3 ms·

It is frankly nothing short of amazing that Apple ships things like homomorphic encryption, and differential privacy, and client-side vectorization, and encrypt

It is frankly nothing short of amazing that Apple ships things like homomorphic encryption, and differential privacy, and client-side vectorization, and encrypted vectors, at the scale that they inhabit... and they still get a bad report card back from consumers about privacy. Comparing Apple and Google, or Apple and Microsoft, it seems to me that Apple's track record on these issues is actually not as bad as public opinion might suggest. Meta doesn't even make the list for comparison, and neither does Amazon. It makes me wonder if picking privacy as a strategy is workable in the first place. People trust TLS; people use banking apps on their phone now without thinking. I remember in 2008 or so when people still didn't quite trust SSL. I'm not sure if Apple will be able to bridge the gap here, though, if _all_ of their competition simply chooses not to ship those features. Do customers know the difference? Do they... care? In theory they want their data to be private, yes. But if they are not willing to educate themselves to perform their counterparty obligation in fulfilling "informed consent," there may be no action Apple could take to avoid catching the same bad rap everyone else does.

jchw 2y ago
> It is frankly nothing short of amazing that Apple ships things like homomorphic encryption, and differential privacy, and client-side vectorization, and encrypted vectors, at the scale that they inhabit... and they still get a bad report card back from consumers about privacy. Personally, I don't shy away from criticizing Google, but that's not the point. Apple makes big claims about their privacy practices that neither Google nor Microsoft make, it would be bizarre to hold Google or Microsoft to claims and standards that Apple set for themselves. > Comparing Apple and Google, or Apple and Microsoft, it seems to me that Apple's track record on these issues is actually not as bad as public opinion might suggest. Meta doesn't even make the list for comparison, and neither does Amazon. > It makes me wonder if picking privacy as a strategy is workable in the first place. People trust TLS; people use banking apps on their phone now without thinking. I remember in 2008 or so when people still didn't quite trust SSL. > I'm not sure if Apple will be able to bridge the gap here, though, if _all_ of their competition simply chooses not to ship those features. Do customers know the difference? Do they... care? In theory they want their data to be private, yes. But if they are not willing to educate themselves to perform their counterparty obligation in fulfilling "informed consent," there may be no action Apple could take to avoid catching the same bad rap everyone else does. I've said this elsewhere, but what I really dislike about Apple's strategy regarding privacy is that they treat privacy as a purely technological problem that could be solved if only we had better technology, but they ignore that a huge component of why users have been subjected to so many flagrant privacy violations is because they have zero visibility and zero real control over their computing experiences. Apple would very much like to retain their iron grip on what users are allowed to do on their platforms, as they make a ton of money off of their control of the platforms in various ways, so they have a huge incentive to make sure we're all arguing about whether or not Apple is better or worse than Google or Microsoft. Because sure, I do believe that if we held Google to the same privacy standards that Apple currently has, it would probably kill Google. However, if Apple and Google were both forced to give more transparency and control to the users somehow, they'd both be in a lot of trouble. Despite all of this effort, I think that user trust in the privacy of cloud computing and pushing data out to the internet will only ever go down, because attacks against user privacy and security will only ever continue to get more and more sophisticated as long as there are resourceful people who have a reason to perform said attacks. And there certainly always will be those resourceful people, including in many cases our own governments, unfortunately.
- sgammon 2y ago
  “Ask App Not To Track” would like a word
  - jchw 2y ago
    Not impressed. I'd be much more impressed if we could run software like Little Snitch on iOS, or install Firefox. Or even just side load apps without pay $100/year. (Note: a Safari webview with a Firefox logo on it does not count.)
    - sgammon 2y ago
      > I'd be much more impressed if we could run software like Little Snitch on iOS You can; or, at least, the APIs are available for this, and have been for some time. > or install Firefox https://apps.apple.com/us/app/firefox-private-safe-browser/id989804926 https://apps.apple.com/us/app/firefox-private-safe-browser/i... Inb4 "you cannot replace the built-in browser engine" https://developer.apple.com/documentation/browserenginekit https://developer.apple.com/documentation/browserenginekit > Or even just side load apps without pay $100/year Then use an Android. Apple has taken a stance on this since it was formed. If you don't know by now that Apple is actively hostile toward hobbyists with their approach to computing, I can't help you. That is why there are alternatives from Google, Microsoft, Amazon, and so many others. To others (me), this is a feature. I would rather the $100/yr fee exist for several reasons.
