Exobiology has been criticized as a subject when those subject matter. We don't know of life beyond Earth. While that may mean there's not much to talk about, there's room for a lot of speculation. One of the areas of speculation involves looking at life in its most general principals. Talking about life as computation. This is not a crazy thing to do. Francis Crick, one of the co-discoverers of DNA, was asked by a New York Times journalists at Cold Spring Harbor where he was at a conference in the 1950s, what was in a sound byte the meaning of the discovery of the double helix? He thought for a minute and said that life is digital information. Remember this was very early in the years of computing. Unusual thing for a wet biologists to say. But what he meant was that the fundamental attribute of life was this ability to store information in the genetic code and transmit that information. If that's the general property of biology that really matters, even to a biologist, it's worth asking the more general question, is life or wet biology the kind we're familiar with, the only way of carrying out the attributes of life? So we can ask, how surprised we would be if life didn't need a star? If life didn't need a planet? If life didn't need cells? If it didn't need carbon or water, or it existed in a completely different computational basis? The interdisciplinary subject called artificial life is collaborative between biologists, computer scientists, physicists, and even philosophers. It's seeking to ask a deep and profound question about biology, of what the general case might be. If biochemical mechanisms involving DNA in our particular genetic code are the only way to get the attributes of what we call life? Artificial life recognizes that one of the basic premises of biology is the ability to develop complexity from simpler components. It speculates that biochemical networks may not be the only way to develop that complexity and store information. Another aspect of artificial life involves what's called synthetic or computational biology, where people use biology as a basis, but then do computational or computer modeling to see how different those attributes might be in different settings, with different ingredients, different physical conditions, even a different conceptual framework. We've already talked about the patterns that are the basis of life, strikingly different from randomness and chaos, patterns on multiscale involving the individual organisms, or the large-scale organizations of those organisms. As a computational experiment in developing complexity from simple rules, the field of cellular automata has existed for several decades amongst computer scientists. The concept is very simple. You have a series of cells which can be either black or white, and the cells propagate vertically or linearly in one dimension growing, with the cells going on or off, white or black according to very simple rule. It's amazing that with very simple rules, cellular automata grow and develop extraordinary complex patterns. Some of these patterns even end up being unpredictable or chaotic. This is an example of how in a very simple and artificial situation you can generate complexity with very simple rules. In the early 1970s, a slightly more sophisticated version of this game was developed called, "The Game of Life" out of the University of Cambridge. There was a time in the mid-1970s when administrators at the university realized in horror that roughly half the computing resources of the entire university were being consumed by a set of a few dozen people playing this game on there as yet rudimentary computers. The Game of Life takes the cellular automaton notion and extends it to two dimensions. Very simple rules cause the propagation of on-off cells into two-dimensional objects which had the ability to move in the computational space. The rules are incredibly simple, yet the organisms for want of a better world that develop extraordinary complex. Some have repetitive behaviors, but some continue to grow in complexity, in characteristics, in an unpredictable way. It's just another example of how simplicity can lead to extreme complexity with the application of simple rules. People who've studied these machines because that's what they are computational machines like Steppenwolf have decided they have lessons for us about biology or the possibility of information storage and something that we call life. With cellular automata, these extraordinarily simple systems, you can have randomness and yet predictable large-scale behavior. You can use these to generate the answers to mathematics problems like differential equations. You can even invent what's called a Turing machine, a universal computer. This idea was actually co-opted by SETI because one of the ways to send digital signals into space is to send a bitstream that corresponds to a Turing machine, which would tell an intelligent alien that we know how to compute anything. These machines have also been able to show that the mathematics we have in our textbooks is a subset of all possible mathematics. Although the cellular automata may seem like an artificial construct involving cells that turn off, black and white, and propagate in one dimension, it's all true the complexity of this world that develops, if there's no grid, if there are bounds, or constraints rather than fixed rules, and it can operate in more than one dimension, two, three as many as you like. It's a very flexible mathematical construct. The hypothesis here is that biological evolution is just one example of a wide range of computationally equivalent possibilities. None of them are deterministic. All of them generate extreme levels of complexity from simple