Most of how I use AI probably won’t surprise you: exploring ideas, researching questions, drafting blog posts, writing code, and updating my website. Increasingly, I also let it carry out tasks across my computer.
The expertise is in knowing what to delegate and how. Is the outcome clear, or would breaking the work into smaller pieces help? How quickly and reliably can I check whether it did the job correctly?
The best opportunities are often tasks where doing the work takes substantially more effort than checking it.
But I also think it’s worth being ambitious. Sometimes I give AI a task just to see how far it gets. It can surprise me. Where its approach differs from what I expected, I learn about missing context, ambiguity, or decisions I haven’t thought through.
When it keeps falling short, I sometimes work through the task myself. That helps me decide whether to explain it better, break it down differently, or recognize that the tool can’t reliably handle it yet.
My clearest boundary is communication. Before work from my business reaches another person, I review and approve it. I want it to be accurate and express what I mean. I don’t want to waste someone else’s time.