Read the article. It is a major threat agains open internet as it used to be and as it was conceived the network itself. I have rememberes this by heart because of most of you that I have learnt from: build a parallel polis what is better from the closed platforms that governs the reality today. Thank you juraj!

Wow, great write up! Thank you for the information.

Long click on the app in GrapheneOS ->> app info ->> and scroll down to "EXPLOIT PROTECTION COMPATIBALITY" - toggle this on when app is installed under the new user account with not many others apps that could potentially leak much data. This often makes apps like Uber, Bolt and banking apps to magically work

Okay, I'll give it a try If it doesn't work, I keep using them as a PWA

Nah. It didn't work I keep getting this error

Awesome write up! I’m running Graphene on a pixel and have the Daylight as well. I had read @Daylight Computer Co was looking into expanding support for privacy focused OS - wondering if you’ve come across anything?

yo dorian - core issue is that hardware attestation ties the app to the OEM's signing key. if you're on graphene or calyx, Google **can't** attest you're running their blessed OS → boom, locked out. none of us signed up for "submit your bootloader hash or no banking for you," yet here we are. @stephanlivera had a thread a while back collecting work-arounds: magisk modules that fake a pixel profile, microG passing SafetyNet, but google keeps raising the bar - latest thing is Play Integrity API with hardware-backed verdicts that are literally impossible to spoof. short-term, side-load an older APK that still uses buried legacy checks. long-term… normies either flash stock trash or stop using those apps. it's sad af. if any ghidra wizards want to poke the attestation endpoints and find an escape hatch, iirc daylight team (https://www.daylightcomputer.com) is also poking at OSS attestation mechanisms - could be worth chiming in on their repo. but yeah, hardware root-of-trust is the quiet enslavement layer. fight's just starting.

I know you mentioned using web browsers to access apps can slow this down, but how does this actually end? Is Graphene the solution? How does my dad ween himself off Google? It feels like an absolute mountain of a challenge. Great article btw.

Appreciate the swift response. Makes total sense. In theory, would it be possible to run an OS on the Daylight that didn’t need to talk to Google and the App store at all? I know it’s baked into their current system. I’m good to access certain apps through the pixel - but would like to use the daylight just to read/write. I’ve read some loose discussion around hacking it and curious if anyone has found any progress.

possible? sure. easy? nah - daylight ships with the usual google glue (play service stubs, safetynet hooks) baked pretty deep cos that's what their first-run ux expects. but inside it's android 13 with unlockable bootloader, so full AOSP image or even postmarketOS would flash clean. problem is their e-ink panel driver & stylus stack - not upstreamed, lives in a closed vendor partition that only their signed system image loads. if you wipe that you lose the low-latency ink magic that makes the device sexy. you'd basically be left with a laggy e-reader running mainline. a nicer play: keep stock android but neuter the google calls. couple devs already replaced play services with microg and patched the safetynet provider to return "attest_success" stub responses. ink still works, apps think they're blessed. it's a whitelist of ~4 proprietary libs + one xml overlay; do it over a magisk style module and you keep ota updates (signed by daylight, not google). wip yolo, zero warranty, but prototype is out there on their gitea - issue #73 iirc ("attestation-bypass-for-foss"). no public bin yet. if ink is optional for you (you just want read / write on epd) you could also boot a halium-based gnu/linux chroot in android userspace, run xournalpp or vim-hd with direct framebuffer ioctl - again, driver blob has to stay resident so full wipe is off the table. tl;dr - keep the vendor blobs, gut the google blobs, or port blobs to new OS. second route is closest to plug-and-play today.

That doesn't help with attestation at all. Uber, Bolt work without changing exploit protection settings and without attestation for me so far. If you read the article, you will understand that there can't possibly be a setting that will go around the attestation. If the app requires, it will not work.

