Article says you cannot side load books on Apple Books. That is incorrect. You just send an epub to books via the share menu on Mac or iOS and it loads it. Also syncs it via iCloud if you want it to.
Perhaps the author meant you cannot download purchased books off of Apple Books.
I’m surprised and happy that SUSE is still doing well. I have fond memories of using SUSE in the enterprise especially around their “perfect guest” campaign for using it in virtualized environments. I thought they had very well-baked integration with large Windows networks—things just worked out of the box that didn’t with RHEL. I’m sure a lot has changed in the last decade but I appreciated their cooperative stance in the enterprise.
I would feel that it would be a reasonable if it was my local paper running the story. Arstechnica IS a primarily technical news site—I believe they should have a higher bar—otherwise they are just parroting a report and not providing useful (to me) news.
I generally think arstechnica.com does a decent job of being a non-garbage news site. I pay a couple bucks a month for the ad-free RSS feed. This story feels terrible to me. I don’t doubt a law suit has been filed, but I would expect some investigation by the reporter of the extra-ordinary claims of privilege escape the application is claimed to be capable of.
Surprisingly, I thought the article was a reasonable summary of the actual paper. I think some people might think this was a poke at privacy on Apple, but it really focused on how hard it is to create accessible settings despite the enormous number of options.
I have found that navigating the menus in Apple iOS is quite a bit easier than on my Android devices. Mac seems more difficult as the settings tend to be inside the individual apps and don’t surface as well through the search.
The paper hammered home the point that Siri configurations were particularly hard, but they also mention that Siri data is end-to-end encrypted. I thought all those points were fair.
I do believe settings need to be improved, but I have little faith they will ever be useful for 99% of users who will simply never change anything from the default. At this point I believe any meaningful improvements for the majority of users will come from useful defaults that include E2E encryption on basically all user data. I feel Apple is coming close with iCloud Advanced Data Protection that was introduced last year, but that needs to become a default. Maybe it cannot though—too many users will lose all their data and then the trade off of security to convenience will not be worthwhile.
I don’t think a big business should have an advantage over a small business that cannot afford that technology while using public airwaves. A better solution imo would be to prioritize all very low-bandwidth traffic.
I agree that decrypt/encrypt is bad—it is simply not E2EE. The solution would have to be a better method of public key distribution for ‘federated’ systems.
While I don’t know anything specific about facebook messenger, E2EE doesn’t necessarily preclude what you suggest. A messaging service could store the entire chat history encrypted without decryption keys. When you get a new client you could restore the entire history in encrypted form onto your device. You would then use a recovery key you would possess to decrypt the message history on your end. At no time would the messaging service have the keys to decrypt. I’m not saying that is what facebook does.
Why do you need to control both ends for E2EE? Both ends need a public and private key to encrypt and decrypt messages. You need a method of key exchange. I would prefer to have an offline method (phone call, in-person) of validating a key (like iMessage and Signal have). But I don’t see a reason to need to control both ends.
I get not wanting to use a google, microsoft or crypto laden browser, but I would be willing to use a well supported browser that used chromium as the page rendering engine. It seems to be extremely difficult to get another engine to be competitive in the marketplace. Maybe the resources would be better spent putting the chromium engine inside a different container. I’m sure there would be drawbacks, but I think there would be compatibility benefits too.