Agile AI

Chapter 2. Understanding AI Tools

This chapter is about how we can understand the technology behind AI. We spend much of our time with clients talking about the differences between terms that are often used interchangeably. That’s an indication that it is valuable to differentiate here what they mean.

We also take the time to talk about how open source projects are critical to the tools and fundamental principles behind AI, and then talk about the trends and risks of machine learning. Both practitioners and leaders in the AI space need to be aware of the effects of models.

Contrasting Machine Learning and AI

Before we discuss machine learning and AI, it’s important to clarify what each means and how to talk about them. It’s essential to do because we’re currently at the peak of the hype cycle for AI, and as a consequence a lot of companies are selling snake oil branded as “AI.” When a term is as overused as AI, people will continue to try to define it so that they can abuse it. So, first, let’s define what these terms and practices mean for technical leaders.

Machine learning is an engineering discipline: it’s the tooling and techniques involved here. At its core, machine learning is categorized as a form of mathematical optimization—because the algorithms are used to perform optimizations to create “learners” from training data. Keep in mind that algorithms are much less valuable than data. In the article “Datasets Over Algorithms”, Alexander Wissner-Gross showed that the mean ...