AOSP Source Releases Are So Late Developers Disassemble Libraries to Report Bugs
github.com
github.com
I think that's the one that made a lot of OEMs pull back on opensource publishing, because their legal departments want to batch the CVE monitoring and reporting obligations into smaller batches.
As long as they start the process and it is not absurd from an industry standards perspective, they should be good. They can have slow processes, including being a good bit slower than their peers, without running afoul of "without undue delay".
Well, the corporate lawyers think otherwise at those corporations, so I'm not sure how much your interpretation counts :)
Maybe the misunderstanding stems from the related common misuse of "as soon as possible" as "immediately"?
A corporate lawyer would, however, always suggest to minimize/reduce/delay any legal obligations so as to minimize risk of liability, regardless of whether said obligation is easy to comply with or not.
Security flaws are discovered almost every day (now that they are automating it with AI there are a ton of reports), pretending the manufacturer releases an update in 1 day is nonsense.
Just building the code alone it would take a lot of time, and of course you have to test it at least a bit before giving it out to the end users, you can't pretend to rollout on a global scale a nightly build without first at least some level of QA risking bricking user devices.
And you have to consider that phone manufacturer doesn't work h24 and 7/7 of course to give out patches.
I think that we are fortunate if the whole process takes less than a month, not a day.
> The CRA requires manufacturers to notify actively exploited vulnerabilities and severe incidents having an impact on the security of their product with digital elements. They need to submit an early warning within 24 hours of becoming aware, and a full notification within 72 hours.
These deadlines are entirely unrelated to the availability of security fixes and delays in deliverying them.
The part that’s probably bothering Google is this:
> 6. Manufacturers shall, upon identifying a vulnerability in a component, including in an open source-component, which is integrated in the product with digital elements report the vulnerability to the person or entity manufacturing or maintaining the component, and address and remediate the vulnerability in accordance with the vulnerability handling requirements set out in Part II of Annex I. Where manufacturers have developed a software or hardware modification to address the vulnerability in that component, they shall share the relevant code or documentation with the person or entity manufacturing or maintaining the component, where appropriate in a machine-readable format.
To put it very simply, the CRA calls on all stakeholders to cooperate in reporting incidents and vulnerabilities, establishes a national coordination mechanism, and aims to address vulnerabilities swiftly in order to maintain essential services within the European Union.
Also, major Android releases and QPR2s have many security fixes (anything not marked high/critical), yet OEMs are not rolling them out (QPR2) or very late (major releases).
Just to say that OEMs were probably violating the act already before Google started withholding source code.
I also think it’s an attempt of Google closing up AOSP more, otherwise they wouldn’t make it so darn hard to get updated Pixel kernel trees.
withholding security updates seems to avoid a law which requires rolling out security updates seems very much like something which needs to be fixed one way or another
Also - the reality of what you think is ok and what a corporate legal department think its ok is... rarely the same -_-
Then there is no time limit to provide a mitigation. It should just be worked out without unnecessary delays. Once a mitigation is published, a report should be submitted with 14 days.
Regulators present an impossible goal to the companies to implement, instead of implementing the goal, the companies try to reward hack their way around it.
The end result is that they basically try to make it impossible to know about vulnerabilities because that would start the clock.
But many Android OEMs are probably terrible at software engineering and have their attention spread across multiple SoCs, etc. It’s their own fault.
What does "timely" here mean? A lot of security patches for all these companies take ~weeks+ to actually make it to customers due to things like QA testing, carrier testing and release processes.
The time between disclosure and release can be rather long, months even for all of them. (I'm counting private disclosure here.)
Google has an embargo period of up to 6 months on AOSP patches. This is why GrapheneOS already has April 2027 AOSP security patches now. Even though some patches are already known for 6 months, they still manage to ship it 1-3 months late. 7-9 months after they got the patches already from Google.
due to things like QA testing, carrier testing and release processes.