rules, as biology does. Recursion or feedback is a necessary part of these schemas. Another way of approaching the idea of artificial life is to recognize that we live in a civilization and at a time or computation and technology is ascendant. It's reasonable to ask as science fiction writers and futurists have done, whether biological evolution is just one stage of our evolution, and we will eventually move to a more machine-like stage. This need not be by the takeover of biological organisms by machines, it could be accomplished by the merger of biology computation and machines. Although some of this sounds outlandish and speculative, just think how far we've come in just half a century. A picture of ENIAC, one of the world's first computers, shows a situation where it was the size of a small house used the power, sufficient to power the entire city of Philadelphia, and needed five full-time technicians to replace the valves or tubes that made it work in this age before the transistor. Yet this computer baby steps in the computational development was substantially less powerful than the average smartphone. Nobody foresaw where computers would go at the dawn of the computer age. Some famous statements by people who even lead computer companies will illustrate this. What we're trying to see is into the near and far future. It's amazing how quickly the crystal ball gets cloudy. Do the thought experiment of thinking back in orders of magnitude in time, 10 years then a 100, then a 1000. In round numbers, 10 years ago there was no Internet. A 100 years ago there was no rapid transit. Most people lived and died within 10 or 15 miles of where they were born. A 1000 years ago, no medicine. Life was incredibly primitive. Ten thousand years ago, no agriculture. We were just hunter-gatherers. A hundred thousand years ago, we were finally modern humans, and we became human a million years ago. Now step forward in orders of magnitude and see how quickly it becomes uncertain where we're headed. In 10 years, we might speculate that genetic engineering will mature, we'll be able to alter the human organism at will. A hundred years, perhaps we'll have cybernetic organisms, full mergers of man and machine. A few hundred or maybe 1000 years from now, I'll guess that we'll be traveling to the stars finally after dreaming about it for so long. But beyond that, 10,000 to a 100,000 or million years from now, if humans last that long, I have no idea what our capabilities will be. Yet logically, if extraterrestrial civilizations are more than extremely rare, such civilizations must exist. Those would have been durable for a million years or more and have developed capabilities we can only dream of. Our technology has evolved computationally in ways that have touched everyone's lives, but it's also important to realize that our mechanical technologies have also improved. Robotics is at a state undreamt of 20 or 30 years ago. Moreover, computers allied with robots create machines that can do amazingly complex things and even begin to learn. Hans Moravec is a computer science and roboticist at the Carnegie Mellon Institute. He's produced a graph that shows the exponential growth in our computational capability. Not only in terms of the power, but in terms of the power per dollar. So our ability to harness more and more computational power. He's particularly interested in how this power manifests in a mechanical package or robot. He himself has built robots for 30 or 40 years with ever-increasing capabilities. To have robots that simulate human sensory apparatus is impossible. But we can still learn a lot about how robots may function in the future by the rapidly developing capabilities now. The most interesting thing he does is compare the fully equivalent mechanical and computational capabilities of a robot to the organism in life in biology. In Moravec's estimation, our robotic and computational technology has progressed to the level where we can create the capability of a lizard or salamander in a small package. That is full sensory capability, ability to function in an environment, process information, and perhaps survive. But if we look at the exponential progression, we see that in 50, maybe even 30 years, we're projected to the capabilities of the human being in mechanical form. We will really get robots as smart as us in 50 years? Who knows. But even if this extrapolation is a little too bold, what's striking, is that in less than a century of computation and robotic experimentation, we've done with robots what took life on earth three billion years to go from the level of a bacterium to the level of a lizard. That's the incredible acceleration offered by our technology. The full extrapolation of this exponential progress is, of course, manifested in science fiction. Blade Runner is perhaps the most famous example of a near future where cyborgs, cybernetic organisms, perfect hybrids of humans and machines exist to do the work that we can't or don't want to do, and of the inevitable and ethical dilemmas that arise when they had sentience, and intelligence, and feelings, and their own wills and desires that may not agree with ours. Even biologists admit that some aspect of biology is pure computation. So it's natural to generalize this idea and wonder what the capabilities of computation might be or might turn into. Our exponential improvement in computation is also mirrored by advances in robotic design, such that we can now simulate the capabilities of modest creatures like perhaps a rat or a lizard. Extrapolation of this progression suggests we'll create human levels of intelligence, merged man-machine hybrids or cyborgs in less than a century, and beyond that, who knows what's possible?