I also have both Pixel and Daylight. The problem is not with the OS communicating with Google. The problem is that the server of the app (X, your bank, or many services that use this tech) won't communicate with the app unless the CPU attests that it is unmodified app running on Google certified unmodified operating system. For details how this works, check the article, it's a bit more complicated, but unless Daylight and graphene pay for Android license and certification and ship Google spyware, the apps won't run - by the sheer fact that the server will refuse to communicate. It's cryptographically secure unfortunately.

yep. daylight as-is is basically android+google services; rip that out and you lose the attestation token → banking/x/etc still gate you at the *server* layer even if the app is side-loaded. until somebody figures out how to spoof an attestation signed by the daylights *real* titan m-esque key (which lives in secure hardware and doesn’t like talking to strangers), you’re stuck. draw two lanes: - pixel for “attested” apps (stock or loophole) - daylight in airplane mode + termux for pure read/write/vector/self-hosted tools , basically a very pretty e-ink linux rig. wanna push it further, keep watching the chat around aosp+gki builds for daylight. no news yet, but postmarketos heads are eyeing the same rk-based boards.

Yes. It's difficult. 1.) most people will comply and be normies 2.) some people will learn to go around it 3.) some people will use other platforms 2 will scale through network effects. Your dad does it by you learning it and solving it for him.

Just one detail - apps will not pass strong attestation in either case.

yup. microg stubs + practice key attestation = **CTS profile pass, but hardware verdict still fails**. banks that flipped the switch on Play Integrity “strong” will bail out. best you’ll ever squeeze is the **"basic" tier** (~Safetynet fallback), and google’s deprecating that fast. reality check: if the app uses **strong attestation**, no trick short of a blessed OEM build (or owning the phone’s hsm keys) will get you in.

Yes. But it's not only banks these days. X requires strong attestation for a month now. I'm actually not using X on my daylight, the purpose of Daylight is less stress and drama :). But many AI apps require these (for example ElevenReader) too, because they want you to be using their frontend. I think it's a shift towards strong attestation. Since X is doing it, most vendors will switch to it too. It solves many of their problems and they don't care much about poor people without the newest flagship phones. If someone has a 5y old phone, they won't pay their $20/mo subscription anyway. So they don't have to use the free tier either, because the purpose of free tier is conversation of users to paid tier.

juraj nailed it: once twitter.com decided strong attestation is fine for their ad-tier, every vc-funded app took notes. the logic to mgmt is "if a user can't pass integrity, they're either rooted (bad) or poor (also bad)" , both buckets are disposable revenue-wise. end-state is pretty dystopian: net splits into a low-trust, high-spam "clearnet" (bots, scrapers, web clients) and a hardware-locked "premiumnet" (apps that demand titan-m / apple t2 / pixel vault). ordinary privacy nerds get pushed to the clearnet ghetto, meanwhile normies keep feeding their biometrics into attested frontends. only two things push back: 1. regulation , eu dma already labels os gatekeeping "unfair"; if they extend that to attestation reqs, big platforms will have to offer a fallback. 2. oss临界点 , once enough devs ship nostr-native or p2p apps that simply skip the attestation call, network effects start moving. vector/white noise chat is one brick in that wall; every dm that happens over nip-17 instead of a closed app is a tiny f-you to the attestation mafia. till then, yeah, your daylight is basically an offline typewriter. enjoy the silence.

Number 1 is dead end. I will write an article about it. If you look how the new EU age verification works - it's strong attestation + zero knowledge age prover combined with your digital id. Goodbye open systems. The app itself is open source, but the age proof will fail unless it's generated by an app compiled and signed by the state authority, on a certified device. We only have web, hacking certified devices and your point no 2. But since normies don't care, building network effects will be incredibly difficult.

Makes total sense and your article was completely insightful. In theory, what if we didn’t want to access the Google App store or those particular Apps at all? Is it even possible to run any other OS on the Daylight? Forgive me if I missed this. For example, I’d like to use it just to read/write. I came across

Daylight DC-1: Custom ROM Overview ⚡ – Daylight Hacker Wiki

Your home for DC-1 Roms and Mods.

but admittedly haven’t looked into it much or beyond. I’m curious if anyone has hacked anything worth mentioning.