Carrier testing is mostly a US thing. If you cannot roll out security patches within a few days, your QA and release processes are just totally broken. GrapheneOS can do it and they generally have far fewer regressions than Android OEMs.
Major releases and QPR2 take some time for testing, but they already have the access to sources months before they are released. Heck, Samsung releases major release betas and QPR2 betas quite a while before Google releases it as stable.
The problem is that the OEMs made a mess. First they just dropped phones on the market without doing any updates at all. Mid 2010s it was very normal for an OEM to do maybe one or two updates. Then Google had to push them to do better, among other reasons because Apple does much better and there were really big fails like Stagefright. Plus users wanted new Android releases. Then they did only 2-3 updates with very irregular software updates. Now they are under pressure from EU regulations and they actually have to support devices for a long time and do regular updates, but they just don't have good release processes. Plus they made their own 'skins' on top of it that they have to rebase, etc.
Companies that have their ducks in a row better, abuse the release process to push sales of new devices. E.g. Samsung withholds major/QPR releases for a while so that only their latest device have it, e.g. recently Fold 8/Flip 8.
People should stop beingapologists for multi-billion dollar companies.
That's the reality - yes, even for Apple.
If you think they could and should be better - I guess I agree, but it doesn't change the reality.
> an early warning notification of an actively exploited vulnerability, without undue delay and in any event within 24 hours of the manufacturer becoming aware of it
> without undue delay and in any event within 72 hours of the manufacturer becoming aware of the actively
> without undue delay and in any event within 24 hours of the manufacturer becoming aware of it
> without undue delay and in any event within 72 hours of the manufacturer becoming aware of the incident
"Sorry your honour, we only became 'aware of it' 2 years later, pinky promise ;)".
So any competant authority would have to prove when they 'became aware of it', whatever the test to pass that threshold means.
“Please stop using this version of my product or this particular feature; it is vulnerable” is an acceptable mitigation measure.
A CVE isn't an active exploitation and don't trigger anything.
Otherwise I would expect Google would also wait with pixel updates
This is not true at all though. OEMs can release anything whenever they want.
https://grapheneos.social/@GrapheneOS/117282080803799576
Since they are working with two OEMs (one for GrapheneOS support and one for getting embargoed security patches), I think they know what they are taking about.
Graphene isn't OEM so they don't get access to Android repository before AOSP drops. OEMs - like Motorola for example - do have access to partner repository and can ship QPRs if they wish to do so. Most often OEMs do not want to push updates (and it took years to make them push regular major Android updates).
> I think they know what they are taking about.
Err, if you look at that account you'll find that the person running it has a penchant for overexaggeration and selective truth telling when they feel aggrieved.
Are reading the same post?
“ The new APIs are currently exclusive to the Pixel OS and aren't available to other Android OEMs. […] Non-Google Android OEMs and AOSP-based projects can ship yearly and QPR2 releases.”
the person running it has a penchant for overexaggeration and selective truth telling when they feel aggrieved.
Here we go again…
Maybe their own source at OEMs lied to them? The QPRs are in the same repo they're getting their security patches from.
That doesn't mean that they are actually allowed to release them. Big difference.
I wouldn't be surprised if large corporations like Apple and Google intentionally use these tactics to pressure regulators through their European users. Siri AI is a prime example. While it's a complicated issue, I certainly wouldn't put such maneuvers past them.
When your system favors people into power that abuse that position instead of furthering the public interest, that provides a feedback signal leading to ever worse people and outcomes.
There is no need for a "conspiracy" to get to a point where it looks like one.
However changes like these are deeply unpopular as they get interpreted as moving more powers to "Brussels" by right wing media.
Their agenda isn't that of "ordinary people". Those rather get used for the interests of those "strategic investors".
Less power for Brussels means more power for rich people.
A more powerful EU would mean less power for big corporations and wealthy individuals.
It mandated USB for charging everywhere, but didn't do a thing about driving on the right side of the road.
