You have probably already heard of this, a robot under development by Honda called Asimo. It’s appeared in a Honda TV commercial or two and there’s been many news stories written about it. It’s a bipedal robot, stands upright, walks, can pick up a newspaper (at least in the TV commercial), etc. Most recently, Asimo has learned to ‘run’, albeit a slow run, more like a jog. Technically its feet never leave the ground together, as human feet do in a jog or run, but it still looks pretty realistic and I’m sure it won’t be long until it and other robots like it are able to sustain faster and longer runs than humans. One day robots like these will deliver our mail and pizzas. It’ll be quite the day when we see the first robotic track meet on ABC’s Wide World of (Robot) Sports. In the meantime, I can’t help but marvel at this thing and how ‘humanoid’ it already is.
Robotics isn’t a brand-new field, but it’s fairly recently that real robots have been acknowledged in the mainstream (outside of sci-fi films, at least). Even Lego has the Mindstorm product line which has been very successful, and is helping to get kids involved with robotics and computer programming at an early age. (Lots of darn cool Lego projects are out there, the coolest one I’ve seen recently being about logic gates built with Lego Technic parts; while it wouldn’t really conform to Moore’s Law, a functional computer built from Lego parts would be way cool.) Stores like the Sharper Image sell toy robots like the RoboSapien, and Sony’s robotic dog AIBO has been out for a while now too. While these are really just fancy toys and have limited practical use, I admit it would be cool to play with these for a while. Maybe program the RoboSapien to take AIBO for periodic walks. They wouldn’t even need to bring a scooper along, but in the event that AIBO does drop any parts, they could always have a RoomBa following close behind.
Working in the EEC and watching our engineers build large test environments with racks of servers and piles of KVM and network gear, I’ve marveled at how an entire enterprise computing environment can get up and running within just a few days (granted, our engineers are awesome at building these environments). I’ve wondered how long it might be until the day when, using robotics, a computer system would be able to build onto itself, adding new machines when they’re needed. Imagine a forest of servers, one domain of which would be connected to robotic assemblers in a warehouse of various parts. When the forest or a domain within it begins to reach a preset amount of member servers and needs to expand, or an application begins to reach the limits of users or storage and needs to scale out, the robotic domain would kick in and build new servers from bar-coded parts in the warehouse, rack them up and connect them to the network. Several scripts would then run to push an image of the OS and other apps via RIS or ADS to the new servers, join them to a domain or create a new domain with them. Humans would only need to be involved for occasional maintenance and for keeping the parts supplies stocked; eventually those roles would be automated with robots as well. This sparks scary visions of SkyNet or the Matrix, or at the very least the lesson from the movie WarGames that advises not to “take the men out of the loop”; I imagine that for those and other political/religious reasons this vision will not be realized anytime soon. But from just a technological perspective, I bet it could probably be accomplished within the next 5-10 years. In the words of Neo, “Whoa.”
One might be inspired from all these robots in the news to look into taking out some Robot Insurance from Old Glory Insurance. “For when the metal ones decide to come for you - and they will.”