How do you use ai today, defines your cognition tomorrow.
I’ve been thinking about the different ways we use AI.
And I don’t think they’re all developing the same capability.
When you do something yourself, you wrestle with the details. You make mistakes. You notice patterns. You build intuition that can’t be downloaded—it has to be earned.
When you prompt AI, you’re still thinking, but at a different level. You’re defining the objective, setting constraints, evaluating trade-offs, and judging the output. That’s a valuable skill too.
But it’s a different one.
The approach I’ve found most useful is a third one:
Think first. Create first. AI reviews last.
Almost every post I publish follows that workflow. The observations, frameworks, and research are mine. Once I’ve finished, ChatGPT helps me improve the flow, tighten the structure, remove repetition, and challenge weak arguments.
In that role, AI becomes less of a writer and more of an editor.
Or perhaps, a reviewer with access to an enormous body of human knowledge.
I’ve found that if an idea survives that review, it usually comes back clearer, stronger, and easier to understand.
The capability I don’t want to lose is the ability to arrive at an original conclusion before asking AI what it thinks.
Because AI can help me communicate better.
It can help me iterate faster.
But if I outsource the thinking itself, I may become a more efficient creator…
…without becoming a better thinker.
As AI becomes more capable, I suspect the real differentiator won’t be who uses it the most.
It will be knowing which parts of your work should remain unmistakably your own.