      - jchw 2y ago
        > You can; or, at least, the APIs are available for this, and have been for some time. Why is there no citation on this one? iOS has a (fairly limited) VPN API but it is woefully insufficient to make an application-level firewall and I don't think any exist. > https://apps.apple.com/us/app/firefox-private-safe-browser/i https://apps.apple.com/us/app/firefox-private-safe-browser/i... That's Safari in a trenchcoat. > Inb4 "you cannot replace the built-in browser engine" > https://developer.apple.com/documentation/browserenginekit https://developer.apple.com/documentation/browserenginekit That requires an entitlement that A.) is only available in the EU for apps distributed in the EU and B.) as far as I know, has never been granted to anyone. Mozilla for example does not have it. If it weren't for EU regulations we would probably still not even be able to change the default browser in iOS, so realistically it appears regulating Apple works great. It's hard to argue against regulating Apple while also defending it by showing the fruit bore from regulating them. > Then use an Android. I own both iOS and Android devices for what it's worth. Anyway... > Apple has taken a stance on this since it was formed. Since... it was formed? Really? I don't remember the Apple II requiring a $100/year developer license to "side load" software. > If you don't know by now that Apple is actively hostile toward hobbyists with their approach to computing, I can't help you. This is a frankly insane thing to say to someone who is in middle of criticizing Apple for this exact hostility. Of course framing it as a developer problem is a convenient way to ignore that developers are users and users are developers, and that these restrictions also have negative impacts on even users who aren't developers. > That is why there are alternatives from Google, Microsoft, Amazon, and so many others. Apple is the same company that ran a tirade about how Android was a stolen product, so I find it amusing that in just ten years it has become the crutch to defend any and all bad practices Apple holds dear. Don't like it? Just use something else! Of course, I do use other devices, but I can't magically not be impacted by Apple's persistent brain-death. For years I have had to deal with the utter stupidity that is the fact that Apple refused to support patent-unencumbered video formats like WebM, and devices that can't play WebM natively and in Safari are still plentiful in the wild. So I can't just pretend Apple doesn't exist. Not that it matters: I am free to criticize Apple however I want, even if Apple fans do not like it. Even just because I feel like it, but even moreso as a paying customer of Apple, and a developer who has dealt with the impact of Apple being involved in the ecosystem. (Thank you Apple, for refusing to support SPIR-V in WebGPU, we love having an additional standard for something for basically no reason.) > To others (me), this is a feature. I would rather the $100/yr fee exist for several reasons. I know. I don't think highly of this position, but I am well aware of it.
Klonoar 2y ago
Consumers seem to think well of their pro-privacy standpoints, even if the devil is in the details with regards to how effective it might be. The people giving “poor reports” are often the hardcore tech users (or LARPers) who grew up in a different era and mindset and are slowly being shoved out.
aleden 2y ago
How did they do it? The idea for homomorphic encryption has been around a long time, but as far as I am aware the time+space requirements have found to be tremendous, thus rendering any serious implementation of it impractible. If they managed to do it, they should open-source the code. If they made a research breakthrough, they should publish. Doing either of those things would give me the peace of mind that I need.
- sgammon 2y ago
  https://www.swift.org/blog/announcing-swift-homomorphic-encryption/ https://www.swift.org/blog/announcing-swift-homomorphic-encr... https://github.com/apple/swift-homomorphic-encryption https://github.com/apple/swift-homomorphic-encryption https://machinelearning.apple.com/research/homomorphic-encryption https://machinelearning.apple.com/research/homomorphic-encry... https://www.apple.com/child-safety/pdf/Apple_PSI_System_Security_Protocol_and_Analysis.pdf https://www.apple.com/child-safety/pdf/Apple_PSI_System_Secu...
  - aleden 2y ago
    Fantastic links. They should be at the top of this thread. This stuff looks really impressive to me. I'm not an expert. I take it that it is implicitly assumed that code running on the server (which I cannot spot the source for) doesn't need to be available, to be trusted? (That there's no way you could come up with an attack through cryptanalysis, involving what the server does, since the security of the data coming from the client is all that matters?)

