Flesh and Machines: How Robots Will Change Us [Paperback]

While inconsistent, there are points in the book that are quite satisfying. After a slow start tracking through ancient history, once Brooks begins telling his own stories and those of his contemporaries, he catches his stride and is captivating. It was late at night, but I couldn't put the book down as he described his laboratory's robots from Allen through Cog and the delightful Kismet (and Cynthia Breazeal! Never miss an opportunity to hear her speak, she can compress ten hours worth of speech into an hour and make it utterly digestible and entertaining.) Brooks lays out his insights regarding his design choices in clear and polished prose, and summarizes a variety of the motivating research without losing the reader in details. Would there were more, though, and more regarding the work of other researchers in robotics. This probably should have been subtitled _How MIT AI Lab Robots Will Change Us_.

There is enough in the book - say between pages 16 and 147 for me to justify the purchase, but after that point, I think it went downhill fast. My views regarding religion are very close to Brooks', but I still found his steamroll through the flower-fields of the almighty rather dull and repetitive. It's a lot like the five-cent tour given by every other pop-sci religion-drubber in the past half-century, and really, it's kind of tired now. If you're a pop-sci writer and you feel obligated to go over it again, please bring up a couple of new arguments or at least an invigorating take on an old one. Maybe you're not preaching to the choir, but we're the ones buying these books.

On the whole, I suppose Brooks did what he set out to do with this book, but I found the poorer writing in Moravec's _Robot_ and Kurzweil's _The Age of Spiritual Machines_ more interesting futurism, Menzel and D'aluisio's _Robo Sapiens_ more interesting and well rounded regarding robotics, and George Dyson's _Darwin Among the Machines_ more thought-provoking history. I can't help but think that Brooks could have written a much deeper treatment of his own research while leaving out his naïve and mild philosophical ramblings and produced a much stronger book. Maybe next time.

As a reader only casually versed in the science and history of robotics, I found the book informative and approachable. The first third of the book held my interest best. In this part, he recounts the early history of robotics with particular focus on a simple robot built in the 1940s nicknamed the tortoise, which combined simple electronics and sensors to create a machine with complex behavior. Brooks then goes on to use the ideas embodied in the tortoise to turn the modern world of robotics on its head. From 1950's though the eighties, robot developers tried to build robots that developed detailed world models, and thus could navigate through them with ease. That was the theory, but it did not work. Robots spent so much time building up these models that they moved slowly and gracelessly. After years of working on robot vision, Brooks wondered what would happen if a robot did not even try to create a mental model of it's environment. What if sensors linked to simple actions, a la tortoise? And what if the actions were guided by simple instructions, layered on top of each other, much the way evolution probably layered behaviors on top of each other? The results were surprisingly agile, frisky, insect shaped robots. I got a little lost with his technical description of how these robots worked, but I got most of it, and best of all I got a good understanding of his creative process. I found this first third of the book the most engaging.

.After that he bounces around between various topics, from his studies of visual perception, to Kismet (a humanoid robot designed to respond to physical and vocal cues), to his adventures in the toy industry. By the time I got to his description of household robots of the future, I was snoozing. Gadget freaks may have a different reaction.

In the final third of the book, he weighs in on the possibility of truly intelligent human made machines. While he offers little hope for people who want to cling to our specialness as human beings, he is cautious about the prognostications of futurists who think we will download our midns into machines in the near future. Brooks says there are a lot of hurdles to jump before we create emotional, conscious machines, or before we are able to port our selves into robots. and we might not have it in us to jump those hurdles ever. But in the meantime, he asserts that we will, through machine implantation and augmentation, and through bioengineering, merge with our technology to the point that we will become robot-people, so that if the machines ever catch up with us, they will find we are already them. All this is put forth in a calm, thoughtful, carefully weighed manner, which made me trust him more than the more entertaining, but frothier, Raymond Kurzweil.

I would recommend the book to a wide audience as long as they are prepared to skip around. There is something for most intelligent, curious people here: a portrait of a brilliant scientist, the basics of robotics, and a vision of the future. And for people who care about vacuum cleaner robots, that is there too. I just skimmed that part.