We have the digital markets act, yet when someone tries to send emails to some big American email providers, they have to register first, otherwise they don't even make it to the spam folder. But hey, let's have a big round of applause for fining X/Elon Musk and Meta because we hate them. Who cares about decades old protocols and ways of doing things?
The PDF for GDPR has 88 pages [1]. And it's just one thing. The AI act has another 144 pages [2]. Who's gonna read them?
[1]: https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELE...
[2]: https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=OJ:L...
https://www.congress.gov/105/plaws/publ304/PLAW-105publ304.p...
We have civil code for everyday life, do you think people read that? I bet no, but anyway it describes rules by which people live. And when there is disagreement we open a book and see whos right. Do you think we should not have such rule book?
It's not like regulators were just bored and felt like fucking shit up, there is a nearly-global consensus these companies are exploiting consumers and markets.
No, they just want to have something to brag about or think they know better (Soviet/communist style). Like, we're lazy and tired of having to carry our proprietary iPhone chargers, we know you're too, let's make USB mandatory.
I can (actually can't, but for the sake of argument let's pretend) imagine why my wife's Air Pods Pro 3 can't supply heart rate data to my phone for exercising. But why can't they report battery level? ffs
Even if Lightning is better in some aspects (USB-C tongues break sometimes), forcing people to have special cables just for apple devices, when every single thing runs on USB-C, feels like done out of spite. And then? Apple devices don't support any fast charging standard. Spite again.
It's hard to see it differently, but there might be better reasons and arguments that I can't or choose not to see. Probably because I don't love the company that forced poor parents to buy their kids expensive phones just so the kids wouldn't get ostracized for not having green or blue bubble in text messages.
> USB-C tongues break sometimes
Unlike barrel chargers which are also simpler. No need for fancy electronics and protocols. Let's not forget non-compliant USB chargers. I remember reading about a Chinese mini-PC manufacturer that sold a computer with a charger which provided 19V or so from the start without any negotiation. Hopefully no one tried to use it with another device.
I actually (or somebody close, friend or family, can't recall) did exactly that. It was a few years ago and I don't remember what got fried. Some small electronics. It was a barrel charger and it fit. Too bad that fitting didn't mean the power levels were ok too.
> I don't like the product because of its charging mechanism, not reporting battery level etc, I don't buy it.
Why shouldn't my wife get headphones because a company is a dick and makes them not work with her husband's phone properly? Looks like a dismissive argument trying to excuse malicious business practices. (I'm not in or from US, maybe that's normal and OK there, so sorry for ranting about it, as it might be fine or even good for you if you're invested in Apple). Anyway, why not buy them, headphones work very well for my wife. Other than price not being adjusted to product performance and quality, AirPods Pro 3 are a very good product. HR reporting is a cool feature. And I don't have bud-type earphones, so I borrow hers for exercising.
Back to "don't buy it if you don't like it" argument: it doesn't argument against apple being dicks for making the product purposefully not work with an Android phone. And if it does, please explain it more clearly, I believe I'm not the only one holding the point of view that it just doesn't exonerate them.
As I see it, it's malicious compliance, as there's no laws YET forbidding what they're doing. (Somebody recently RE'd AirPods BT protocol/calls IIRC and found them reporting battery on non standard ?variable name?. He then got Android phone to show battery level after a small modification.
For comparison, I would bet a beer (I see betting more as gambling) that if Samsung created headphones with HR monitor, I would still be able to see battery level and HR in my Oura app on my OnePlus. And vice versa for 1+ headphones on Samsung. Do you see my point?
And for the record, apple has recently released many phones, both low and high end, that weren't overpriced. And I vehemently oppose anyone whom I'll have to tech support later, getting any phone different than apple - for a non technical user these phones are perfect. Good default's. Non customizable - perfect from point of view of tech support. Just works. Just to show that I'm anti anti consumer conduct, not anti apple.
