preface
When I first started exploring machine learning, the idea that a computer could learn the algorithms I needed instead of painstakingly coding them felt magical. Leaning into my laziness, I loved that I could define a model and let the system discover the internals for me. I had no idea then how far things would go. Today, models can even code themselves, and we’re limited more by the clarity of our prompts than by the machinery itself. The pace of progress has been astounding, and part of what I want to do in this book is peel back that sense of mystery.
My first experience with AI was back in college. In 2016, I purchased a few technical books, much like this one, on machine learning and deep learning. At the time, I tried learning Scikit-learn and TensorFlow but found the learning curve quite steep. Whether due to the libraries themselves or my own inexperience, I struggled to get past the basics, barely managing to build simple models and often feeling stuck. Then PyTorch was released on January 18, 2017, and it immediately struck me as different. It was the first framework that hit the right balance between ease of use and power.