> An business should be able to buy cheap Android phones in bulk, install Linux
Good luck with that. Unless you settle for Android middleware underneath your OS (like some existing projects do), it's a massive undertaking. You still need to invest a lot into R&D while you lose what's differentiating you in the process (in case of Librem 5, there are things like hardware kill switches, PGP card reader, replaceable M.2 modules etc.) so you can chase SoCs that will already be considered obsolete once you're done with the software anyway. Even Purism with their prices couldn't afford to use an unsupported SoC and relied on NXP's upstreaming efforts.
I suspect that if I was wrong and it was actually a viable path it would already be taken by someone. Maybe we'll see something appearing in coming years, as these days more vendors actually started to care about mainline support, but as it is today the landscape is still not that great, especially when it comes to SoCs that are actually targeting smartphones.
Lots of companies build/buy/sell cheap phones and presumably put some work into getting Android up on the hardware.
I think maybe that's where Purism went wrong. They tried to go pure from the get-go. But think it would be better to get a profitable product out first, and then invest in opening components, one by one.
I don't need a perfect device. Rather hardware getting better every year. Then I could upgrade, and they'd make more money.
Getting Android up on the hardware is easy - you just grab the unmaintainable kernel fork provided by the vendor and you're 90% there.
I'm not interested in Android though, so I just wouldn't buy such phone no matter how cheap would it be. I'm not interested in upgrading every year either. Librem 5 is great, but unfortunately I'll need something to upgrade to in a few years anyway, as the Web has a tendency to only get heavier with time.
And that's what I'd love to see. It's not as easy as you make it appear though. An obvious upgrade path for Librem 5 would be a newer SoC from i.MX line as there have been several released since, but all of them are even less suited for a smartphone than i.MX 8M or come with downgrades in some important areas, such as GPU - so this now calls for a complete redesign around a different vendor. Not impossible, but not cheap either.
Grabbing some cheap Android devices is a path to nowhere. Otherwise we all could be just grabbing them and putting postmarketOS on them ourselves, as there's no shortage of such devices on the market.
A defeatist mindset. The point is to make money, then invest in making more.
For example my new starlite linux tablet is awesome and star is making money hand over fist. Because they finally built a good one and stopped making excuses. They develop firmware with coreboot and upstream patches to Linux.
That’s what a successful business is, you invest then reap the rewards.
Good luck with that. Unless you settle for Android middleware underneath your OS (like some existing projects do), it's a massive undertaking. You still need to invest a lot into R&D while you lose what's differentiating you in the process (in case of Librem 5, there are things like hardware kill switches, PGP card reader, replaceable M.2 modules etc.) so you can chase SoCs that will already be considered obsolete once you're done with the software anyway. Even Purism with their prices couldn't afford to use an unsupported SoC and relied on NXP's upstreaming efforts.
I suspect that if I was wrong and it was actually a viable path it would already be taken by someone. Maybe we'll see something appearing in coming years, as these days more vendors actually started to care about mainline support, but as it is today the landscape is still not that great, especially when it comes to SoCs that are actually targeting smartphones.