I've always believed that if something can be expressed clearly as information, out of context it is something a computer could be taught as well. Therefore their capacity for improvement is limited only by our capacity for expressing knowledge. And only for the time being.
The human brain is the most complicated system known to mankind, but all it is, is a chemical computer encased in an organic carbon construct. Yes, we are computers, but we also have the distinction of being self aware, something computers cannot claim.
So what else makes us different from computers?
There are several ways in which we can see if something is "alive" in the sense of the word that we have come to understand:
Movement
The ability to move, as a result of stimulus is a fairly obvious one. Whilst plants don't conform to this in the most obvious way, they can still adapt to their surroundings by growing towards light or nutrients.
So how far have computers come in this category? Quite far. Many modern cars now come complete with sensors and an on-board computer that helps drivers avoid collisions with deer and slow-moving pedestrians, the US army use robotics to sense explosives without risking troops, boats have auto navigation. All of these display movement as a result of external input.
Respiration
A slightly more difficult one. It's not breathing as some people think, though breathing is a part of the process of human respiration. Respiration is simply the process of deriving energy from organic molecules. Now, bio-diesel doesn't really count because it's not usually totally organic. But there are machines that can derive energy from organic sources, you could argue that burning refuse for electricity even counts. However, few systems are truly mobile. Here computers and machines fall short.
Sensitivity
The ability to be aware of the world is something that is really broken down into two parts: seeing and believing. It is irrelevant if a computer "sees" something but does not understand it. And until recently, this was pretty much how it worked. But in the past two years or so, steps forward have been made. By telling the computer the context of an image, it can be trained to identify parts of an image. For example, telling a computer that it is seeing something in the context of tennis enables it to find a tennis ball, the players, the court etc. in an image and outline them. It is still a big step away from truly understanding what it sees, but it would be hard to believe a computer could do this twenty years ago.
Growth
There's a difference between plugging in some extra RAM and a machine actually growing. At first it sounds impossible. But, we are chemical computers, how do we do it? Our DNA contains instructions on how our bodies should change and expand as we age. External influences also play their part, such as food and light.
However, what if a computer was given a room full of spare parts, a mechanical arm, power and a zip file full of instructions on how to build itself an extra circuit board? Our DNA is nothing more than cleverly "zipped" information, telling us to grow in a set way.
Reproduction
And here we come to the big one. If something cannot reproduce is it alive? No. But computers would be silicone based lifeforms if they were ever classed as such, we cannot apply our standards to them. Steven Hawking has already postulated that we have created life with the computer virus, as it is a master of reproduction. What if you were to stroll down to your nearest computer factory? You would see computers building other computers. You'd probably see lots of people in white coats too and the occasional technician guiltily installing Windows Vista. But in effect computers are reproducing. However, this is more like the first strings of amino acids combining to form a cell than true generational reproduction. The machines building computers are no more than parts run by a simple program, not life forms trying to propagate their race.
Excretion
Here computers stumble a little. They don't really produce waste. They consume electricity and generate heat as a by product of their calculations, but they don't drop little USB sticks on the ground or leak coolant when nervous. But once again, I can't help but feel that we might be judging them too much by our standards. Who's to say a silicon-based life form wouldn't be more efficient?
Nutrition
Machines need power to run, but don't generate it themselves. We consume food, and from that generate electricity. Computers skip the middle process. And they don't need nutrients. If they are damaged they are unable physically repair themselves. The closest a computer can do is skip over a damaged section of a disc, barely an allegory for mental scarring.
The computers that look the closest to us, in the form of human-shaped robots (think Honda's ASIMO robot) have the ability to move and react to external stimulus on an impressive scale, but are lacking in other ways. It may be that steering computers into being more similar to us is taking them in the wrong direction; we can't expect a new life form to be the same as the old, at the most basic, even Darwin's theories were based on new and different characteristics setting one species above the other.
Computers still have a long way to go before we can class them as alive and they may never reach the requirements currently set out for classification as a life form. However, we should be open minded, we can't judge something so different by our own standards and in some ways they may even be superior to us.
As science stands, an organism would have to be "alive" before it can be self aware, but it seems that computers may achieve these things in a different order. A vast computer with the ability to learn and change its own programming may be far closer to being alive than the typical type of machines we like to think of as “the next step”.