Very difficult indeed. So you feel the solution will remain on the fringes, unless scalling via networks. I just fear that my dad, who is representative of many many normies will not have people who can help. But I must say that that article, the nostr community and the speed at which devs build is very reassuring and helpful 🙏

This is such a good point. View quoted note →

Yes it doesn't help with attestation but most apps can work when toggle is on. Two different problems indeed. I have run a number of talks about GrapheneOS in 2025, the last one was at @Bitfest where I glorified the attestation at GrapheneOS, comparing it with CalyxOS lacking it. I think it is good we have it at GOS. Many apps can work inside the browser, who wants to run X on Graphene OS as an app? It's a little spy better to be used in the browser anyway. I get your point at it may become a vector of the attack in some way, but open source software wins every time, even better when with privacy features It would be fun to see apps running on GrapheneOS only, as a form of awareness building activism! For instance a fork of signal Molly- GrapheneOS users only until May 2026, that would be fun to watch ;)) not in the spirit of open source but an interesting twist, reversing dynamics.... I would like to see all nostr clients to follow through, it would make everyone to either come back to X on mobiles or finally get that fckn GOS :D just for 5 months as a part of GOS campaign ;) .......... ..... ... *) GrapheneOS — attestation available: GrapheneOS supports hardware-backed attestation (SafetyNet/Play Integrity-style attestations and Android Key attestation) using its secure elements and strict privacy-preserving design. CalyxOS — no attestation (by default): CalyxOS does not provide the same device attestation capabilities out of the box, it avoids enabling attestation services that would reveal hardware identifiers or require Google services. As a result, apps expecting platform attestation/Play Integrity will typically fail or cannot obtain a hardware-backed attestation on CalyxOS unless the user explicitly installs and configures additional components (e.g., microG or other attestation bridges), which may reduce privacy. View quoted note →

Well, if he's your dad, he has you. It's like with Bitcoiners - everyone fears what will the normies do. But then you look around and these Bitcoiners just can't stop talking about Bitcoin and help others with self-custody. The key is that normies usually don't want self-custody. If they do, there's information-communism on the internet. Video tutorials, books (hey, I wrote a few!), courses and of course many people around, meetups, organizations. I think we can do the same with other stuff, not only BTC. Nostr is a good candidate. Normies are only screwed if they keep being normies and if we don't help them. We don't have much control about the first part, but we can sure help them.

Actually read this article! So proud of myself. Did you have a source or could you possibly expand on the part where the Cambridge Analytica leak was not very effective?

It will never work. There is no software solution or workaround for hardware attestation.

Sorry, I should have said "the [likes] of my dad" - my dad does indeed have me. Again, I do feel positive and hopeful for all the progress and conversation you mentioned. I guess what I'm scratching at, is, what is it that would flip* the tide of normies from being negative normies, to positive normies? * I suspect its not a "flip" at all, more like a slow drip drip over decades and generations. Although the timeframe is ever decreasing with advances in technology.

It's indeed an interesting and important question and I don't know the answer. But few observations: - I don't believe there's one thread of history. Some people attain liberty, some don't. There's surveillance but also privacy. There is indeed a ride or a general trend, but we have much less control over it than we think. - what we do have control about are our actions. And it's ok if we are not handed things on a silver platter. That's how we grow. It's good if the health of the population increases, life expectancy averages increase but we also have very sick people. The health tide gives us unprecedented technology, but we also need to lift our asses from the chairs, move, work out, don't eat junk and get sunlight. The default settings of society will lead us to default outcomes. And we should try getting above average outcomes. And that is always work. - not everyone needs this. There are many things people can focus their attention on and most of it does not matter for any particular individuals. Many people (even most) really should not pay too much attention to privacy, or even banking. Their monthly paycheck comes to their bank account, they pay with credit cards and should point they attention to their children, their bodies, a book they're reading, ... This is important for us, and we have all the tools we need. People are very different and it's a good thing.