jchw 2y ago

> It is frankly nothing short of amazing that Apple ships things like homomorphic encryption, and differential privacy, and client-side vectorization, and encrypted vectors, at the scale that they inhabit... and they still get a bad report card back from consumers about privacy. Personally, I don't shy away from criticizing Google, but that's not the point. Apple makes big claims about their privacy practices that neither Google nor Microsoft make, it would be bizarre to hold Google or Microsoft to claims and standards that Apple set for themselves. > Comparing Apple and Google, or Apple and Microsoft, it seems to me that Apple's track record on these issues is actually not as bad as public opinion might suggest. Meta doesn't even make the list for comparison, and neither does Amazon. > It makes me wonder if picking privacy as a strategy is workable in the first place. People trust TLS; people use banking apps on their phone now without thinking. I remember in 2008 or so when people still didn't quite trust SSL. > I'm not sure if Apple will be able to bridge the gap here, though, if _all_ of their competition simply chooses not to ship those features. Do customers know the difference? Do they... care? In theory they want their data to be private, yes. But if they are not willing to educate themselves to perform their counterparty obligation in fulfilling "informed consent," there may be no action Apple could take to avoid catching the same bad rap everyone else does. I've said this elsewhere, but what I really dislike about Apple's strategy regarding privacy is that they treat privacy as a purely technological problem that could be solved if only we had better technology, but they ignore that a huge component of why users have been subjected to so many flagrant privacy violations is because they have zero visibility and zero real control over their computing experiences. Apple would very much like to retain their iron grip on what users are allowed to do on their platforms, as they make a ton of money off of their control of the platforms in various ways, so they have a huge incentive to make sure we're all arguing about whether or not Apple is better or worse than Google or Microsoft. Because sure, I do believe that if we held Google to the same privacy standards that Apple currently has, it would probably kill Google. However, if Apple and Google were both forced to give more transparency and control to the users somehow, they'd both be in a lot of trouble. Despite all of this effort, I think that user trust in the privacy of cloud computing and pushing data out to the internet will only ever go down, because attacks against user privacy and security will only ever continue to get more and more sophisticated as long as there are resourceful people who have a reason to perform said attacks. And there certainly always will be those resourceful people, including in many cases our own governments, unfortunately.

