This is the future of space exploration. Which is why we have to stop talking about "exploration" as a justification for humans in space.
[Update in the evening]
Commenter Paul Dietz recommends >Saturn's Children
as a relevant book on the subject. If it's like most of Stross' work, it's hard to go wrong.
I don't buy it. Even with robots ten times more capable than the best available today, they can still only look for stuff they're built to see -- in other words, they can only find what the designers expect to find.
Now, absolutely robots can be a tremendously powerful tool for exploration, but ultimately a person has to go look and poke around. I don't see that changing any time soon.
This looks like an interesting tool, but it will enhance rather than replace human exploration, provided that the project succeeds. So exploration is still a viable justification for humans in space.
No, Rand is correct. 2020 seems a little ambitious to me, but I don't think it's terribly off either. The Grand DARPA Challenges for self-driving cars really "drove home" (ha ha) to me how quickly this sort of technology is advancing. As quickly as Moore's law propels the hardware, advances in software efficiency and intelligence are even faster.
I would not be surprised at all if one of the mini-rovers that's going to be launched as part of the Google Lunar Challenge is self-directed.
One of the key understandings here is that sending a probe of this nature will not be better than (or even half as good) as sending a geologist with 20 years of experience and a fully stocked lab. Not for many decades, at any rate. But you simply can't feasibly send a geologist to Io or Ganymede; so given the paucity of our knowledge of those places, probes will be far better than nothing.
Plus, when you're no longer limited by the C&C capacity of Houston, you can do cool stuff like mass produce the probes for cheap per unit prices (like a Dell computer driving a Model T) and send three to every moon in the solar system and map the entire asteroid belt. Even if a real geologist is 10x as good, you can make 100x as many.
"Even with robots ten times more capable than the best available today, they can still only look for stuff they're built to see -- in other words, they can only find what the designers expect to find."
This is fairly common misconception about machine intelligence, which is not the case at all. Look a bit into modern adaptive and genetic algorithms and machine learning. An algorithm isnt by far limited to smarts of its designer, when it has a huge database/knowledge set to draw upon, and if it is permitted to learn and evolve.
Algorithms are already designing better hardware than humans can for example.
Case in point
http://technology.newscientist.com/channel/tech/mg19526146.000-evolutionary-algorithms-now-surpass-human-designers.html
The fact is the probes don't have to be that smart. The A.I. in your average strategy video game has almost had the require abilities for years, which is to say map any area that you haven't visited yet and collect samples or data for anything that isn't similar to what you have already sampled. Make it a priority not to waste fuel.
Trying to make a probe as smart as a man is not going to happen any time soon. It's enough that many cheap dumb probes that give us choices to consider so we can decide what's worth sending people to investigate.
Several different matters in the discussion:
Exploration, especially given rapidly improving robotic capabilities is a weak justification on sending humans to space ( i believe thats Rand's original point which i agree with )
Robotic capabilities have for long surpassed humans in quite a few areas, and are reaching human levels in other areas ( yes, also in areas of "creativity" and adaptiveness which is my main point ) which further weakens the case for humans exploring.
And yes, probes dont have to be as smart as humans. But if technology permits it and it doesnt cost anything extra ( software unit cost is nonexistent ) it doesnt hurt either.
As for humans in space, this is an entirely separate matter. We go because we want and can, soon.
This, and other reasons, are why I always use the terms spaceflight and space development. For most people, when you say exploration, you mean science.
I would argue there is more to exploration than just science, but that is how the general public views it. We need to adopt development, or even colonization/settlement, rather than exploration
In the case of North America, exploration and development went hand-in-hand.
It is a helluva lot easier to explore a place if you live there already.
No, in North America exploration came first and development came second. See also, Columbus vs. Conquistadors, Lewis & Clark vs. Manifest Destiny.
Of course in Star Wars (a universe with econimical FTL ships), probes were sent out with the explicit intent of sending a human (or other biological) to follow up on positive responses...
Not off-topic for this thread...
If you are a Heinlein fan, I recommend reading Charles Stross's latest book, _Saturn's Children_. It's a nice late-Heinlein pastiche with no biologically human characters, but plenty of non-biological ones. It has nifty/realistic (if somewhat consternation inducing) space technologies, and the obligatory female character with a nipple going "spung".
We need to keep in mind that in North America there already existed some infrastructure due to native Americans. A number of early exploration and exploitation efforts were driven by the existing infrastructure. For example, the Spanish wouldn't have worked so hard to colonize the New World if it wasn't for the wealth of the existing empires of that region.
And Lewis and Clark would never have made it without help from the natives.
Neither of which changes the fact that before capital can (will) be committed to development, you need information on the lay of the land.
Also, both are good arguments for why robots should go first - to establish forward operating bases. They can be the advance teams.
http://selenianboondocks.blogspot.com/2008/06/westward-ho.html