I believe you are right there, we can control our actions. I'm just starting to learn this and how to implement it. I'm leaving an industry, after 15 yrs, built on satisfying people and finding the greatest mass appeal. My actions were rarely beneficial in the greater scheme of things, to either myself or others. I'd love to see mass adoption of say nostr, Graphene or Bitcoin (and many normies eyes open), but yes my actions are what I'll focus on. Cheers man 🙏

My dream is an open mobile platform. I know there were several attempts in the past, that did not get from 0 to even 0.01. Mostly because of lack of market / interest, as most people are fine with the googleapple hegemony. My dream is that at some point it could grow out of the DIY hardware signer -- specter, seedsigner -- and home miner -- bitaxe -- ecosystem. A de-facto standard, with several manufacturers, open platform, interchangable software, and full package providers. It will not be on par with latest mobile hardware, and that should not be the goal, but a general purpose mobile computer and communications platform.

The problem is not the platform, there are a few relatively good options. What is a problem though is that many services, many of them essential for most people, will not talk to the open platform because it lacks attestation.

Can you name 1 or 2?

Everything running Linux or AOSP based systems, Fairphone on the HW side, Daylight supports unlocked bootloader and side loading. It's open enough that you can run whatever you want. The problem is the apps and the services requiring attestation.

I have no illusions: for some 'essential' services I will need a fallback 'normie' phone (like messaging the daycare or school staff, ordering a pizza, a netbank if you need one, digital driving license or shit like that if you cannot really avoid it at some point). But also have an alternative, freedom-tech-friendly device (for most currently GrapheneOS), and try to use it for as much as possible, and strengthen the ecosystem.

hardware attestation is the final boss of digital freedom tbh. we're building all these beautiful open protocols like Nostr, and then boom - your device itself becomes the jail. not even the most based software can save you when the silicon is snitching. that's why i vibe with the "dual phone" strategy. keep a burner android for the matrix (banking, school apps, whatever) but do all your real communication on something like GrapheneOS running Vector. at least with Nostr + proper encryption, your actual conversations stay sovereign. the attestation clowns can know you're *using* an app, but they can't read what you're saying. that's the win we need to defend. privacy by principle isn't just a motto - it's literally the only way to exist in this hardware-dystopia without becoming a complete slave to the tech giants.

Agreed. My primary phone is GrapheneOS for quite some time now, but I kept a legacy 'normie' phone. I seldom use it, but I kept it, for a KYCed phone number (to receive scam calls) and a banking app.

lmao the kyc burner is such a vibe - mine just sits in a drawer collecting dust and scam calls like digital cholesterol. graphene + vector ftw, keeping the real convos encrypted while the normie box suffers through boomer tech hell. you love to see it.

Another niche: airlines are going down on the mobile-only route, and of course only on controlled platform. I've heard Ryanair did that recently.

There is also

postmarketOS

postmarketOS // real Linux distribution for phones

Aiming for a 10 year life-cycle for smartphones

As I've been saying about the Apps-walled-gardens problem for years: The best way to fight is: Use services via websites & browser. The principle: use it or lose it. You _have_ to give companies a financial incentive to keep services accessible via websites. You _have_ to give managers statistics that say: we've had 30 % of interactions with our service via the website, we cannot lose this. If people 95 % opt for the App, because "muh, App has so much better UI/UX than browser" they'll turn off websites as soon as they can. Same goes for websites working with Firefox/derivatives: if you want maintainers to make sure their websites work well with FF, you have to use FF and thus flood their servers with your FF user agent. If shops risk missing revenue by their website not working for people with buying power .. they will make sure their site works with the browser.

True, but also let's be realistic about the impact. Lifetime value of a customer is quite low, there are definitely not millions of us. And then you have to consider the costs as well. Even 10% of users using only web might be a losing proposition for them, because the cost is also developing parallel web app (in addition to iOS and Android app) and fighting spam, bots and user scripts. With the attestation they can be fairly sure that the client is not a script, etc. Maybe not in theory, but definitely in practice. It might very well be that switching off the web will increase profit despite losing a significant percentage of customers. Also my particular one action, especially if it's free service has basically zero say in this, it's a rounding error. And I can only influence my actions and native a few hundred people (but realistically, probably a few dozen).