pull down to refresh

Others have mentioned this but this article is a pretty good summary.

Is AOSP going away? Google says no

As promised, Google published the source code for Android 16 this week, allowing independent developers to compile their own builds of the new operating system. This source code was uploaded to the Android Open Source Project (AOSP), as usual, under the permissive Apache 2.0 license.
However, multiple developers quickly noticed a glaring omission from the Android 16 source code release: the device trees for Pixel devices were missing. Google also failed to upload new driver binaries for each Pixel device and released the kernel source code with a squashed commit history. Since Google has shared the device trees, driver binaries, and full kernel source code commit history for years, its omission in this week’s release was concerning.
These omissions led some to speculate this week that Google was taking the first step in a plan to discontinue AOSP. In response, Google’s VP and GM of Android Platform, Seang Chau, refuted these claims. He addressed the speculation in a post on X, stating that “AOSP is NOT going away.”

How does this impact developers of OS's like Graphene and CalyXOS?

Previously, Google made it simple for developers to build AOSP for Pixel devices, but that support is now gone. Developers simply had to “pull the configurations [that] Google created,” add their customizations, and then build. Now, however, they will need to take the old device trees that Google released for Android 15 and “blindly guess and reverse engineer from the prebuilt [binaries] what changes are needed each month.”
So this is not good at all but not game over.
These omissions led some to speculate this week that Google was taking the first step in a plan to discontinue AOSP. In response, Google’s VP and GM of Android Platform, Seang Chau, refuted these claims. He addressed the speculation in a post on X, stating that “AOSP is NOT going away.”
He also confirmed the omission of Pixel device trees is intentional, stating that “AOSP needs a reference target that is flexible, configurable, and affordable — independent of any particular hardware, including those from Google.” Instead of supporting AOSP builds on Pixel devices, Google will support the virtual Android device “Cuttlefish” as its reference target. Cuttlefish runs on PCs, allowing Google and platform developers to test new hardware features. Google will also continue to support GSI targets, which are generic system images that can be installed on nearly any Android device.
And this little tidbit seems very concerning.
Furthermore, Google’s decision to squash the kernel source code’s commit history also hinders custom development. The Pixel’s kernel source code was often used as a “reference point for other devices to take features, bug fixes, and security patches from,” but with the history now reduced to a single commit, this is no longer feasible.

My Take

Google is doing this intentionally and strategically to limit the reach and competition it is getting from alternative ROMS. The popularity of Graphene and other ROMS has grown since 2020 when I first started trying these systems out.
This is a pattern we have seen in companies built on open source over the past few years. There are business reasons for some of the changes but also there are technical reasons. I don't like it one bit. Google has long claimed to be a champion of open source and honestly while the custom ROM space is growing the idea that its a threat to Google is laughable.
It is possible that this is just Google tightening the belt and reducing dev overhead. Either way its not good for people who develop more privacy respecting mobile computers and signals a massive need for open hardware and a truly open phone OS.
152 sats \ 1 reply \ @freetx 13 Jun
In the long run this is good.
We need to push outside of the google/apple duopoly. A real-free hardware platform running a real-free OS would benefit everyone.
reply
Indeed. Its just not easy to build hardware. I think if the market for these alternate OS's grew we would see more focus on open hardware. Its a chicken egg problem.
reply
Maybe @final has a suggestion regarding how we can help in the most useful way?
reply
56 sats \ 1 reply \ @nichro 13 Jun
Hmm good info. But I doubt the competition from custom ROMs was really making them sweat. it's a tiny part of the market.
Besides, I've bought 3-4 pixels in my life and I only bought them because GrapheneOS exists. I'd be willing to bet a significant part of said custom ROM market participants would also say the same
reply
That's an excellent point on GrapheneOS driving pixel sales. Its pretty much the same for me.
reply
I've been using GrapheneOS for a little over a year, so it's a bummer to hear that the future outlook of the project doesn't look good. It feels like whenever I take steps to improve my privacy and security with new tools and services, they eventually get cracked down on or don't become as effective.
reply
I think the project will survive. Its just gonna have to adapt and I imagine right now its pretty depressing for the devs.
reply
One of the main devs is in the front lines...
And now they have to do much more extra work to support Android 16 because Google didn't release the code for the devices...
Seems like a bad time for the project, but they're continuing.
reply
I think in the end GrapheneOS will need to use their own OS, and hopefully their own hardware. They tried doing the hardware bit some time ago but it didn't end up working for some reason. Maybe it's a good time to revisit that.
In terms of the OS, they discussed this a bit and the long term plan is to use something along the lines of Redox
reply
I didn't link to them but the GrapheneOS guys have been talking about this on Nostr and it lines up with what they have been saying. There's a reason GrapheneOS focused on Pixel devices. It was much easier to do than other devices. That is over now.
reply