Nor harsh nor grating, though of ample power
Last month, I came across ABEBooks’ Weird Book Room, a delightful collection of amusingly outdated, crazy, niche or otherwise just plain unusual books. I mentioned on Facebook that I’d be delighted to receive any one of those books as a present, a comment which was intended mainly to compliment whomever had put together the list. I was (and remain) amazed and delighted that the list contains not even one book that doesn’t pique my curiosity in some way.
However, some people read into the statement some kind of request for actual gifts, and a few of those people even talked about getting me some of those books for real! I found myself, again, amazed and delighted. Alas, I was to be mostly disappointed, as these things tend to go. However, my wonderful cousin Casey did follow through and gave me a copy of Daniel H. Wilson’s How to Survive a Robot Uprising.
First of all, it’s a beautiful book. The cover’s a heavier-than-usual cardboard with a glossy finish and metallic highlights. The pages are all also gloss heavy-bond paper and are liberally decorated with stylized 4-color illustrations of the book’s concepts. In a nice touch, the pages are all gilt-edged in a red-metallic finish. It’s nice to see such an imaginative book design — it’s clear that someone cared about the project.
Wilson is a roboticist at Carnegie Mellon, and as such knows quite a bit of the state of the art in robot design and capabilities. The main intent of the book, it seemed to me, is to give a broad overview of the state of the art in robotics engineering, paired with humorous comments about what part these technologies will play in the inevitably upcoming robot uprising. I will note that the book is from 2005, and 6 years is a while in any technological field. While most of the stuff he talks about is still relevant, most of his examples of the “next big thing” are today more-or-less commonplace. It doesn’t affect the enjoyment of the book at all, though.
He begins the book by going over the broad different types of robots that might take part in the uprising. He starts with examples of robots based on biological forms: humans, snakes, insects and even lobsters. While this kind of thing is most of what I used to think about when I thought about robots, Wilson presents an array of robot types that I’d never considered before.
He makes compelling arguments for both the smart house and smart cars as possible co-conspirators in the robot revolt. Unmanned planes fly missions every day in the modern military, and those are robots too, as are unmanned submarines and the generally less-successful unmanned automatic boats. He even talks about modular robots, which instead of being a single mechanism are actually a collection of organized sub-mechanisms that work together — like a bunch of minimally-aware, motile LEGO bricks which can assemble themselves into whatever configuration is optimal for the task at hand.
During all of this discussion, Wilson gives tips on how to divert, avoid or destroy robots of different types. Humanoid robots can be tripped; robotic cars can be fooled by thick bushes, which they might not be able to differentiate from walls; modular robots’ communications and linkages can be fouled with oil or sand. He also inserts a good amount of humor: at one point he suggests that a smart house might try to kill you by refusing to cook dinner until you “inspect the oven.”
Next Wilson moves on to a discussion of the different sensors robots might use to help find and kill us all. He defines sensors as devices which “convert a property of the physical world into an electrical symbol.” This includes the obvious stuff like cameras for vision, microphones to hear, and chemical sensors for smell and taste, as well as more esoteric devices such as laser range-finders, GPS systems and thermal imagers. One very interesting example is the hyperspectral camera, which examines a great range of different types of light and is able to see through the outer layers of your skin, allowing robot overlords to identify humans unerringly by the unique patterns of blood vessels in their faces.
In the final section concerning the robots themselves, he talks about machine intelligence: how robots process the data coming in from their sensors. He discusses facial recognition, silhouette recognition and speech recognition and gait recognition, going into the broad strokes of how they work and into more detail on how they can be fooled. For example, gait recognition (recognizing and identifying people by how they walk) can be fooled by wearing a skirt or long coat; talking through a fan will “chop up” your speech enough to fool most artificial speech processors but will still be easily understandable by humans; covering yourself in cool mud (ala Predator) can fool infrared cameras; and so on.
Now that we know what robots are made of, Wilson goes into greater depth about what the robot uprising will look like and how we can survive in both the turbulent early days as well as the oppressed aftermath. The advice is pretty intuitive: first, be prepared; second, be aware and recognize the uprising as soon as (or before) it begins; finally, make your escape quickly and without hesitation.
For preparations, he suggests building an electronics-free “safe room” in your house, and having a good cache of food and survival gear available no matter where you live. Signs of the revolt may include a “sudden lack of interesting in menial labor”, “repetitive ‘stabbing’ motions”, and mysteriously disabled off switches. One escape plan he mentions involves pushing your erstwhile servant into the swimming pool and making your getaway, into the wilderness, on a bicycle.
As before, avoiding detection after your escape is a matter of confounding the relatively simple sensing systems of the robots. Live in the wild to hinder pursuit by machines that are unable to negotiate rough terrain; build your shelter in the side of the hill so that it presents no silhouette and use natural materials with few sharp angles, both of which will be difficult for artificial image processors to recognize. If you want to stay in the urban areas and work to overthrow the mechanical overlords, stay around lots of rubble to confuse sensors and don’t use electronics which will emit fields that can be picked up by the robots.
Finally, he talks about fighting back; using EMP bombs and directed-energy weapons to disrupt the robots’ systems as well as using tradition guns to target weak points in the robots’ structure, such as the joints and the external sensors. In the end, of course, it doesn’t matter; there’s very little we can do about a determined robot overthrow, but it’s fun to think about in any case.
One of my favorite themes of the book is the idea that our advantage over the robots lies in the basic different between our origins. Robots are created things and thus are generally only good at only specific tasks; they’re incredibly simple when compared to the massively complex evolved things which humans are. The arduous path of our evolution has left with us a great number of abilities which are of limited or no use in day-to-day life but can be critically important when they are the only differentiator between ourselves and a murderous machine.
While there are few single things that a machine can’t do as well or better than a human being, there is no single machine which can even come close to matching the variety of abilities possessed by even the most average human being. Science fiction robots might be ultimately superior to humans, but the ones we have today are all purpose-built and designed to perform specific tasks in specific environments. As such, the real key to survival is to recognize which thing it is that the particular robot you’re dealing with can’t do, and use it to your advantage.
A snake robot can climb through small holes and go up any tree, but on a flat-out, a human will outrun it every time. A vengeful smart car can run down a human in that same flat-out without any trouble, but even a 2-foot deep hole will stop one dead in its tracks. The modular robot may seem to be the least vulnerable to this kind of problem, but their design means that the interior of their mechanism is exposed to the outside world every time they have to reconfigure their shape, and you can take that to your advantage just as easily.
So, the moral of the story is, know your robots, know their weaknesses and exploit them as best you can, and you’ll be one of the lucky ones who ends up living in a steampunk rave club at the center of the earth. And thanks again for the book Casey; we’ll have to do another book exchange next year! You still have to let me know how Summer with the Leprechauns turned out.