Automating the boring parts with a little AI

I’ve been using language models for small automation tasks for about a year. Some experiments stuck; most didn’t.

Where it helps

• turning messy notes into a checklist
• drafting a first version of a script I then rewrite
• explaining an unfamiliar log format

Where it doesn’t

Anything that must be exactly right every single time is still a plain script, versioned and tested.

Rule of thumb

Use the model to start things. Use code to finish them.