rethink your priors

Is prompt engineering the most important AI skill?

The short answer

No, and it was never the whole skill. Nobody has the job title "Google search engineer", even though searching well is useful. Prompting is heading the same way. Models keep getting better at working out what you meant, so clever phrasing matters less every year. What matters more is knowing the workflow and the outcome: what the work actually involves, what done looks like, what can go wrong, and how you'd check the result. Someone who understands invoicing will get better results from a model than a prompt expert who doesn't, because they know what to ask for and can tell when the answer is wrong. Learn the work. Write down what good looks like, with two or three real examples of good and bad output. The prompt follows from that.

Outcome first

Want this applied to your actual workflow? That part isn't a cartoon. Work with Eduardo. Longer writing: The Compounding Founder.

More priors