The split the author talks about is about people who enjoy coding and people who enjoy having a working program. I can understand why he thinks that’s it, but I personally disagree for some reasons.

The craft-lovers and the make-it-go people sat next to each other, shipped the same products, looked indistinguishable. The motivation behind the work was invisible because the process was identical.

I think this thought process starts off wrong because a bunch of those make-it-go people became managers, directors, requirement analysts and other stuff that doesn’t directly involve code. I personally see myself as someone in the “make-it-go” camp, I want code that serves an actual purpose, even as a simple game. I enjoy coding most of the time, it’s “grunt work”, and there are times where I just groan, but I despise seeing my code go to waste. I also won’t work with AI code until it can actually pump out working code that I don’t need to double check for possible leaks, holes or bugs, I already have plenty of my own to contend with

Now the puzzle is architecture, composition, directing the assistant. It’s different. It’s still satisfying to me, anyway.

Welcome to the life of management. Or, as is right now, managing an unpaid intern that has limited memory, cannot actually learn from its mistakes, is overeager and can cause catastrophic failure with a stupid “smile” on its text response. There is satisfaction in doing correct management, but pretending “you” did it is fucking stupid.

I think a good analogy is that AI agents are all Tommy Wiseau in regards to acting and you have to direct him in your movie (program). The very best you can get out of it is something like “2 Fast 2 Furious” or “Transformers the Last Knight” - it “works”, but nobody will be happy. Similar to finished movies, the director can only take so much credit and a lot of the blame