You forget that the worst offender took its toys and went away not too long ago.
They were fined because they broke the law. Many consider these fines, a slap on the wrist. Personally I support much harsher and detrimental punishments in such cases.
Sure it has. Consider cookie banners: widespread malicious faux-compliance from the start, and approximately nothing was done about it for most of a decade, while people were tricked into believing they were GDPR’s fault. Even now when a little has been done about it, the significant majority of cookie banners are very obviously not compliant with GDPR or ePD requirements, and no consequences occur.
Correct? I suppose that depends on your definition. From a corporate perspective, perhaps. But from a consumer's perspective it's most certainly not the desired outcome.
They care about their partners and about following the law. Google is not being spiteful here like some commenters are saying; instead, they have changed the way they work to make it compatible with a new law that applies to them.
It’s tiresome to see the same kind of snarky comments in every thread because they are not interesting at all. They are just lazy, and begging for votes.
Even if we set legality of their behavior aside, they deserve scorn for how they've been managing Android for the past couple of years - every Android user is worse off because of them. They broke off productive relationships with grapheneos and other projects which contributed patches and reported vulnerabilities upstream, they're engaging in anti-competitive conduct and imposing limits on OEMs, giving Pixel devices preferential access to new APIs.
They're simultaneously putting their own customers (Pixel owners) at risk of known, exploitable security vulnerabilities. They do this to help other OEMs legally skirt (maybe) the 4 month timeframe for patches which is very lax to begin with. Thereby opening a whopping 8 month window where Android users can get hacked by known vulnerabilities.
Their Android 17 QPR1 release was so buggy, it made many Pixel devices almost unusable and then they went radio silent. https://grapheneos.social/@GrapheneOS/117372129125340408
How cynical do you have to be to defend their stewardship of AOSP?
> Google is not being spiteful here like some commenters are saying; instead, they have changed the way they work to make it compatible with a new law that applies to them
Read the second link I posted above, this is patently false.
I worked at MS and was involved in several decisions to not ship in the EU because it just wasn’t worth the internal and external regulatory headache. There was no secret agenda, Satya never told me to stick it to the EU by withholding a feature in the hopes the masses would clamor for it and demand regulatory changes, which would eventually allow us to… ship the feature in a decade or so.
I’m super sympathetic to the EU’s intent, but their execution is so obviously self-defeating that it doesn’t take elaborate conspiracies to explain corporate response.
Broadly, I agree, and good grief do I wish people would follow this more often. But Google specifically has acted incredibly anti-competitively for several years now [0], especially regarding Android. This has gone on for such a long time that they have, in my mind, lost the benefit of any doubt.
As they continue to lock down Android [1], things that would have once been met with a "Wait and see" approach are instead met with suspicion and a default assumption that it's primarily (though not typically exclusively) done to serve their own interests.
So generally, yes! That which is asserted without evidence can be dismissed without evidence. However, Google's history is all the evidence I need to assume their actions are not done in good faith.
[0] https://arstechnica.com/gadgets/2018/07/googles-iron-grip-on...
[1] https://manualdousuario.net/en/blocking-unverified-apps-andr...
Well no. It’s not self defeating. It’s being defeated because these are US tech companies, protected by US monopoly power, and the fact that US lawmakers are captured by lobbyists and are incapable of passing similar laws to protect their constituents. Big difference.
[1]: https://news.ycombinator.com/item?id=45911673
[2]: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=OJ%3AJOL...
