Foreword

I am super-excited to write the foreword for this ground-breaking new book combining two massive industry themes: Generative AI and C++ programming. Having spent almost forty years with enterprise technologies I have been fortunate to participate in the evolution of machine-based intelligence and its adoption by organizations around the world. Enterprise software tools including ERP, CRM and IT management systems enabled organizations to collect and control their data to gain valuable insights. Machine learning and predictive AI allowed organizations to use this data to predict the future and make more intelligent decisions — from setting dynamic alerts in IT management tools like PATROL to providing intelligent forecasting and sales recommendations in tools like DYNAMICS. Generative AI takes the intelligence to a much broader audience, making it available to everyone in an organization and potentially beyond, through a native language interaction.

Lately, I have been amazed to watch Generative AI explode into consumer consciousness with breakthroughs such as ChatGPT and Microsoft Copilot. It seems like there are new capabilities being added to these technologies literally every day. Personally, I feel that Generative AI is making me more productive with searches giving me desired information instead of raw data; and emails, notes and reports coming together much quicker.

I have known David Spuler, often known as “Dr. Dave,” since the early nineties when we worked together to build the most advanced applications and systems management product of that era — PATROL. He wrote an immense amount of C++ code for the product with its revolutionary intelligent autonomous architecture. Even back then, Dr. Dave had already written three C++ programming textbooks. I cannot think of anyone better to write about a complicated new technology like Generative AI.

Generative AI in C++ is not only the most advanced book on Transformers and LLMs with C++, but to my knowledge, it may be the first such book. Although the material is inherently complex targeting true experts, the book introduces the topic well with a general overview section about Generative AI in business, with coverage of models such as OpenAI's GPT and Meta's Llama, along with other open-source options. The bulk of the book is about programming of the internal Transformer components in C++, and then optimizing them. Finally, it goes far beyond any other Generative AI book, with the last section summarizing numerous AI research papers in a full literature review.

I wholeheartedly recommend this book to anyone with an interest in Generative AI. The first part of the book is an insightful overview of the business issues of using LLMs, including AI PCs and AI Phones. DevOps engineers will benefit from its coverage of optimizing architectures and deployment of AI engines. This book is a must read for any C++ programmer who wants to learn Generative AI technology. It offers an immense amount of material from parallel C++ coding of CPU and GPU acceleration to AI software development practices to the low-level C++ coding of Transformer engines.

This book will certainly serve as a major contribution to the Generative AI industry. I am sure readers will enjoy it and I wish Dr. Dave the best of success!

— Kirill Tatarinov, March 2024.