Software and service vendors can't "go around security and privacy protections", they can do exactly what the operating system and Apple allow them to do (short of actual bugs and vulnerabilities which would exist regardless of distribution method).

Either those protections are technological, baked into the OS, and therefore apply equally to all installation sources, or they don't exist. There's no in between.

Look at the history on the PC and Mac desktop side. Ever see someone who had Firefox or VLC, only the binary they got was loaded up with things not shipped by the real developer? Notarization prevents that shady phished from talking your dad into installing “a critical security update!!!” from their own server and then either having it immediately get access to his stuff or walking him through logging into his password manager, etc.

I'm not against notarization as long as it's free (akin to Let's Encrypt) and strictly used for combating outright malicious software like you described, and not as a way to keep competitors off the platform, rent seek, or ban apps for "philosophical" reasons (like NSFW content).

They're intentionally conflating these objectives to give themselves an excuse for maintaining their stranglehold on users and developers alike. They need to give up some ground if their security concerns are to be taken seriously.

I'm sure all the smart people in Cupertino (and elsewhere) can figure out some really great solutions for protecting users in an honest manner, if only their leadership didn't instruct them otherwise.

There’s in-fact an in between, which is humans enforcing rules. It’s what’s in place now. It does have an actual effect, it’s not like it’s imaginary or doesn’t do anything. Some of the rules aren’t practically enforceable by software alone, at least so far (things like “don’t try to fingerprint the user or device in unauthorized ways”)

Those rules are even less enforceable by human reviewers because they don't employ people to reverse engineer your app, never mind any subsequent updates.

Your contention is that the review process entirely fails at enforcing privacy and security rules that cannot be achieved entirely through automation, or fails at such a high rate that it may as well be entire?

That doesn’t reflect my experience submitting apps, nor as a user of Apple devices. It’s certainly imperfect, but it achieves a lot more than if they simply stopped doing it.

[edit] and in fact, some of the automated checks wouldn’t be practical to run on a user’s device—are those also totally ineffective?