4.7. Super-intelligence (machines surpassing human intelligence)
First of all, it is funny that some believers of the theory of evolution believe that super-intelligence can evolve in machines:
• while all machines are (intelligently) designed;
• while both intelligence and design originate from the bible;
• while the theory of evolution has no satisfying explanation on the origin of intelligence;
• while scientists fail to define intelligence in a natural way (as a set of natural laws);
• and while some evolutionists even deny the existence of intelligence.
Besides that, we really need to distinguish “machines surpassing human capabilities in a limited domain” from “machines surpassing human intelligence”.
Humans are able – to a certain extent – to play chess, to play the game Go, to move an amount of sand, to contribute to business meetings, to cook diner, to clean the house, to babysit, to tell bedtime stories, and so on.
Super-intelligence requires a machine to surpass humans in all human capabilities. Otherwise, it is only surpassing humans in a limited domain.
First, let's consider a few systems that surpass human capabilities in only one limited domain:
• Deep Blue was able to beat humans in playing chess, but not in playing the game Go;
• AlphaGo is able to beat humans in playing the game Go, but not in playing chess;
• Bulldozers are able to beat humans in moving an amount of sand. But they are unable to play chess, unable to play Go, unable to cook diner, unable to babysit, and so on.
Now, let's assume we want to build a machine that surpasses humans in a more than one domain. Let's consider to integrate a chess computer with a bulldozer. In this way, we will get a chess-playing bulldozer, or a sand-moving chess computer, that surpasses humans in both playing chess and in moving an amount of sand. It is possible. But it has not been done yet, because such a machine – integrating the capabilities of different domains – is very unpractical.
Even if we try to integrate systems of domains that are less different – like a car and an airplane – we will eventually discover that the design of the integrated vehicle is weaker than the individual designs of the car and the airplane. So, it is possible to make a road-legal airplane, or a car able to fly. But the integrated vehicle will always be weaker than the individual designs of the car and the airplane.
Apparently, we are limited to design systems that surpass human capabilities in a limited domain. It proves to me that we have a divine origin.