I'm much more interested in a realistic bang for your buck discussion about software value for effort and diminishing returns rather than a conversation about whether it's possible to build great software.

Also notable that not up for discussion was any second order effects about user interaction with software; software always must serve its user in the immediate first order. Why? I suppose it's because if you attempt to examine nth order effects the proposition completely falls apart because the world is much more complicated than "software must serve the end user".

Finally, the purpose of software isn't determined by the end user, nor necessarily its author. A user can have their goals, the author(s) theirs, the funding body theirs, and so on. The "purpose" is diffused in this way until what's left is what it does. Saying it _must_ serve the user is a principled stance reflecting a world view, but it's just that. That are other coherent worldviews where software has more than one master. Or no master.