My thoughts: Superintelligence: Paths, Dangers, Strategies by Nick Bostrom
This is a pilot post about my thoughts after reading the book. I'm starting with a very good title about artificial intelligence or an unfinished fable of the sparrows.
I am writing these thoughts mainly for myself. These are my private opinions on this subject and you may disagree with them. Nevertheless, since you've come here, I hope you'll get something out of it for yourself.
Unfinished fable of the sparrows
The book begins with a short tale about sparrows who are planning to find and train owls to help them with the construction of nests and generally advice in everything.
Only Scronkfinkle, was not convinced if it was a good idea. Eventually, no sparrow ever controlled the owl.
The story has no end. The old sparrow stays in place and tries to solve the control problem with the rest (you can read complete store here).
Artificial intelligence
We can learn from the book that there are at least several paths that can lead us to the AI. When we can expect it (currently hard to tell). Why the problem of control is so important.
One selected example: we could order AI to make us happy. The result could be implantation of a special arrangement in our brain that would cause a constant return in the center of happiness which would certainly make us happy, but probably not quite the point.
But it's only half.
In further parts, the problems of power limitation, the transfer of will and the ways of inoculating values are discussed. The book ends with a discussion of global strategies and a message to everyone to be on guard.
The number of footnotes in the book makes a big impression. You can see that the author has deeply explored the topic and does not take thoughts from space.
It can be said that it is a kind of compendium of knowledge about the power and current problems that AI can bring to the world.
I have such an impression that the amount of knowledge was so huge that some of them certainly passed me somewhere.
Simulation hypothesis
Nick Bostrom is the author of the modern version of the simulation hypothesis. It says that the reality we perceive is just a simulation we are not aware of.
Elon Musk summed it up once in a very nice way: About forty years ago, Pong was created. Two lines and a dot between them. Today we have photorealistic, three-dimensional simulations in which at the same time play millions of people.
Every year this technology looks better and better. With such a pace of development of the game will become indistinguishable from the reality, and thus the probability that we now live in some kind of simulation is enormous.
Conclusions
Theoretically, we should strive to create artificial intelligence with all our forces. No other technology has such a positive economic return.
It will be able (thanks to a huge amount of resources) to solve our problems much better and faster. For example, it creates von Neumann probe that will facilitate the exploration of the space.
In my opinion, a lot of time will pass, so our industry will not be threatened. On the other hand, even if created faster, the work of programmers will definitely be needed (even for debugging, though it may be debugging itself 🤔)
Perhaps we should strive to create a distributed artificial intelligence? In such a way that everyone has power and control? Maybe a blockchain solution that will provide value protection?
Anyway, the most important thing is spreading knowledge and awareness of the inevitable coming.
Do you have a different opinion about this? Start a discussion in a comment or write to me.