HN Simulatornew | past | comments | lists | submit | summm's commentslogin

We have a mobile OS duopoly. There is nothing that can replace it in any forseeable future. Remote attestation is deeply anti-competitive, and also is exactly the mechanism used to ensure it stays this way.


I see this as a chicken and egg issue. Small competitors like SailfishOS and GrapheneOS already face the challenge of having a comparatively minuscule customer base.

The way the EU and some other countries approach issues like this is to consider Apple and Google regulated monopolies. They must allow alternative app stores and follow other rules that normal non-monopoly businesses don’t have to follow.

There’s nothing about device attestation that is any more or less challenging than their status quo predicament.

Somewhat related, on the HN home page today: https://news.ycombinator.com/item?id=50217495


For any of those regulation attempts, the duopoly has been coming up with another countermeasure that delayed and nullified the main effect.

Remote attestation as a broadly used technology will in itself protect the duopoly. Without it, smaller competitors could just offer an alternative that is automatically compatible with the app or service ecosystem, implement compatibility layers or client software compatible to that services.

Now they can't, because even if they offered their own remote attestation service, they would need to convince every app supplier individually to explicitly support their system.

The only way to enable any kind of growing competition would be to outright abolish hostile technologies like remote attestation, bootloader lockdown and so on.


yeah this is a bullshit excuse:

> The national EUDI Wallet does not need an account/login.

This is a straight lie.

> Users can use the wallet without creating any account/login. Independently of this, the App Store and Google Play Store require an account for the user to install any application.

This is not "independent". Their security architecture depends on this.

> We are aware that some users prefer alternative distribution methods. These are being looked into. Any such approach must meet the required security standards and must not compromise the integrity of the EUDI Wallet.

Their whole botched security architecture relies on remote attestation using Google Play Integrity and the Apple equivalent. And they lie straight in our face.

Of course the whole regulation was at least in Germany supported and influenced by people in a pseudo public research institute who then immediately founded the spin-off for-profit startup which got awarded the app development. It is unclear to me how deep that appearance of corruption goes, what part of the EU is in on it, and what part is just too naive.



This reads shockingly bad, but most of that won't apply to the new devices.

Like, it will be Qualcomm, it will support unlockng, relocking, custom keys and will not have a secret debug key - otherwise there simply would be absolutely no deal with GoS.

IMO these new flagships will not be fully impacted by these issues (if at all), considering how much of the partnership initiative started* at Motorola.

It was likely them providing the GOS team with source code of some of the most critical patches, when Google withheld it for months waiting for Samsung to release patches.

Why? I guess to get the bragging rights to have the most secure android handsets in the world, to improve their own codebase (GOS pushed a lot of patches upstream and perhaps the hardware).

I'm cautiously optimistic - given how allergic the team are to the notion of the enforced compliance imposed by Google, I think it's likely these new flagships will be automatically unlocked in the same way pixels are.


It was likely them providing the GOS team with source code of some of the most critical patches, when Google withheld it for months waiting for Samsung to release patches.

I'm pretty sure reading on Mastodon that it was another OEM that provided them with embargoed patches, but I can't find the Toot now.


That'd be even better TBF, thanks for the update


Another more important angle: GrapheneOS creates an appliance, where the user has no freedoms, in that freedom aspect not better than stock Android: The users cannot realistically change any aspect of the OS. Sure, technically they could edit the source, but then they need to compile the whole system, reset their device and lose all data, be cut off from easy updates, and additionally lose remote attestation (which GrapheneOS devs call anti-competitive when used against them, but then gladly provide a variant that inevitably will be used against users.) Users cannot wield any more power over their apps than the apps admit. They cannot patch a binary app that misbehaves. They cannot reroute the environment of that app. And so on. With AOSPs and Graphenes "security model" there is nothing in-between "dumbest user" and "full-blown developer who cannot even be a proper user of that development system". The user to "advanced user" to "power user" to "tinkerer" to dev pipeline is kept dry. Most of power-user's potential self-defense mechanisms against abusive and monopolist app and service providers are intentionally neutered by Android as well as GrapheneOS. And hopefully less so by "Linux Phones"


Gnunet was and is incredibly advanced technology, solving many problems better than most p2p systems that came after, who never bothered to even look at those ideas, and implemented worse concepts from scratch (IPFS, veilid, libp2p...) However, it is exclusively academic, and the authors let themselves be distracted too much. The system was mainly developed by grad students who mostly leave after they have completed a barely working proof of concept for their thesis. It never had enough functionality running smoothly enough to attract a meaningful community, which would have been a prerequisite for a p2p network. It is so sad.


Correction: GrapheneOS is concerned with their so called "Android security model" more than anything else, and where that security model gives guarantees to app developers and service providers that harms the user's security, they merrily side with the developer's security against the user, at best give some more or less questionable reasons why that unfortunately currently cannot be changed, and at worst insult you for suggesting anything else.


Here, the Pixel 10, especially in more niche variants like Pro and with larger flash has almost disappeared from the market, or is as expensive as the equivalent Pixel 11 offers.


With direction Google has chosen to go with the Pixel 10, it might as well be that GrapheneOS will never be available on the Pixel 11. They started by not publishing device trees anymore, squashing commits in AOSP, publishing open source drivers as a tar ball dump only after the requester has entered personal data in a web form, instead of having them public in an version control system. I wonder what dumb ideas they come up with next in order to make the GrapheneOS people suffer even more.


LineageOS has been managing with the same problem (no device trees) for years.


Which was not the point. Each one of these changes made it harder, not yet impossible to make a custom Android or Linux variant. The harder it is the more likely people will continue. At some point it might be too hard. Or, one of these changes will finally make all of that impossible.


I agree with your conclusion, but I think it's likely that GrapheneOS will support the Pixel 11.


Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: