The most common version of AI-assisted work I see looks like this: person has a task, asks an AI to do the task, pastes the output, moves on. The AI did the work. The person did the logistics.
That’s not what I’m trying to do.
What I use it for
I use AI for things that are well-defined and low-stakes - first drafts, reformatting data, generating options I can react to, explaining things I don’t understand yet. Tasks where the output is a starting point, not an endpoint.
The decisions, the strategy, the judgment calls - those stay with me. Not because I’m precious about it, but because that’s where the actual value is. Anyone can generate text. Knowing what’s true and what matters in your specific context is harder.
The discipline it requires
You have to be willing to edit, disagree, and throw things out. The moment you start accepting AI output because it’s “good enough” and you can’t be bothered to fix it - you’ve stopped thinking and started transcribing.
I catch myself doing this sometimes. It’s a habit worth breaking early.
The honest upside
I build things now that I couldn’t have built two years ago. Not because AI writes the code for me, but because when I get stuck, I have something to think with. That’s the version of this that’s actually useful.