> Clause 1.2:
> (6) Operating system updates:
> (a)
> from the date of end of placement on the market to at least 5 years after that date, manufacturers, importers or authorised representatives shall, if they provide security updates, corrective updates or functionality updates to an operating system, make such updates available at no cost for all units of a product model with the same operating system;
> (b)
> the requirement referred to in point (a) shall apply both to operating system updates offered voluntarily by manufacturers, importers or authorised representatives and to operating system updates provided to comply with Union law;
> (c)
> security updates or corrective updates mentioned under point (a) need to be available to the user at the latest 4 months after the public release of the source code of an update of the underlying operating system or, if the source code is not publicly released, after an update of the same operating system is released by the operating system provider or on any other product of the same brand;
> (d)
> functionality updates mentioned under point (a) need to be available to the user at the latest 6 months after the public release of the source code of an update of the underlying operating system or, if the source code is not publicly released, after an update of the same operating system is released by the operating system provider or on any other product of the same brand;
> (e) an operating system update may combine security, corrective and functionality updates.
EDIT: I see that was already brought up multiple times when you posted about it 11 months ago.
either way if that is true it looks like a very intentional abuse/misinterpretation of existing regulations
> security updates or corrective updates mentioned under point (a) need to be available to the user at the latest 4 months after the public release of the source code of an update of the underlying operating system or, if the source code is not publicly released, after an update of the same operating system is released by the operating system provider or on any other product of the same brand;
https://eur-lex.europa.eu/eli/reg/2023/1670/oj/eng#anx_II
So by delaying AOSP releases by 4 months, Google is allowing OEMs to leave their users exposed to known vulnerabilities for up to 8 months.
The second part explicitly addresses this, Pixels getting QPR1 starts the timer.
My understanding is as follows: They're delaying source code releases for security updates, but they're privately sharing source and binary security updates with partners. Source releases are embargoed, binary releases aren't (I assume it's to avoid triggering the ecodesign clock)
When patches go out to partners in October, they are immediately released by GrapheneOS for users who opt-in for "security preview" releases. Stock Pixel users might get them in December (triggering the ecodesign clock), and other brands could get them as late as April (per ecodesign requirements).
So the timeline could look like:
June: Vulnerability discovered
September: Vulnerability patched
October: Source and binary patch sent to partners. Patch can now be reverse engineered and actively exploited.
October, shortly after: GrapheneOS releases an update with the binary patch
December: Vulnerability is officially part of the December bulletin, Pixel officially gets the update
January: Samsung releases the update
April: Everyone else (legal worst case)
Not completely. Android Security Bulletins only have fixes for vulnerabilities marked high/critical in resent years. To get all other security fixes, you need QPRs and major releases.
- 45 kWh per household per year
- 11.7 TWh of electricity in 2020 EU-wide
Literally just from better designed consumer vacuum cleaners.
https://energy-efficient-products.ec.europa.eu/product-list/...
I'm not familiar with the concrete wording of the regulation, but I would wager that the all the customization layers (like "One UI") that OEMs put on top of AOSP is the component for which the update clock should tick, and they should be unaffected of an underlying open source component (AOSP) being released.
The only interpretation that matters is that of the Authority of enforcement and judgement. The implementer’s interpretation doesn’t hold much weight.
We can say what we earnestly believe the regulations mean, and how they can be met, but if the authorities say “not good enough” that’s all that matters.
And why should a regulation in EU affect rest of the world. Or for that matter, any US regulation. The arrogance of these nations and the ignorance of tech companies is ruining the entire Internet.
It's simple economics. The companies are also allowed to either do special versions to comply with local laws or even to pull out of the market. They just choose not to for reasons.
But surely, unless the regulation is written in an especially thoughtless way, the clock would start at either the time the device manufacturer has access to the upstream fix, or the time that the manufacturer provides an update to any of their devices (which implies access to an upstream fix).
So public source release would only impose additional burden on device manufacturers if it is the first time that they are seeing the upstream fix.
Right?
See https://news.ycombinator.com/item?id=50112059 for a more detailed explanation.
Not surprisingly, I get slight hints of Claudism in the text.
AI "opening" the source of software is one of the best things I've seen going for it so far.
Who beside the open source community develops new features if they can be easily copied?
That may well be the future for most commercial software (and software engineers) as margins collapse and AI tokenomics moves more power from labor to capital.
But why invest time and money when your features get copied as soon as you publish it?
