Название: The AI Pocket Book
Автор: Emmanuel Maggiori
Издательство: Manning Publications
Год: 2025
Страниц: 200
Язык: английский
Формат: pdf, epub (true), mobi
Размер: 10.1 MB

If you’re worried about your tech career going obsolete in a world of super-powered AI, never fear. The AI Pocket Book crams everything engineers need to know about AI into one short volume you can fit into your pocket. You’ll build a better understanding of AI (and its limitations), learn how to use it more effectively, and future-proof your job against its advancement.

In The AI Pocket Book you’ll find no-nonsense advice on:
Deciphering AI jargon (there’s lots of it!)
Where AI fits within your field of engineering
Why AI hallucinates—and what to do about it
What to do when AI comes for your job
The dark side of AI—copyright, snake oil, and replacing humans
Balancing skepticism with unrealistic expectations

The AI Pocket Book gives you Emmanuel Maggiori’s unvarnished and opinionated take on where AI can be useful, and where it still kind of sucks. Whatever your tech field, this short-and-sweet guide delivers the facts and techniques you’ll need in the workplace of the present.

About the Technology:
You don’t have to know everything about AI to get a big payoff! Whether you’re looking to boost your coding speed, generate ideas for your next project, or just get a helping hand with your next Medium article, there’s an AI-powered tool ready to assist. This fit-in-your pocket guide tells you everything you need to surf the AI wave instead of drowning in it.

About the Book:
The AI Pocket Book takes a peek inside the AI black box and gives you just enough on key topics like transformers, hallucinations, and the modern ecosystem of AI models and tools. You’ll get handy techniques to select AI tools, learn when putting AI first is the smart move, and pick up some excellent tips for managing the inevitable, potentially expensive, screw ups.

The first part of this chapter is a high-level explanation of how large language models (LLMs) such as ChatGPT work. Its sections are ordered to roughly mimic how LLMs themselves turn inputs into outputs one step at a time. The code is moderately easy to follow if you understand some Python coding and the TensorFlow library and start from the bottom of the file. I also recommend you read a guide called The Illustrated Transformer to learn the details of the architecture.

The middle part of this chapter discusses machine learning, which is the technique that makes computers learn from data to create LLMs and other types of AI. Note that AI and machine learning don’t mean the same. AI is a research field that tries to create computer programs to perform tasks in a way similar to humans. Machine learning may or may not be used for that goal. However, machine learning has been the preferred methodology in AI for at least two decades. So, you might hear people use the terms AI and machine learning interchangeably. When I speak of AI in this book, I mean current AI methods, and these methods involve the use of machine learning.

The last third of this chapter discusses how AI works outside language generation. Specifically, I give an overview of how AI analyzes and generates images or combinations of text and images. We also comment on current developments in AI-based video generation.

What's Inside:
Deciphering AI jargon (there’s lots of it!)
Evaluating AI tools
Why AI hallucinates and what to do about it
How and when to use AI

Who should read this book?
Two main groups of people should read this book. The first one is software engineers—aspiring, novice, and seasoned ones—who want to understand the effects of AI on their careers and prepare for it. The second group includes people related to or interested in the software industry, even if they’re not engineers themselves. For example, these are product managers and startup entrepreneurs. One of this book’s reviewers said he thought the book would be useful not just for software engineers but also for “software sympathizers,” which I thought was a good way to put it.