A surprisingly comprehensive introductory text to understanding what machine learning is and more importantly, its limits. As a practitioner and professor within the field one of my main concerns is to demystify what "AI" and specifically machine learning actually is and how it works; how bias can be encoded in the data and statistical methodologies used to create predictive models to catastrophic and life-ruining results, how every other AI startup is just phrenology repackaged, how we're nowhere near close to general-purpose AIs, how often these terms are deployed as nothing more than marketing buzzwords and how in general we can't entrust decision-making to algorithms as much as we'd like to believe we can without human intervention. This book doesn't go as hard as I'd like on those subjects, but it touches on them in really easy to parse and humorous ways and it doesn't shy away from providing more dark examples when appropriate.