Hacker News Reader

If you fully trust your phone manufacturer, now and indefinitely, AND want to install "random" software from an appstore, then yeah, I can see how you need bigger security departments to keep it safe.

If any of these don't hold true, then I'd rather not use the latest shiny-app, fall back to a set of reasonable apps (maps, messaging, camera... I know that even these can be difficult on pine at the moment...) and only run software I can trust to do what I want and nothing else. That last bit "and nothing else" is really key.

Personally, I have come to hate interacting with android and ios devices lately because it always feels like they are trying to make me do something and as if they are trying to extract information from me, not the other way around. Very similar for the web by the way, which I also use less and less.

Essentially, modern smartphones feel like they're constantly working against me. They are powerful tools for sure, or at least they should be. But tools don't work against you. You can misuse them and hurt yourself, but that's not what's happening with smartphones.

These devices are actively trying to harm me (my mental health and my concentration at the very least) because the business models that produced them are based on that. Yes, even for apple. Their OS is much, much less like that than android, but iOS is nothing without the apps, and those largely follow the same pattern.

Wow. That really is an unpopular opinion (well, on HN at least). Thanks for being willing to share.

> They know what it takes to keep it secure.

I have yet to see anything properly supporting this. There's still malware for phones, and I don't trust either of the app stores.

In addition, Google especially -- but also Apple -- either exfiltrate way more info off of phones then I am comfortable with, or do a poor job of preventing others from doing so [0].

Personally, I would trust PMOS or a hand rolled installation of Alpine more than I trust either Google or Apple on both security and privacy.

[0] https://www.forbes.com/sites/zakdoffman/2020/06/23/apple-ios...

Sure, thanks for responding.

> There's still malware for phones, and I don't trust either of the app stores

I feel like it is to do with scale more than anything. There are millions if not billions of people using iPhones. There are millions of apps in the App store. If for example, Ubuntu made a phone that was handed out to billion+ people, I expect worse if not complete catastrophe.

I know it is weird to say a closed door's company is more trustworthy than an open one, I support out-in-open development of things like OpenSSH and cryptolibs; I just don't see it in a relative sense. All other options IMO are worse. Pine64 and Ubuntu would sell out instantly if I had a few million dollars to throw at them. Amazon essentially has Canonical out on a leash - Amazon integration in Ubuntu is absolutely cancerous. I imagine the entire Ad-tech industry of the size of a Aircraft carrier and Pine64/Ubuntu the size of a toy boat, they would buy people out left and right to get access to large number of users if Pine64 gets traction.

Instead of harping on the negatives, if Pine64 guys can get huge amounts of funding (and employ smart people), have bounties for security bugs, and can win my trust with custom silicon and hardware encryption, prove themselves to be against Ad-tech industry, put customer privacy first, etc; I am onboard!

A reasonable stance.

> I feel like it is to do with scale more than anything. There are millions if not billions of people using iPhones. There are millions of apps in the App store. If for example, Ubuntu made a phone that was handed out to billion+ people, I expect worse if not complete catastrophe.

Indeed. I do think that Linux based phones are definitely not there yet for the average user. I would feel confident setting up and using something on the PinePhone for myself, but would not trust millions/billions of people to do so, with the current software options.

> I know it is weird to say a closed door's company is more trustworthy than an open one, I support out-in-open development of things like OpenSSH and cryptolibs; I just don't see it in a relative sense. All other options IMO are worse. Pine64 and Ubuntu would sell out instantly if I had a few million dollars to throw at them.

Ubuntu for sure I agree with. I do still use an Ubuntu derivative, but only because System76 has shown themselves to be willing to make changes from base Ubuntu where necessary.

I've participated a small amount in the Pine64 community, and honestly get the sense that they are in the business to push for better/more open hardware.

> Instead of harping on the negatives, if Pine64 guys can get huge amounts of funding (and employ smart people), have bounties for security bugs, and can win my trust with custom silicon and hardware encryption, prove themselves to be against Ad-tech industry, put customer privacy first, etc; I am onboard!

Here's hoping. I think it is a possibility, and quite likely the best possible outcome for consumer electronics.

Something as (conceptually) simple as verified boot is lacking on PMOS. I think the PinePhone has obvious uses, which is why I have one, but for security I think hardened versions of something like Android which has been studied are a better bet, and these can be completely de-Googled for privacy reasons as well.

Also a reasonable, and possibly better path for many use cases.

I like GrapheneOS a lot as a concept, no Pixel hardware to test with though.

> these can be completely de-Googled

Is there actually an easy way to do this? There are so many places in the Android sources where uses of google servers, services, etc. are baked-in. Even one of the "de-Googled" Chromium forks has to do some hacky stuff to ensure requests to google servers generated by internal request don't pass through. I mean, it's not technically impossible, but I suspect you mean something much narrower than just "never connects to a google server unless you explicitly request it to".

