I’ve always had the misguided tendency to outsource my agency—rely on my friends, family, and the internet instead of making decisions for myself. Once ChatGPT came out, I felt the need to bring it everything: personal problems, work decisions, even “what should I write in this birthday card.” We talk about this problem in education—the kids are cheating!—but it’s become increasingly easy to “cheat” at life too.

I am fortunate to have a fully developed brain while the AI revolution is happening, because I can recognize this problem and fix it for myself. So what’s the fix? Large Language Models are incredibly capable, but prompt engineering matters because it “activates” the pertinent weights in the model. Humans are pretty similar! We have a ton of knowledge, wisdom, and decision-making ability—or at least the neuroplasticity to develop them—and sometimes, all it takes is the right prompt.

Basically, I find that LLMs are very good at prompting me. When I’m facing a big decision, I’ll explicitly ask it for follow-up questions. For example, ChatGPT might ask me “If you looked back five years from now, how would you like to remember this moment?” and that simple question will evoke a newfound insight and clarity on my end. I had tunnel vision, and this prompt activated a different set of neurons. Now I’ll be better poised to take a long-term view on my personal problems in the future.

Of course, sometimes you just want an answer, but consider whether you’re losing out on the opportunity to use and shape your own mind. Ask your LLM to prompt you!