Software is a False Realestate Business
In the real world bricks are laid, and cement in poured into a finite set of rooms and apartments (n). Those are then sold to their respective new owners, and the value of each each unit (u) is x / U*n
Although software finiteness is more flexible, and scaling costs almost nothing, it follows more real estate logic than we realize.
Model View Controller (Reenskaug, Kay, Goldberg 1979) is de facto the only server architecture we use. It has 2 main, and unchallenged characteristics: stationary databases, and proprietary interface points.
In our building, it means that the basement has a bunch of servers, and that for anyone walking in needs to be signed in by the doorman, or at least have you in the book (register, sign in).
Then, once in, every step is followed, and every opportunity for attention (new door open, idle time in the elevator) will be monetized (by ads).
The owner of the building might have different ways of funneling you through a building you don’t own. Remember you didn’t really need this building in your life (if you did you would have built it, or bought a part of one).
Maybe they will get some of your friends together, put together a music show, or teach you a new langauge for free.
Once you left the building (browser tab closed say) you have left data behind which contributes to the overall nudging capability of the building. More efficient attention grabbers will be in place next time you visit. More rooms, more concoctions, and more time you will spend in the building (instead of playing outside).
On the other hand, plumbers (decentralized, edge services – more on that later) do none of this. A plumber comes to your house, fixes your pipes (likely of a different set up than their earlier home visit), gets paid (services, not attention) and leaves (no network effect).
I am lacking the language or the mental models to think of a version of a product, that does not effectively make it a service offering. Especially as a lot of the current network effect (MVC) type systems are easy to automate (consider that all it is big data optimization against trial and error).
What would Facebook look like as a plumber? Maybe a delivery service of photos from your family? Or a weekly reminder for a video call catchup with family around the world? But without a dimension of services (opinion, conviction, brand) it seems incredibly flat.
Maybe abundance means a flattening return on incremental products work. Consider why you choose one piece of software over another - what is driving your trust and decision?