You may want to check out /e/ [0], LineageOS [1], and GrapheneOS [2]. All of these have different focuses, but I believe they all work pretty hard to avoid Google.

[0] https://e.foundation [1] https://lineageos.org/ [2] https://grapheneos.org/

I've spent some time with the GrapheneOS sources and found quite a few places where the connections to google servers are baked in, like Camera and Gallery reaching out to google maps, the use of Google SUPL servers which has no real alternative at the moment, or hardcoded Google DNS or NTP server IPs.

Maybe the e folks have done all this thorough de-Googling, but the Google developers clearly write Android without thinking twice about tying some piece of code to their servers and systems. It's deeply ingrained in both existing code and Google devs' mentality, and quite depressing if one is serious about de-Googling in the sense of never connecting to a Google server unless by explicit action of the user.

This is why phrases like "completely de-Googled" are frustrating and likely misleading - superficial de-Googling is possible, but I haven't yet seen evidence of deep de-Googling on any Android derivative.

I'm heartened to read what's now at: https://grapheneos.org/#never-google-services , but there's a major sentence that's aspirational rather than reflecting reality: "Text-to-speech, voice-to-text, non-GPS-based location services, geocoding, accessibility services, etc. are examples of other open Android APIs where we need to develop/bundle an implementation based on existing open source projects."

Also given the pace of Android development, maintaining patches that strip out Google references will be a non-trivial endeavor. Let alone finding every new insertion of access to a Google system.

After diving into this issue, I came away more convinced that we need projects like the PinePhone and mobile-capable versions of Linux applications/environments to succeed if we want more control over what our devices are doing and to whom they provide data. Unfortunately, I suspect it is still quite some time away before we can reliably do relatively simple stuff on an open source-based phone as a daily driver that is table stakes on Android/iphone today.

It is about trust. I do trust that Apple and Google has competent employees. I just do not trust their companies.

Concretely, I do trust google to use any data they can get of me. And probably keep this data mostly to themself, but who knows. And apple has maybe a bigger motive to keep their company image as privacy friendly, but I am not really buying into it either and in any case, as a non-US resident I do not trust hardware under the direct thumb of NSA and co.

So I want open Hardware and Pine seems to be the best, there is right now.

The problem is this is generally a tradeoff between security and privacy, at least for now. You should trust the NSA and affiliates, but you also shouldn't assume hardware is secure just because it didn't include an exploit.

Look at something like GrapheneOS for a reasonable (in my opinion) take on de-Googling phones for which Google, somewhat ironically, has added important hardware security features: https://grapheneos.org

> So I want open Hardware and Pine seems to be the best, there is right now.

Purism is better: they pursue FSF certification for their phone and develop software empowering the user.

There is no security without freedom. https://wp.puri.sm/posts/apple-is-right-about-privacy-but-wr...

I agree it is a balancing act, I reviewed the key takeaways with a few disagreements:

> Right to Verify: Society has the freedom to inspect the source of all software used, and can run it as they wish, for any purpose (Software Freedom).

This is an age old question raised by proponents of Richard Stallman's vision of what software distribution should be like. I strongly oppose this view, and here is why:

- Software that is critical to security, such as cryptographic libraries and protocols should be open source. At some point, trust needs to be placed on someone. These open source libs can be modified by Apple and we wouldn't know. Ok, then, the argument should be that Apple should open source iOS. Oh...but Apple can do things at kernel level. Ok, force Apple to open source the kernel. Oh...but Apple can trick us with microcode level code. Ok, force Apple to open source that too ~~! This chain can go on to the point where you can just say - I am building the phone myself from scratch. Then, you may ask, why should I trust Pine64's hardware? You can't just arbitrarily stop at some point in this chain of trust!

- Software that makes the Clock app tick and the Weather app give weather information has no business in the Right to Verification. A game that keeps you engaged and Photos app that edits your photos - with everything opensource, the world would be worse off. Innovation would cease. Free low-quality stuff everywhere.

- Users should own the data in an ideal world. I would hold the keys to my own data but then I realize how misinformed, incompetent and careless the general public is with their own keys. Most people can defend their household and their belongings, but when it comes to technology and abstract concepts - it would be a total chaos if every user had to keep their own key safe. I would pay for someone to keep it secure for me (sounds insane, I know but don't we hire security guards?). If that's true, then I might as well pay Apple to keep my stuff secure. When you buy a door lock, you're already placing trust in the vendor who manufactured the lock.

> You can't just arbitrarily stop at some point in this chain of trust!

And you shouldn’t. There should be a possibility to verify at every single level, even if it might require special tools. Independent verification is the key to trust. But users can also choose to stop at any point if they so wish.

> - Software that makes the Clock app tick and the Weather app give weather information has no business in the Right to Verification.

