Chapter 8. Using Your Data as a Differentiator

In the last chapter, we spent some time giving you a point of view on the power (and potential) of small language models (SLMs). We introduced the notion that one model doesn’t have to—and won’t—rule them all. We outlined how humongous models are clunky to operate, expensive, and center power on the few (vendors) that can afford to build them. But, what’s more, they won’t help you take advantage of your data (unless you give it away) to generate value tailored to your business—in short, they help you to be an AI User as opposed to an AI Value Creator. We posit, and will continue to prove, how highly focused models can do some incredible things. We want to see an AI future that is open; hence, we oppose the notion that one super LLM (large language model) should rule them all.

A fundamental premise of this book is the only way for you to become an AI Value Creator is to first see your data as a dormant superpower. To maximize what you can do with AI and create value, we believe big bets must be placed on fostering a collaborative ecosystem across your company that can put your data to work, creating value for you. In fact, we think this notion is so important, it literally became the title of this book: AI Value Creators.

In this chapter, we look at how developers and domain experts in your company can leverage new techniques in model customization to contribute to your company’s Gen AI models, driving defensible and differentiated ...