When I use any moderately complex piece of software, for instance a word processor, the UI is stacked full of things I don’t use, making it less convenient for me to use. At the same time, simple features that would be useful to me are not present. Software is currently aimed at the highest common factor so that it appeals to as many people as possible, which paradoxically makes it suboptimal for everybody.

If I wanted to build something that is specific to my needs, this would be prohibitively time consuming and expensive. Even today with all the latest models – even if what I want is relatively mundane.

To add on to that, what would be produced would be ideal for me but less ideal for other people. Other people need things that I don’t, and they don’t need things that I do. And people’s needs change over time. So the actual range of software that there is appetite for is the result of a huge combinatorial explosion of features, for every single type of application out there.

The appetite you are thinking is satisfied today is merely “there is an app that does X” but the appetite that is actually present once we are able to create software much more efficiently is more along the lines of “everybody gets their own custom app that does X”.

I don’t think the appetite for software can be quenched until we have just-in-time feature generation. That is definitely not within present day capabilities.