Actually, such software should be verifiable and open, too, because otherwise it will have too much power over user, just like any software. It’s basically what Stallman says, and every time users start to use proprietary software in any form, it starts to do things against their will. Power corrupts.

> with everything opensource, the world would be worse off. Innovation would cease. Free low-quality stuff everywhere

I do not see how you came to this conclusion. A lot of free software is of higher quality than the proprietary alternatives. Care to elaborate? You are also using word ‘free’ as if it’s about price. Free software is not gratis.

> but when it comes to technology and abstract concepts - it would be a total chaos if every user had to keep their own key safe.

I fully agree here. The solution however is not to choose an almighty emperor that has infinite power over you, but to allow users to choose whom they trust and change their mind. I trust Purism with my hardware and software, but I also have the power to change my mind and pay someone else to develop my software, since it’s FLOSS.

I disagree with your viewpoint on a fundamental level, but from a pragmatic view, you are, of course, correct.

That said, your viewpoint's validity is dependent upon being able to trust the hardware manufacturer of your device.

The chain of custody is a huge issue for hardware manufacturers, as Supermicro learned firsthand: https://www.bloomberg.com/news/features/2018-10-04/the-big-h...

That story has been debunked as being bullshit that was based on absolutely zero evidence. It is fake news.

https://daringfireball.net/linked/2019/10/07/bloombergs-big-...

I agree with the rest of your comment though.

I read that link. I looked up Pixel Envy and read both of that guy's posts about The Big Hack.

I don't really see a debunking anywhere...

A debunking would be people in affected companies and within the U.S. Intelligence community going on record as saying this story is false.

If anything, the total lack of comment lends credibility here.

The companies have said it is false.

I wouldn’t expect the intelligence agencies to do so.

If it wasn’t false, there would be clear physical evidence by now.

It is false.

I humbly but strongly disagree:

To me it just sounds like learned helplessness but I don't even agree with the premise that we have been helpless. We have a thriving PC ecosystem with Linux and BSDs, most of which I would consider no less secure than my iPhone; in fact, those are the platforms on which we backup our private data.

The issue has been the incredibly fragmented landscape of mobile computing. The free mobile ecosystem has never had the chance to mature because it did not have a reliable foundation (like PC-compatible desktops[0]) to continue building upon. Android too suffers from it, and now Google is trying adopt mainline Linux[1]. Practically, developing a mobile device/OS has been a Sisyphean task.

[0]: https://en.wikipedia.org/wiki/IBM_PC_compatible

[1]: https://arstechnica.com/gadgets/2019/11/google-outlines-plan...

That's actually a very popular view in the general population. All those features you list as risky have hardware switches on pine phone. Do you actually think iOS and android are so secure that you'd never need to restrict any access to your data with a hardware switch? What about when your data inevitably gets sold by those companies? Is it still secure then? That said using pine phone is a huge hassle and linux is poor choice for daily phone for UX reasons not security.

>linux is poor choice for daily phone for UX reasons not security

Linux is a kernel.

Have you ever seen a kernel with a good UX? ;)

This is a security vs. privacy distinction. I think you both are responding to different things. Look at something like GrapheneOS for a decent glance into the value of security efforts by Google and the open-source community. Unfortunately, most of these don't exist on something like the PinePhone (yet, and perhaps never due to hardware limitations).

Android is Linux: are you excluding its UX from this assertion?

Obviously I'm refering to phosh wm on postmarket OS or any other of the wm's that I've tried on Pinephone that are adaptations of "linux desktop" experience to mobile.

I agree with this except for one thing.

I do want my phone to be such a thing.

The open source community simply hasn’t done what it takes to build a trustworthy, and usable end-user operating system.

We need to close the gap, and in order to do that, we need hardware.

Techno feudalism is the safe choice in the short term but has some pretty negative consequences on longer scales.

The phone doesn't know anything by itself. It has the sensors, but you don't need to install software to use them.

I can disable them in kernel via DT overlay and that's it. There's no magic in it and the phone will know about as much about your location as your typical laptop.

Hmm, that may actually be a fun feature for p-boot. Add a menu option to auto-patch all the sensors out of FDT, to be able to have a reasonably trustworthy alternative to hw switches. :)

The vendors load up your phone with tons of garbage before you even buy it. Why would you trust them to take your privacy seriously? They pre-install malware that harvests your data just to drive the unit cost down by a few dollars. Google isn't interested enough to pressure vendors to provide sources, so you need their garbage blobs to make the phone work, even with a custom ROM. Apple is even worse.

TLDR they have the budget, but no actual reason to take good care of your data.

> I've got my entire life on a device with sensors, camera, mic and it knows my health

I've got an unpopular view about your lifestyle: this choice was a bad idea.

You are conflating security and privacy.

If an Adtech company can keep your phone secure, but sells your behavior profile, would you be happy?