sgammon 2y ago
“Ask App Not To Track” would like a word
- jchw 2y ago
  Not impressed. I'd be much more impressed if we could run software like Little Snitch on iOS, or install Firefox. Or even just side load apps without pay $100/year. (Note: a Safari webview with a Firefox logo on it does not count.)
  - sgammon 2y ago
    > I'd be much more impressed if we could run software like Little Snitch on iOS You can; or, at least, the APIs are available for this, and have been for some time. > or install Firefox https://apps.apple.com/us/app/firefox-private-safe-browser/id989804926 https://apps.apple.com/us/app/firefox-private-safe-browser/i... Inb4 "you cannot replace the built-in browser engine" https://developer.apple.com/documentation/browserenginekit https://developer.apple.com/documentation/browserenginekit > Or even just side load apps without pay $100/year Then use an Android. Apple has taken a stance on this since it was formed. If you don't know by now that Apple is actively hostile toward hobbyists with their approach to computing, I can't help you. That is why there are alternatives from Google, Microsoft, Amazon, and so many others. To others (me), this is a feature. I would rather the $100/yr fee exist for several reasons.
    - jchw 2y ago
      > You can; or, at least, the APIs are available for this, and have been for some time. Why is there no citation on this one? iOS has a (fairly limited) VPN API but it is woefully insufficient to make an application-level firewall and I don't think any exist. > https://apps.apple.com/us/app/firefox-private-safe-browser/i https://apps.apple.com/us/app/firefox-private-safe-browser/i... That's Safari in a trenchcoat. > Inb4 "you cannot replace the built-in browser engine" > https://developer.apple.com/documentation/browserenginekit https://developer.apple.com/documentation/browserenginekit That requires an entitlement that A.) is only available in the EU for apps distributed in the EU and B.) as far as I know, has never been granted to anyone. Mozilla for example does not have it. If it weren't for EU regulations we would probably still not even be able to change the default browser in iOS, so realistically it appears regulating Apple works great. It's hard to argue against regulating Apple while also defending it by showing the fruit bore from regulating them. > Then use an Android. I own both iOS and Android devices for what it's worth. Anyway... > Apple has taken a stance on this since it was formed. Since... it was formed? Really? I don't remember the Apple II requiring a $100/year developer license to "side load" software. > If you don't know by now that Apple is actively hostile toward hobbyists with their approach to computing, I can't help you. This is a frankly insane thing to say to someone who is in middle of criticizing Apple for this exact hostility. Of course framing it as a developer problem is a convenient way to ignore that developers are users and users are developers, and that these restrictions also have negative impacts on even users who aren't developers. > That is why there are alternatives from Google, Microsoft, Amazon, and so many others. Apple is the same company that ran a tirade about how Android was a stolen product, so I find it amusing that in just ten years it has become the crutch to defend any and all bad practices Apple holds dear. Don't like it? Just use something else! Of course, I do use other devices, but I can't magically not be impacted by Apple's persistent brain-death. For years I have had to deal with the utter stupidity that is the fact that Apple refused to support patent-unencumbered video formats like WebM, and devices that can't play WebM natively and in Safari are still plentiful in the wild. So I can't just pretend Apple doesn't exist. Not that it matters: I am free to criticize Apple however I want, even if Apple fans do not like it. Even just because I feel like it, but even moreso as a paying customer of Apple, and a developer who has dealt with the impact of Apple being involved in the ecosystem. (Thank you Apple, for refusing to support SPIR-V in WebGPU, we love having an additional standard for something for basically no reason.) > To others (me), this is a feature. I would rather the $100/yr fee exist for several reasons. I know. I don't think highly of this position, but I am well aware of it.

Klonoar 2y ago

Consumers seem to think well of their pro-privacy standpoints, even if the devil is in the details with regards to how effective it might be. The people giving “poor reports” are often the hardcore tech users (or LARPers) who grew up in a different era and mindset and are slowly being shoved out.

aleden 2y ago

How did they do it? The idea for homomorphic encryption has been around a long time, but as far as I am aware the time+space requirements have found to be tremendous, thus rendering any serious implementation of it impractible. If they managed to do it, they should open-source the code. If they made a research breakthrough, they should publish. Doing either of those things would give me the peace of mind that I need.

sgammon 2y ago
https://www.swift.org/blog/announcing-swift-homomorphic-encryption/ https://www.swift.org/blog/announcing-swift-homomorphic-encr... https://github.com/apple/swift-homomorphic-encryption https://github.com/apple/swift-homomorphic-encryption https://machinelearning.apple.com/research/homomorphic-encryption https://machinelearning.apple.com/research/homomorphic-encry... https://www.apple.com/child-safety/pdf/Apple_PSI_System_Security_Protocol_and_Analysis.pdf https://www.apple.com/child-safety/pdf/Apple_PSI_System_Secu...
- aleden 2y ago
  Fantastic links. They should be at the top of this thread. This stuff looks really impressive to me. I'm not an expert. I take it that it is implicitly assumed that code running on the server (which I cannot spot the source for) doesn't need to be available, to be trusted? (That there's no way you could come up with an attack through cryptanalysis, involving what the server does, since the security of the data coming from the client is all that matters?)