Because you still get money?
Nothing about getting copied means you don't still get money. It probably just means your legal team has some people to annoy into going away if you're a big company.
The only thing that made commercial software more innovative is most developers never had time to build and maintain a big open source project without support from their job. Now we have agents and it takes a lot less time to implement your ideas. That means slop filling up GitHub, but also that people who came up with good new features can make them real in their free time.
ReSkate is a good example, it's a project 90% coded by AI but the people working on it are passionate and most of them are experienced game modders. They know what made the old Skate games special and wanted to bring it back to the new one. Now it's better than vanilla Skate 4 and has tons of new features never seen in the OG trilogy. That project would never be possible without Claude taking care of the hard and boring RE work.
But I guess they will also think twice if big companies can rip off the software with clean room AI implementation to circumvent the open source licenses
IMHO it's the best outcome.
Not perfect, of course. All that hype about the vibe-cloned Adobe suite is, today, hype. Next year it might not be, though that depends on questions like "will investment bubble pop?" etc.
But we are at a level where, if you can describe a feature or process well enough to test it, an LLM can develop and optimise for that feature/process. That creates opportunities for businesses to have custom just-for-them business software, which nobody copies because nobody outside the business even bothers to learn exist.
At this point I'm willing to argue that the owner of a device is whoever is legally allowed to log in as root. When I pay for server colocation, I'm the owner - and not whoever runs the building - because I decide what runs on the box, and if they tamper with it they can go to jail.
With phones it's the same, except that you pay Google for the privilege for keeping their device safe, charged, connected to the Internet, and walking it around every day with you, so they can collect better data. The only thing they can't do is snatch it from your hands, but why would they - you're doing so much useful work for them.
And of course if you try to root it, you're breaking a digital lock, which is punishable by up to 10 years in prison and $1,000,000, so the colocation analogy holds there too.
Google should not be allowed to abuse its duopoly position in this way. I recommend EU citizens to report every case they are (potentially) affected by to the EU DMA team.
But to me big tech succeeded also lobbying the EU like they already did in the US, in fact they violate multiple EU rules (GDPR, etc.) without facing any consequence, and when they try to announce they would do something at the other side of the ocean they treat us that they will double or triple tariffs so...
iOS being so difficult to jailbreak means that most apps don't have any verification at all. Even most DRM doesn't even try to verify that the device isn't jailbroken in order to access the highest resolutions or whatever. But if we ever add a first-party option for the device to be modified, then we start running into the Android situation where every app has some invasive root/tamper detection thing. Some apps already have anti-debugging things, but the situation on iOS could be far, far worse.
(I was reading recently about the SEP's behavior after a memory authentication failure -- it completely stops accepting requests until the processor is reset -- so I thought that applied to trusted boot violations as well. Seems not.)
I suppose that also explains why every time I open Settings, the Apple Pay section is missing for a moment before it shows up; the OS is apparently able to function without it. Interesting bit of trivia.
Someone should make a complaint about ePD violation in relation to those. I doubt it's "strictly necessary" for those applications to read that information (if they even fulfill the "at distance" requirement).
At any rate, “developers are disassembling it to find bugs” sounds extreme, but this is something Claude will do almost spontaneously, in my experience. I ran into an issue last week where a library wasn’t doing what I expected, and asked Claude to sort out why; I came back a few minutes later to find Claude tearing apart the library’s disassembly to reveal a mismatch in struct sizes between the library’s expectations and my headers. And it did solve the bug!
But you're right, that report is definitely written by an LLM. Humans don't typically put parenthetical postmodifiers in headings ("Mechanism (from the shipped libc.so)"), but LLMs just love them too much.
We need an Android alternative, something that respects users' freedom.
Alternatives exist. They just need signifiant adoption.
"EU Age Verification Project Mandates Hardware-Bound Attestation"[1]
[1]: https://www.reddit.com/r/Android/comments/1vds3jm/eu_age_ver...
