I have been a fan of language learning recently, and today I think I added a new one; prompt engineering. Not fluent, by any means, but at least I have started.
Prompt engineering is a lot of things, but at its most basic, it’s talking to LLMs in a way that maximises efficiency. Think of it like direct instructions; instead of asking ChatGPT to give you a recipe, you ask it to give it you step by step, using Celsius, and in a way that someone who is inexperienced in cooking might best understand. In both scenarios you get the recipe, but one is clearly better.
The proof will be in the metaphorical pudding tomorrow, when I present my findings to our AI squad. Fingers crossed.