blueprints) that make it a more capable printer. Two, the printer accumulates and stores common blueprints that it has printed thus allowing future versions of those blueprints to be transmitted using fewer bits. And three, if the printer is intelligent it can certainly evolve on its own. Although, from an engineering perspective, you probably want this intelligence to be more like a guardian sworn to the principles of the portal, whatever those are. (One sensible requirement is that it shouldn't wander away from where the sender expects it to be.)
Time and Information Flow
Although this form of travel is effectively at light speed (and consequently instantaneous from the perspective of the traveler), the vastness of space separates points of interest (such as our planet) greatly across time. Distances across the Milky Way are typically measured in tens of thousands of [light] years. Enough time for an alien civilization to miss the emergence and demise of a civilization on a far off planet (hopefully not ours). Assuming intelligent life is prevalent across the cosmos, even with on-site monitoring, word gets out late that a new civilization has emerged.
Earth has been an interesting planet for about a billion years now and should have been discovered well before humans evolved. It's not unreasonable to hypothesize that one or more GPPs were parked nearby long ago. Those GPPs would have had plenty of time to evolve--sufficient time, perhaps, for the singular culture the Zoo Hypothesis requires to take hold.
Kurzweil predicts that humanity's artificial intelligence, manifested as self-replicating nano-bots, will one day, soon on a cosmological scale, transform the face of celestial bodies about it, and the universe with it, in an intelligence explosion. Here I take the opposite tack: an intelligence explosion leaves little trace of itself.
For once you can beam blueprints and physically instantiate (print, in our vernacular here) things at arms length, there's little reason to keep physical stuff around when they're done doing whatever it was that they were supposed to do. As long as the memory of the activity (of meddling with physical stuff) is preserved, the necessary machinery (like that mining equipment on the now depleted asteroid) can be disassembled and put away. An information-based intelligence has little use for material things; it is more interested in their blueprints.
This is not to say super intelligent ETs do not build things from matter (and leave them there). They likely need to build much infrastructure to support their information based activity. But as communication speeds are important in any information based activity, this infrastructure would have to be concentrated in relatively small volumes of interstellar space. In such a scenario, there's little incentive to build far from the bazaar.
I don't particularly like its original form because, as Ball also notes, the hypothesis doesn't make falsifiable predictions: "[It] predicts that we shall never find them because they do not want to be found and have the technological ability to insure this." The step forward, it seems to me, is to attempt more specific postulates (such as the printer portal introduced here) that are still in keeping with the broader "deliberately avoiding interaction" theme.
If the hypothesis is broadly true, then there must be a point in a civilization's technological development beyond which they (the metaphorical zookeepers) will no longer eschew interaction. Which suggests a protocol to start the interaction.
The search for extraterrestrial intelligence ought to aim to systematically confirm or rule out the zoo hypothesis. A zoologist looking to document a new species might well parse tribal lore and anecdotal evidence for clues.