Hacker News Reader

It seems to me (at the moment) that the 'engineering' of "prompt engineering" happens at the level of taking a word prediction model AI and turning it into a Q&A AI, and the art of decomposing requests into state and behavior that can produce an A from a Q. Also the chaining of prompts, self prompting, etc.

It's functional composition that's interesting. A system of prompts. Not the phrasing of a single question, no matter how clever.

View on HN

Agreed, there’s a lot more design space to explore with prompt chains/flows. But a single prompt requires design too especially if it’s dynamically generated. You have to choose what information to provide, frame the question or task you want it to complete in a way that provides good answers and is also extractable from the completion, and fit within the token limits (32k tokens is a lot but expensive and still small for some tasks).

This article only covers the simplest type of prompting you might do.

Reply on news.ycombinator.com