Review: Superintelligence: Paths, Dangers, Strategies

Superintelligence: Paths, Dangers, Strategies by Nick Bostrom
My rating: 3 of 5 stars

This book has been a tough read  for me. Because the language used by the author is very esoteric, and this book is filled with so many abstruse concepts such as Orthogonality, Optimization. Recalcitrance etc. I picked up this book because Elon Musk, the legendary entrepreneur and Sam Harris, the neuro scientist have recommended it. Alas, it was a disappointing read. Not that the book was boring or so, but may be because my own knowledge of technology is not that strong. The author has tried his best to explain the most complex concepts in the simplest  way possible. But, the simplicity was not enough. A layman will definitely find it hard to understand this book.

This book does not fall into the league of futurist books such as the Physics of the Future for instance, written by Michio Kaku. Mind you, the author is more a philosopher than a scientist. Moreover,this book is not about when will reach to the point where the AI exceeds the intellectual might of humans.Rather, it is about what will be the fate of humanity if  it happens so. Yes, clearly Nick Bolstrom is thinking ahead of his times. But, this book was too complicated for me to digest.The author makes it clear that he does not want to make use of the word Singularity for the reason that the word does not fully capture the essence of the phenomenon where AI fully overtakes the humans,which has far reaching implications for the mankind.

But yes, there are certain points which I liked in this book, and they were in the first few chapters.
The concept of whole brain emulation was really interesting. He talks about modelling the AI superpowers by simulating the human brain. But that too seems to be a very distant possibility going by the current situation. At present, to the best of my knowledge , with all the advances we have made in the neurotechnology, we hardly know only 5-10% about the human brain. I don’t know how many years it will take for the Connectome project to finish, which aims at mapping the entire 100 billion neurons of the human brain. And, it is not going to be easy. Because it is a cross-disciplinary project.

The second and a very controversial chapter is on biological cognition. Nietzsche proclaimed to the world that he will teach the secret of how to become a superman. That is what the author propounds in this chapter in a way. Biological cognition talks about the enhancing the intellectual capabilities of human through various intentions such as genetic engineering, drugs that would increase one’s cognitive horsepower etc, so that our brains would be better suited for modelling for superintelligence.But such interventions can lead to unwarranted side effects and have unintended consequences.

The author also writes about how brain-computer interfaces will become the norm of the day in the future if at all we achieve superintelligence .In the near future, humans should learn to co-exist with machines and AI enabled devices. This is something which has already started happening.But my objection is that(I may be wrong); generally in the market, only those technologies will thrive that complements the human ability and not replaces the need for human ability. In that way, the concept of existential risk seems to be very apocalyptic to me. Because, on the other hand, we have technology evangelists like Tim o Reilly who say that technology will always complement the human ability and the probability of technology or AI overtaking humans sounds like science fiction to me.












View all my reviews