Month: May 2014

The future of war

Although drones are already being used on a daily basis in war zones (in some countries the noise of their rotors is a daily experience), they still are piloted by humans. Drones allow technological advanced countries to move their armies within minutes to locations all around the world, an important point in the upcoming oil wars, that this author is predicting for the near future too. It also allows more conflicts than the people of those nations usually tolerate, as own casualty numbers can be kept low.

Google already has an AI that can drive a car through normal traffic. A task that actually is more complex than flying a drone in a war zone. We have algorithms that can recognize human faces and voice. Soon robotics and artificial intelligence will be capable enough for those drones to make decisions no longer controlled by humans. This will give them several seconds of advantage over their foe. An advantage, that will become necessary, when drones fight other drones. Will we allow them to decide about taking a human life?

This blog is about trying to predict the near future. With near we mean 25 years, a quarter of a century. Trying to predict more is bound to be too far off to be of any practical meaning. So, although we believe that drones eventually will be allowed to make decisions of ending a human life (simply because the past has taught us that what is technically possible and will change the balance of power, will eventually happen), I hope and believe that this will be not within the next quarter of a century.

Humans still will have to approve an engagement command, very similar to human pilots drones will have to wait for this command. But we will probably soon have drones that given this command will be able to decide how to aim and to pull the trigger. The next generation might be able to seek and destroy without human intervention, staying within a specified radius of course. Will these drones be able to distinguish children from adults? Probably not so soon. Will they be able to detect fear or a threatening situation, distinguish an ape from an old woman or a Shepard? Probably not, but pilots looking on a screen through a camera probably also not. This engagement command will at first not be allowed in their own country of course, but outside, to prevent terrorism? It will happen, because it still will feel better than sending your son to war.

The Future of Computers

Thirty years we got used to the fact that computers are getting faster every year. The CPU in your mobile phone is probably thousands of time faster than the computer on board the Apollo missions. But we already reached a boundary in computer manufacturing that has forced the whole hardware and software industry to change and that is making it more and more difficult to enhance computer power every year. Computers nowadays have the same frequency as the models of last year and the year before. Ten years ago we still could compare the frequency of CPUs to tell which is faster, this is no longer possible. Raising the frequency of the CPU raises the power consumption and heat production over linear. This was the easiest way to raise the performance in the past decades, but the industry had to find other ways. That was the time multi code CPUs where starting to emerge. Nowadays some mobile devices have 4 or even 8 cores. Cores are nothing else but the inner part of a CPU, so instead of having one of those in a CPU, the industry started to integrate 2 or more of these cores into one CPU. Two cores should theoretically have double the performance of one. Unfortunately this is not true. Software traditionally is written to make use of one core. Adding more cores to the CPU will not speed up this kind of software. You need specially written software that makes use of these additional cores.

But we will hit another performance wall soon and industry analysts already can see this wall as the industry is working hard to keep it at bay. The miniaturization of electronic gate structures onto the silicon chip is becoming more and more difficult with every generation. The cost for a new fab is becoming too expensive for one company very soon.

So, we will have to get used to computers not getting faster every year. The industry will try to compensate, they will try to sell us other features like lower power consumption, but eventually computers will no longer dramatically change in performance. These revolutions in the virtual world that we came to get used to will no longer happen, or become very rare. Computers will still be cheep and they still will be everywhere more than now, but new inventions and breakthroughs will slow down to the pace of other industries.

Humanity will adapt. We will pic computers for their design instead of their performance. Come to think of it, we are already doing that. Finally the computer industry will become an industry as all the others and the enormous human labor we invest into it will be reduced to that of other industries.