I keep digging into the rabbit hole that is artificial intelligence and machine learning - and it turns out the books on this field (when they aren't just spreadsheets and code) are extremely insightful to the understanding of machine learning systems. Also, all the books explain the problem with the term "Intelligence" which by definition cannot be programmed, but must be emergent.
This book takes a casual and pop approach to explaining the subject, and mentions several interesting stories about machine learning in various applications. It skilfully avoids the subjects of political elections, but a highly important examination of the IBM Watson system when faced with the TV show "Jeopardy" is presented. This was my favourite part of the book.
The latter part of the book takes on predictive sales analytics by focusing on making the important measurable, as opposed to making the measurable important. The problem is that statistical ex-postfacto approximations are the best that can be achieved, and that other gains in predictive analytics can only be made by measuring more and more relevant aspects of the sales process.
Naturally, the book cannot tell us what it does not itself know, and the very end of the book leaves one blue-balled in a way. In the future, I'll look into case studies of predictive analytics in large companies to help understand how sales can be advanced through this system. As for the book, I recommend it to anybody new to the field.
9/10