Book review: Superintelligence

I've read the book Superintelligence: Paths, Dangers, Strategies by Nick Bostrom and it is a book that talks about artificial intelligence. The reason I decided to read the book was that Elon Musk recommended it through his Twitter account. Elon Musk believes that artificial intelligence will be one of the things that in the future will change the world. 

But we begin from the beginning. The very basic idea behind the book Superintelligence is that the world in a near future will change. We are not talking about a small change, like when we invented the car or the aircraft, we are talking about a total transformation. The reason behind the transformation is that we have developed an artificial intelligence that is much smarter than what we are. This transformation point is called singularity. Those of you who have seen the movie series Terminator can relate to this when I say that this superintelligence is Skynet.

In the Terminator movies, Skynet destroys larger parts of the world. There's a large probability that the superintelligence we develop in a near future will also destroy the world. For example, let's say we have a superintelligent machine and we tell it to stop e-mail spam. Which is the easiest way to stop e-mail spam? The answer is to kill all humans, because we humans are the sole reason behind e-mail spam.

So how can we control the superintelligence in a way so it can only do good? The answer is that no-one really knows because this superintelligence will be so smart. You can compare the difference between a human and a dog with the difference between a human and superintelligence. We humans will understand concepts the dog will never understand no matter how hard the dog is trying, like writing. This superintelligence will also understand topics we humans will never be able to understand no matter how hard we are trying. Yes, we could try to insert this superintelligence in a computer and then unplug it from the Internet. But what if the superintelligence find a way to communicate with our nuclear weapons though the electric cable?

You may think that this superintelligence will come to life years from now. But some scientists argue that it may happen 10 years from now, even though most of the scientists argue that it will happen in 2040. Hopefully, we have by then found a way to control the superintelligence, because it will make our lives much better. It will probably find a way to cure cancer, it will find out how to produce cleaner energy in abundance, and we will probably never have to work again because intelligent robots can do whatever we are doing for free. I don't think we will need money anymore because robots take care of everything, hopefully without needing humans as slaves.

Back to the book. About 80 percent of the book Superintelligence is about: What will happen after we have developed a superintelligence. The author can only speculate because no one will know for sure. This speculation gets a bit boring after a while and I guess the author could have talked a little less about it. And I also think that you have to be interested in artificial intelligence and singularity to be able to enjoy the book.