Foundations of AI

Before we can explore the vast applications and navigate the significant dangers of AI, we must first build a solid foundation of understanding. This first part of the book is designed to demystify the technology behind the headlines, taking you on a journey from the classical pillars of the field—like search, optimization, and knowledge representation—to the powerful generative models that have captured the world’s attention. Our goal here is to look under the hood to understand not just what these systems do, but how they work.

At the heart of the current revolution are Large Language Models (LLMs), the sophisticated engines powering the generative AI that most of us interact with daily. We will dive into the theory of how these models learn to understand and produce human-like text, tracing their evolution from simple statistical methods to the complex transformer architecture that made them possible. By understanding their core mechanic—predicting the next word in a sequence—we will see how surprising capabilities like translation, summarization, and question-answering emerge.

Finally, we will explore how these raw models are integrated into the digital world to become genuinely useful tools. A powerful model is only the starting point; harnessing its potential requires a set of specific techniques. We will cover the essentials of prompt engineering, the art of crafting instructions that guide the model effectively; context augmentation, which allows us to ground the models in new or private knowledge; and agentic AI, the paradigm that gives models the ability to use external tools and act on our behalf. By mastering these concepts, you will gain the fundamental literacy needed to harness the power of AI responsibly.