The people around me who understand AI the best are often the most critical of what it produces.
And I don't mean they believe less in AI. Quite the opposite. Many of them use it heavily.
I notice the same thing in my own work.
Recently I had AI working on a task for me for hours. It kept trying different routes, making things more complicated and somehow getting further away from a simple solution.
At some point I stopped it.
I looked at the problem myself, suggested another approach and gave it a few clear directions.
Around 20 minutes later it was done.
Even the AI basically acknowledged that my approach was much simpler than what it had been trying for hours.
But this happens both ways.
Sometimes AI gives me a solution where I think: okay, I would never have approached it like this. And I learn from that. A lot actually.
AI corrects me. I correct AI.
When I correct something important, I also explicitly tell my AI to learn from it and remember it. I don't want to have the same discussion again next week.
I see people saying they use AI for 90 or 95% of their work. Sometimes I think they actually mean 110%.
And yes, I still receive emails and documents where ChatGPT or Claude instructions somehow made it into the final text.
For me, using more AI doesn't mean I have to think less.
Sometimes I let it run. Sometimes I stop it. Sometimes it teaches me something, and sometimes experience tells me there is a much simpler way.
Maybe that interaction is the part of AI I enjoy the most.
I learn from AI every day.
And I make sure it learns from me too.