Though I see signal usage creeping upwards. Maybe just my circle
Google can kill them in an instant, they need the AOSP source to survive.
So really, the alternative is proper GNU/Linux (or as I've recently taken to calling it...)
It is as close to GNU/Linux as Inferno was to UNIX System V.
Which is why Termux has to work around what is actually available as public APIs on the NDK.
Termux has to work around security restrictions in a way not dissimilar to how you would need to work around an extremely securely configured SELinux setup on RHEL.
Which is what Termux has to do, just like they have to call into specific Android services via JNI, as the NDK only exposes official NDK APIs, alongside the C and C++ standard libraries.
See also:
https://news.ycombinator.com/item?id=50057296
I look forward to "Linux users and/or developers benefit from Android but..."
What does "Android is Linux" mean to you? It uses Linux? Why does it matter?
You asserted further up that "Android uses Linux but it has no consequences for users or developers."
That simply isn't true.
The alternative are other AOSP-based OSes like GrapheneOS and Lineage. Once they are used widely enough and funded enough, they could fork AOSP when needed. In some sense they have already soft-forked in that they use major and QPR2 releases, but put a lot of patches on top to fix issues, improve security, etc.
GrapheneOS has an estimated 500,000 active installs now, apparently LineageOS has 1.4M active installs. Probably around two orders of magnitude more than traditional Linux desktop-based phones, which have been trying since, checking notes, Maemo in 2005?
They are mostly relying on LineageOS now and putting their own cloud, microG, and App Store app and branding on top of Lineage. Even now they are not able to roll out Android releases and WebView updates timely. Them maintaining an AOSP fork would probably end up in an outdated, broken mess.
My money is on GrapheneOS and maybe LineageOS to do it when needed.
I have old Chromebooks and Nexus Tablets that are EoL, but I can install linux on them and give them a second life.
The source code for my old EoL pixel phones is out there, it's just a matter of effort to port Ubuntu or PureOS or one of the other Linux phone distros to them.
It's Apple that is being as evil as possible when it comes to stomping down alternatives, they don't release iOS source code, they bricked most of my old devices when they EoL'd them by not letting me unlock the bootloader and install my own OS, they don't release drivers.
I'm deeply unhappy with Google's current direction, but they're still lightyears from the level of evil Apple inflicts upon us.
But that's not the case for other manufacturers, including ones way more popular than Google. Especially Samsung.
On average, a new Android device is almost as locked as iPhone is.
As per september 2026: https://developer.apple.com/news/upcoming-requirements/
Why does that matter? Even if it was released 20 years ago, why the hell should Apple (or any company for that matter) dictate if a developer can make an app for an old device or that a user wishes to use that app!? With Android, the developer doesn't need to depend on the "App Store", or even care about it, to distribute their app - any Android user can directly download and install the app through other means on their old or "outdated" device.
s/lightyears/5 years/
This is the TiVo situation. Does open-source require allowing third parties to install their own software / OS and have unlimited control over the device? (GPL v3)
Or does it regard to access to the source code but not access to the device? (GPL v2)
TV’s, routers, IoT, etc vendors aren’t even allowed to ship devices with user-alterable software!
The more successful Android alternative - iOS - that people love so much doesn't do any of that and it's apparently even more secure and earns even more money.
See post downbelow: https://news.ycombinator.com/item?id=50083579
So why? :)
You can support independent Android development (for example, GrapheneOS)
Isn't one of the main limitations to more open mobile platforms due to opaque drivers? The very recent trends in reverse engineering might provide some solutions to that.
Android has ~4b cumulative users, iOS 2.5b. Combined 1-2b phones are shipped annually, so phones are replaced over 3-6 years or so. Total production numbers of Sony PS2/Nintendo Switch are ~160m each, and there haven't been enough of those to be a viable alternative consumer platforms. China moves couple hundred million units per year of de-Googled phones internally(Google services are default banned in China except through Hong Kong VPN and such), and that doesn't seem to be changing situations much. Windows Phone was like few hundreds total.
