As eager as he is to invent robots that can travel to a moon of Saturn or Jupiter, and function autonomously in these hostile environments, Rodney Brooks would love a shot to explore space himself. “I made an offer to Jeff Bezos, Larry Page and Sergei Brin that if they would fund a one-way mission to Mars, I’d go on it,” says Brooks. But he knows that robots are cheaper to send than us, “big bags of skin with biological processes requiring replenishment of all sorts.” Under the Bush Administration, NASA first laid out an ambitious program in robotic technology, involving sending machines to reconnoiter the moon and Mars and prepare habitation sites for humans. “Robots would dig channels, then lower habitation modules into them, and when people come, they’d live like moles underground,” says Brooks. But why send people at all if these robots can accomplish so much? It turns out that there’s a dangerously long lag time between sending a command to a robot and having the machine perform a function. Ultimately, human senses and timing will be needed on site.

But now NASA’s grand robotic research plans are on hold, says Brooks, blocked by the difficulties and enormous expense of designing a new launch vehicle. The future of sophisticated robotic work seems earthbound, says Brooks. First, there are military innovations -- Congress has mandated that by 2015, 1/3rd of all US military missions should be unmanned. Also, the oil industry is pushing for machine-based solutions to such gritty problems as deep-ocean drilling and oil-well maintenance. And don’t forget the new billionaire space cowboys, who dream of mining platinum fields on asteroids (for fuel cells on earth), or building space tourism businesses. But, Brooks reminds us, we have a way to go: After 40 years of research, “the generic object recognition that a two-year-old child could do, we can’t do with our robots.”