It's not either-or. I don't buy the argument that if we didn't shovel garbage that we call "software" today, we wouldn't have equivalent but better software at all. It's a multi-agent problem, and a lot of it is driven by business dysfunction, not even actual complexity or programmer laziness.

In my - perhaps limited - work experience, there's enough slack in the process of software development that I don't buy the "time to market" argument all that much.