Include guidance for working with LLMs to generate code

[alexander-newton]

Is your content or feature request related to a problem? Please describe.

There is lack of guidance for how analysts in the CS should exploit LLMs. The value proposition of using LLMs in combination with rigorous quality assurance is phenomenal - RAPpers are particularly well set up to exploit the new technology.

Describe the solution you'd like

A full chapter on how to use LLMs well, how to make sure the code they generate is useful and safe, and how to guard against common risks (e.g. prompt attacks, exfiltration risks, etc.)

(p.s. I'm willing to work on contract to write this - say hi to those who remain from me)