if you’ve ever been intrigued (but perhaps intimidated) by artificial intelligence, machine learning, or other similar buzzwords, read this book. hannah fry breaks down the technical details brilliantly, all while carefully exploring the grey area that emerges when man meets machine.

one of the main take homes for me is about managing expectations; an algorithm doesn’t have to be perfect, it just had to be on the whole better than the current state or the art (be it another algorithm, a purely human-guided system, or a human-machine hybrid). we seem to judge algorithms much more harshly even though they may objectively lead to better outcomes than what humans can do on their own.

often the best result occurs when we leverage the advantages of each component: algorithms and humans. AI-assisted diagnoses by pathologists seem to be the best example of this in the book, but i think we’re poised to see more such success stories in the future.