Before AI, as a head of product (who has always written code), I did this thing where when I was thinking through an idea or a product direction, I built the solution three or four times before I found the shape and direction that I liked. And once I liked it, I put it on a roadmap for one or more of my teams to execute on.

Candidly saying before AI is a little disingenuous, because since AI has gotten better in the last year at coding, my workflow has gone back to exactly what it was when I had a 40-person team reporting to me.

I still go through three, four iterations before a final direction is picked. It still takes me two, three weeks to think through an idea. Three things have changed.

1. When I think of a possible direction, a new version gets spun up within minutes to a couple of hours, usually in a single shot. 2. I can work through more big ideas which require some amount of coding-based ideation than I could previously. 3. And when a direction is decided on, the idea comes in to deliver the outcomes at a much quicker pace. Previously, it could have been 1 month of ideation + 2-8 sprints, now it's 2-4 weeks of ideation and 1-2 days to final delivery.

All in all, while I can see where the author is coming from, the grief has been different for me.

I've had a lot of good developers, product managers, product owners, and designers that have had the privilege of helping develop their skills in the past. That was the necessity of ensuring that we were developing talent who would then go on to produce good work on our teams.

And I'm at a stage now where a three-person team that I have can produce more than the 40 could, and I am likely never going to need to develop the skills the way I used to. The loss is not from coding, I thoroughly enjoy how that's evolved. The loss is from the white space around it.