Home | Writings | Resume | Links | RSS Feed
A Proud Member of the Reality-Based Community
About Me :
I'm a grownup nerd living in the Boston burbs. I write computer programs for a living and plays for fun.
I'm married to a wonderful woman, and we share a nice little house with our daughter and our cats.
I'm a humanist, a technologist, an artist, and an idealist. I believe in reason, freedom, love, equality,
and democracy. (Did I mention that I'm
an idealist? I did, OK.)
I'm also a pragmatist and an empiricist. I reject ideology and dogma, especially when they conflict with
practical facts (i.e., pretty much always). I particularly hate willful ignorance, which tends to go hand-in-hand
with ideology and dogma.
Like the alignment of the planets, this blog gets updated as I have the time, inspiration, and inclination to do so.
Like the alignment of the planets, this blog gets updated as I have the time, inspiration, and inclination to do so.
Saturday, December 20, 2003
The Future Is Sooner Than You Think
I am astonished by Sony's Qrio robot. This amazing machine is, unfortunately, not for sale just yet. Sony was said to be mulling selling this robot, at a price mentioned as about that of "an expensive car", but has decided, for the time being, to use the robot as a development platform and a showcase of Sony's technological prowess instead. What a showcase! This amazing machine is a humanoid bipedal walker that can balance itself, walk over uneven surfaces, protect itself when it falls, and get back up! Sony has demonstrated the machine dancing and even jogging! It's not very useful -- not yet -- but it has demonstrated an amazing range of abilities.
This machine doesn't exist in a vacuum -- there are lots of other similar robots out there, including humanoids from Fujitsu and Honda. Humanoid walking robots have been under serious development for nearly two decades now. That might seem like an awfully long time, but it's more or less normal for new, fundamental technologies. I think we're approximately at the cusp of robot development, where the performance/price curve just begins to take off.
The primary limiting factor in making an independent walking robot is the amount of computing power that can be economically applied to the control task. (The secondary limiting factor is the amount of electrical power that can be carried around. I notice that the Sony and Honda models have both shrunk in their latest incarnations, and it's no accident: it takes a lot less power to move a smaller robot around than it does to move a large robot.) Both of these technologies are making dramatic improvements, but especially the computing part. It just takes a lot of computation to do the sorts of things these machines are capable of doing. How much computation? Nobody really knew before now, that's how much. One of the benefits of building robots is that it quantifies the computational requirements of robots. As computers continue to yield more cycles per second per dollar, as memory and power densities improve, you can expect that humanoid walking robots will become more capable and more affordable.
Some people think that we're at the same point in the evolution of personal robots as we were in 1975 regarding personal computers. The market was ready for PCs in 1975, and the technology was relatively weak, but the stage was set for a revolution. Look at your computer now, and compare it to an Apple I, and you might see where robots are going, and how long it'll take to get there.
I am astonished by Sony's Qrio robot. This amazing machine is, unfortunately, not for sale just yet. Sony was said to be mulling selling this robot, at a price mentioned as about that of "an expensive car", but has decided, for the time being, to use the robot as a development platform and a showcase of Sony's technological prowess instead. What a showcase! This amazing machine is a humanoid bipedal walker that can balance itself, walk over uneven surfaces, protect itself when it falls, and get back up! Sony has demonstrated the machine dancing and even jogging! It's not very useful -- not yet -- but it has demonstrated an amazing range of abilities.
This machine doesn't exist in a vacuum -- there are lots of other similar robots out there, including humanoids from Fujitsu and Honda. Humanoid walking robots have been under serious development for nearly two decades now. That might seem like an awfully long time, but it's more or less normal for new, fundamental technologies. I think we're approximately at the cusp of robot development, where the performance/price curve just begins to take off.
The primary limiting factor in making an independent walking robot is the amount of computing power that can be economically applied to the control task. (The secondary limiting factor is the amount of electrical power that can be carried around. I notice that the Sony and Honda models have both shrunk in their latest incarnations, and it's no accident: it takes a lot less power to move a smaller robot around than it does to move a large robot.) Both of these technologies are making dramatic improvements, but especially the computing part. It just takes a lot of computation to do the sorts of things these machines are capable of doing. How much computation? Nobody really knew before now, that's how much. One of the benefits of building robots is that it quantifies the computational requirements of robots. As computers continue to yield more cycles per second per dollar, as memory and power densities improve, you can expect that humanoid walking robots will become more capable and more affordable.
Some people think that we're at the same point in the evolution of personal robots as we were in 1975 regarding personal computers. The market was ready for PCs in 1975, and the technology was relatively weak, but the stage was set for a revolution. Look at your computer now, and compare it to an Apple I, and you might see where robots are going, and how long it'll take to get there.