Software development is complex, for any non-trivial project. But complexity is hard. Overwhelmingly, when we in the software field talk about developing software, we’ve dealt with that complexity by ignoring it. We write code samples that fit in a tweet. We reduce interviews to trivia challenges about algorithmic minutia. When we’re feeling really ambitious, we break out the todo app. These are contrivances that we make to collapse technical discussions into an amount of context that we can share in the few minutes we have available. But there seem to be a lot of people who either don’t understand that or choose to ignore it. They frame the entire process of software development as being equivalent to writing the toy problems and code samples we use among general audiences.
I tend to be skeptical of claims that software development is some sort of inherently different art from other pursuits, that it is uniquely complex. However, I think this analysis is spot on.
Moreover, I think that one of the big problems with the broader approach to the use of LLM-based tools is that the tendency for their proponents to mistake the output of labor for the labor itself.