How to Save Android

No. You build one version; this is the premise behind the VM environment that runs your code, which is "write once, run everywhere". If you depend on system APIs only provided by Google or Amazon, you have an additional version for each system that exclusively provides those APIs.

> consumers face a bewildering array of user interfaces.

Not in the apps themselves. Conflating the system UI with the app UI?

> They have made no real effort to fix the root problems – especially around fragmentation, security and privacy.

I'll push back on this without expending too much effort: Treble, Mainline, and the new Privacy Sandbox are actual efforts that have improved things on the security/privacy front. AndroidX makes fragmentation a non-issue beyond the most esoteric of app requirements.

> Google should make Android truly open source.

I mean, I am able to submit patches to AOSP and AndroidX, have them reviewed, and merged. There are uncountable AOSP forks out there. I'm just not sure how much more open-source you can get? Unless you mean "open governance", which I could get behind.

What? No it doesn't. Which Android devs are building hundreds of versions of the same app? Who is this guy?

>> We know many software developers who insist on using an Android phone out of principal, but their green message bubbles stand out as exceptions

Oh right, someone who thinks America is the whole world. The green iMessage bubble and all the madness it leads to is a US specific problem. Nobody I know cares about iMessage at all.

>> Consumers, especially young consumers (aka customers of the future) prefer iOS by wide margins.

No they don't. Android dominates with 70% market share. Even in America iOS is only ~60% market share, which is not a "wide margin".

>> The result is that iOS users monetize for app developers much, much better than Android users

Yes because young people are famous for their abundance of spare cash. Eyeroll.

The chromebook grade application processors lead the charge. But we're seeing wifi router & phone chips also getting upstream support.

Enabling people to build neat devices & alternatives can generate net new demand, open markets. But people need access to dev tools & chips at reasonable prices with low barriers, and they need software support to build on top of. For a long time the whole universe of making connected devices of any kind has been deeply foreboding, filled with barriers everywhere, and that's finally getting a little better. Buying chips is, however, still really really really hard.

Android themselves getting to be more up streamed & less special stack is another slow burning change that I think will help their platform. Chromebook doing Android apps is a genetic infusion of possibility into a kind of sterile corporate ware os, bridges a gross gap. Android is still distinctly Android & uninterested, but at least there's some technical possibility of hybridization that's now starting.

It'll never happen because they won't be able to agree on which cloud these backend services will sit in.