Brief notes on: Homo Deus

Homo Deus is Yuval Harari’s follow up to the popular Sapiens. It begins by discussing the idea that humans now have the capability to solve the three major problems that have beset it throughout history: famine, plague and war. What problems should we attempt to solve next? Longevity, tedium and divinity.

In part two, the book explores religion and meaning. Humans have traded meaning for freedom by means of liberalism. Though this part presents few new ideas about the future, it sets up the final part of the book nicely.

Part three finally discusses the future of humanity. Because we can make algorithms that know us better than we know ourselves, we will promote algorithms from oracles (should I take the Central line or the Piccadilly line?) to agents (tell me the fastest way home) to sovereigns (the algorithm routes all people to the benefit of everyone). We trust them to make an ever increasing amount of our decisions because they can aggregate incomprehensible quantities of data and are not subject to the same biases and fallacies that hamper our own thinking (see Thinking, Fast and Slow). This ultimately gives rise to a new religion, Dataism, in which experiences have no intrinsic value except for the data they provide. Humans’ purpose is then to create the ‘Internet-of-All-Things’ into which they are subsumed.

The book is thought-provoking, though the ending seems rushed. I would have liked to hear a more indepth treatment of the potential alternatives to the dystopian portrayal of Dataism, which does not push the envelope far enough if you are familiar with the topic. Harari frequently cites Kahneman’s peak–end rule, which is just as well: I found the middle too historic for a book I assumed was about the future. I expected a more technical viewpoint (more akin to Bostrom’s Superintelligence), and was disappointed. But this is entirely my fault – the book promised a history of the future and that is exactly what is delivered.