The third option has to ship like 1000m - 1b units total and/or annually to be viable. Code quality... isn't going to matter, but it'll have to be literally 10-100x more popular than Nintendo. That is going to be the bar.
The bar for what? The parent talks about the start. I already daily drive Librem 5 without waiting for "the bar".
My point is that they need to be adopted at scales that almost dwarfs Nintendo, in the realms of billion devices in circulation(why does that phrase remind me of a certain purple-blue-yellow UNIX company...)
Google removed Pixel support from AOSP with the release of Android 16. They did so by no longer pushing the device trees, kernel drivers, userspace drivers/HALs and Pixel exclusive releases. That means there are no longer device branches pushed to AOSP for the OS releases used by recently released devices or variations from the standard releases. The monthly and QPR1/QPR3 are now considered fully Pixel exclusive and no longer pushed to AOSP either. That's why this code isn't pushed to AOSP. The remaining non-Pixel-exclusive releases of Android are still pushed to AOSP at launch.
Google is abusing their control over Android to give Pixels inherent advantages over other Android devices. However, in doing so they've also harmed the value prospect of Pixels by taking away official AOSP support which has made it much harder to support Pixels for other operating systems and even for app developers.
Pixels are now on Pixel exclusive QPR1 and QPR3 branches for half the year without sources available. For the other half of the year, they're on the same major yearly and QPR2 branches as other devices. Android 17 QPR1 is the latest major release for Pixels but Android 17 is still the latest for non-Pixel devices. Android 17 QPR2 will be the latest for both in December 2026.
They're not pushing the monthly, QPR1 and QPR3 releases to AOSP with a delay but rather not at all since those are only for Pixels and AOSP doesn't support Pixels anymore.
NDK seems to be handled like a 20% project in terms of allocated resources, I bet it has not more that like 2 FTEs, given its progress since Android 2.1.
That's the huge, huge carrot that Google carries in front of everyone thinking to fork Android.
Google does have a card to play here, I don't disagree but they shouldn't overstate their hand either. The Google bits of Android are good enough to keep open-source ROMs and small manufacturers away but they aren't a silver bullet and won't save them if China stops playing the game.
I highly doubt that. People have far too much money already invested in Play Store purchases, that's a lot to give up for a marginally cheaper phone.
Also, I don't necessarily think the phones will be marginally cheaper, just that the market will adapt.
Though given how little those companies have respected their users so far, I feel like this is a bit of a monkey's paw.
In a way, it's similar to Samsung's Tizen, except Huawei's OS sees actual use, though not so much in the ecosystem of the USA of course.
Man, the early '10s were wild, man.
just hang up the hat and never look back
Both aren't.
I would argue though that even Samsung phones give more user control than iOS phones (not that I would buy either one). Even though you can't install your own remix of Android, you can disable or remove most of what you don't like and install anything that targets the Android APIs.
Unlike on Android, there aren't that many third party iOS apps, though.
Plus, even with Google delaying shipping source code, at least they ship the source code. The traces of source you get with the open Darwin code are nowhere near comparable.
Doesn't apple still "verify" (and can decide not to allow) those supposedly 3rd party applications?
EU users can also add alternative app stores to their iOS device, e.g. Epic Games Store, AltStore, or some corn app.
Being able to have a fully functional version of KDE Connect would also be a requirement.
Check out Blink Chrome on iOS on Agelia developer Gyuyoung’s Weblog: https://blogs.igalia.com/gyuyoung/2024/08/08/chrome-ios-brow...
The Google has documented how to compile your own build. Note the ‘building blink for iOS’ section.
It’s as easy as adding ‘use_blink = true’ to ‘ [gn_args]’ in your ‘.setup-gn’: https://chromium.googlesource.com/chromium/src/+/main/docs/i...