There has been progress.. but most of that progress is nothing when offset against the vastly more complex things we're building nowadays

My reaction was even stronger. Forget "order of magnitude", I question whether the world wide web, for all its other gains, has had even a net positive effect on "programming productivity/reliability/simplicity". The latter two axes seem fairly uncontroversial? For productivity it would depend on how you define it. If you chose a metric of "LoC generated" we're doing amazing. But that metric seems misguided. In general, the terms of Brooks's essay map to today's world in non-obvious ways. We have started "growing software more organically" as Brooks recommended. And that has generated new problems that OP seems to largely ignore.

LoC generated is a liability 🙂 If anything, a problem we'll face is we're producing more stuff than we can maintain, or that the people that depend on the software can pay to be maintained