Much prompt engineering advice online lacks concrete evidence. The reason: no-one really knows how LLMs (Large Language Models like ChatGPT) work, nor how what they are capable to do. The best you can do is to discover things yourself.

The key to mastering ChatGPT is to consciously use it to enhance your life.

Thinking Outside of the Box: Solving Real-World Problems

It's easy to get stuck into thinking ChatGPT can only help with a few tasks. This often comes from not thinking outside of the box.

Finding great ways to apply ChatGPT to a problem often requires creative thinking. Often, ChatGPT can help not by directly solving problems, but automating necessary but tedious tasks. This way, you can focus on what you enjoy and are uniquely good at.

For instance, ChatGPT can assist musicians not by writing music, but by organizing music sheets, noting recording tasks, or creating tour tech riders. This leaves all the more time to actually make music.

Not so Obvious: Remembering to Use ChatGPT

Too often, I forget to use ChatGPT to solve the problem in front of me.

Remembering to use LLMs requires practice. It's common to overlook that a problem can be solved with an LLM: as a software engineer, it seems obvious that I should engineer a solid prompt to help me review code, yet it took me 10 months to think of it!

Make it a habit to think "could I use ChatGPT for X?" and to write down all the ideas.

Thinking About an Idea Takes More Effort than Just Trying it Out

With ChatGPT, trying something out often takes less than discussing it.

When you catch yourself talking (joking?) to colleagues about something ChatGPT could do, grab your phone, open your laptop and just try it out. Successes will add to your prompt arsenal, and failures will provide valuable insight, all written up in a transcript.

ChatGPT is not a one-trick pony but a versatile tool that can be applied to most tasks and questions; the hard part is remembering to use it!