I always find it odd when people talking about high level policy stuff like EuroStack completely ignore end-user devices - Like, if you have fully sovereign EU AI cloud apps, where are you going to access them from? Your iPhone?

https://ecfr.eu/publication/get-over-your-x-a-european-plan-to-escape-american-technology

Get over your X: A European plan to escape American technology – European Council on Foreign Relations

Foreign technology companies cannot be entrusted with meeting Europe’s growing digital needs. This includes American big tech firms. Here’s what Europeans shoul

ECFR

@tbernard and because what is happening in the US can happen in Europe (or has happened in some countries, even), as a citizen I want to use international, transparent, open tech.

I don’t want tech nationalism.

@thibaultamartin @tbernard I agree with you. I think that while we’re seeing a bit of tech nationalism, the current positive attitude toward FOSS in Europe will hopefully lead to more transparent, open technology.

At least for now — until politicians decide that open-source projects developed in the EU are competitive enough and start lobbying to close the sources.

@tbernard I will access it via my @murena phone running @e_mydata on a @Fairphone or use my @tuxedocomputers running @debian . There are great European solutions, you just need to buy them ✌️🇪🇺💙
@spipau how's that going to work if Android goes proprietary?
@tbernard @spipau And with an SoC that still has an American backdoor in the CPU and cell modem.
@blobjim @tbernard step by step we have to take back control and nothing will every be perfect straight from the start. For my part, I can't wait to buy open RISC-5 processors, but they are not ready yet...
@spipau @tbernard yeah I was looking at RISC-V single-board computers and they have the same problem as ARM devices, where you can't install a generic Linux distro on them because they don't support ACPI and there's no way to make a generic Linux distro boot with a specialized devicetree easily. So the designers provide a special build of a Linux distro, which is totally impractical for real world use.
@blobjim @tbernard hopefully this will change in the future. @frameworkcomputer is doing great work in this field: https://frame.work/at/de/products/deep-computing-risc-v-mainboard
Mainboard (DeepComputing RISC-V)

DeepComputing is creating a RISC-V Framework Laptop 13 Mainboard powered by a StarFive JH7110 processor.

Framework
@tbernard while the Android OS itself is open source and cannot go proprietary, the rest which makes phones smart (PlayServices: push notifications, location, maps, etc) have always been proprietary. Also the cloud ecosystem around is closed source and this is where Murena steps in. Replacing all proprietary parts (e.g. PlayServices with @microg or cloud with @nextcloud ) and make it as easy as possible so non engineers can use it as well.

@spipau Of course it can go even more proprietary. Major parts of what typically is part of a phone OS, like the Clocks or SMS app, have been proprietary for a while now. Android 16 QPR1 (the latest version) was released to various devices in September, but the source code was only published 2 months later. And security updates are already being delayed in AOSP by up to three months vs the proprietary version.

@tbernard

@spipau At any point, Google can decide to stop updating AOSP. Stop publishing security patches and new major updates. Of course you can still use the old code that's already published today, but it will be a security nightmare in just a few months and will stop being able to run new apps in 3-5 years.

@tbernard

@larma @tbernard very valid and true for any open source project. If you'd stop working on microg then the same would happen to my Murena phone - thank you very much BTW, an awesome piece of work and I am one of your donors 🙏❤️

Seems like Google gets more and more afraid of our liberation movement, which should encourage us to continue. Europe and the rest of the world might have to fork Android and continue without Google, although this would be a tremendous effort! 😅

@spipau @tbernard microG could be continued without me. There are other contributors and certainly enough developers that could contribute. But that is because microG is a free software community project.

Android in contrast is exclusively developed by Google. AOSP is merely a code dump, it's almost impossible for non-Googlers to contribute to Android or AOSP. Even if you were to fork it, you won't have the necessary developer and security researcher power to compete.

@spipau @tbernard Thus, rather than forking Android, we'd be better advised not to build on top of AOSP and instead use the free software platforms we already have as a basis, making use of the huge amount of developer effort and research put into free software Linux distributions and apps running on them, and making that available to smartphones.
@postmarketOS is laying the groundwork here and initiatives like @modal are pushing to build the remaining pieces to have something that can compete.
@larma @tbernard @postmarketOS @modal Yes, very much agree and @postmarketOS is great! I don't think that it will be long and they are a very valid alternative
@larma @spipau @tbernard @postmarketOS @modal like it's not even funny how much hope (and money) I'm putting into PostmarketOS. No pressure 😅

@larma @spipau @tbernard yup recently android is becoming more like closed source. Their development is behind closed door. And they are planning to remove features like side loading (installing apks).

Which is really bad if you think about it, you do not own anymore your own mobile phone.

@larma @spipau @tbernard while I agree with most of what you say‚ "will stop being able to run new apps in 3-5 years" is very false‚ typing this from 8yo phone.
@rangelovd is your OS 8 years old or just your phone?
@dantescanline the os is ancient and insecure as well and shouldn't be used. However‚ it suits my little needs and the bootloader is closed leaving me with very few options. My next device will most likely have linux with Gnome though

@rangelovd @dantescanline 8 years ago, Android 8.1 was just freshly released. Here's what Android Studio defaults to for the minimum SDK version for new apps. Even if some developers might change this, most won't, so most new apps simply won't work on Android versions older than 10.0.

And that doesn't even consider that some apps will opt to even higher version requirements simply because they require or wish to use newer features (like secure key storage, specific hardware APIs or similar).

@larma @dantescanline That's fair! Clearly I am not an average user so maybe that's why I haven't ever noticed.
@larma @dantescanline though I'm not sure how is this still relevant. I'm just an unfortunate user of unsupported device like people with Nvidia Pascal GPUs. Many modern phones receive OS updates up to 7 years. This it 7 years of the most latest OS version after which there is still a lot of time before things stop working.

@rangelovd @larma i think you've misunderstood what's being discussed here then?

degoogled open source forks of android (murena etc) completely rely on google releasing each android versions source code, if google cut them off then all users of those open source forks would decay similar to your own unsupported device. apps that target newer APIs will not be able to run without herculean effort from a small open source community.

@rangelovd @dantescanline We were talking about AOSP stopping to receive updates, not the proprietary Android version devices run officially. If the device offers 7 years of updates, sure, it will receive 7 years of *proprietary* updates.

But I'm not aware of any device coming with 7 years of AOSP updates. Because device manufacturers offer those 7 years based on a contract they have with Google for their proprietary Android. And that doesn't cover AOSP.

@larma @dantescanline ooooh I see. Thanks for explaining this to me 😅

@tbernard 100% this.

In 2 years a small team at Mistral managed make a competitive(ish) LLM, but in 20 years nobody (not even Microsoft) managed to make a decent competitor to iOS & Android.

So maybe the EU could start to realize that the hard part is not AI, it's the platform

@tbernard our experience as @ev when talking to policy makers on these topics are:

1. We're doing it later as non-US cloud is already hard enough (most common one)
2. They're not aware (basically ignoring what they're using every day)
3. They think it's the easy part as long as the apps they develop are #OpenSource for Android

…but I'm positive that we can move some goal posts in 2026 in that regard.