[jmorrison]

Thinking out loud:

(1) The nameless model of reality I think is implied by "marginal cost of production is zero" simplifies away some essential aspects of reality (as explained by the late, great, RIP Manny Lehman model of "software evolution"). As the article points out, bug fixes and updates of the existing copy are necessary (to wit, IoT botnets), and in this model one never speaks about the continuing cost to maintaining the zeroth copy.

(2) The anticipated rejoinder to observation (1) is that, given the source, the end-users can do all maintenance, customization, and extension themselves. What we know about complexity (again, thanks, Manny) of any non-trivial application puts the lie to this.

(3) The anticipated rejoinder to observation (2) is existence proofs such as the Linux kernel. My reply to this rejoinder is that Linux is largely corporate-supported as a complement to the supporters products/services.

[TheOtherHobbes]

The initial cost of production is far from zero, as anyone who has spent more than few months (years) building a non-trivial application will tell you.

As you say, updates/fixes also cost. So do marketing, web infrastructure and so on.

It's usual in business to charge for utility and perceived value ("As much as the market can bear"), not for the cost of mechanical assembly/reproduction.

OP is correct that Apple has created a new(ish) distribution model. It was a gold-rush for a year or three, but has become extremely unfavourable to new product development.

But this has become a curation and discovery problem, not a distribution problem. Apple's 30% would be a lot more palatable if discovery and search worked better, if buyers could opt-in to a mailing list to support newsletters, cross-sales, and future sales, if sellers could set up trials - and so on.

None of this is likely because the App Store is an example of the first law of aggregation - be the feudal landowner, not the peasant smallholder.

The App Store is immensely profitable for Apple. The law of aggregation means that Apple has no serious incentive to make it better.

The App Store won't change now. Eventually the entire platform will collapse and be replaced by a new technology in a new market space. But I wouldn't bet on that happening within five years, and possibly not within ten. (My best guess would be somewhere between those two.)

Meanwhile it's still possible to do well selling app development as a service instead of apps as a product.