Most Apple users will definitively want those on, and will break your system in subtle ways if you disable them. I wouldn’t make such assumptions about whatever you use the service or not. If you don’t use a service, it typically will have very little traffic if any at all.

Like I said this was all from the last 3 hours and I haven't used any of those services listed. It's also missing the point a bit, I'm not bandwidth poor on a 3G connection trying to save my datacap.

I don't want apple turning my laptop into a thinclient for their cloud systems. I don't want telemetry and meta data going to them every ~30 seconds (the little snitch icon flashes a red X every time something is blocked, it's on a per minute basis).

Of all those services I use calendarsync. I miss having airdrop a little bit but everything else I don't need.

Also, the way you've phrased some of those, especially the notarisation

> getting updated security signatures

is a bit disingenuous when there's a massive privacy implication in that it allows apple to know every single application I run on my computer.

Most Apple users wouldn't agree to that if you stuck it in their face and the fact that it breaks the OS in subtle ways is a user hostile position to argue from. Hence why I'm giving up my Apple addicition.

AKA phoning-home to Apple what apps you launch in realtime, in an unencrypted manner visible to your ISP/hotel/government too.

It also connects to all of that crap even if you have LS off, analytics off, iCloud/FaceTime/iMessage/AppStore off, et c.

Press F8 and your serial number gets transmitted to Apple.

IP Addresses: 2600:1402:e::b833:965b

  2600:1402:e::b833:9661

  2600:1402:e::b833:9669

  2600:1402:e::b833:966a

  + 17  more

Protocol: TCP

Connected: no

Connects: 0 allowed, 7,359 denied

First Activity: 2/11/21, 22:20

Last Activity: 5h 29m ago

Yea sorry, sneak's right, it's still going port 80.

https://support.apple.com/en-us/HT202491 (bottom)

What data is actually going over OCSP? I thought it was only ever checking for revocation and not actually sending app data, and I believe only ever happens when you want to take an app out of quarantine versus every app launch. You can always disable gatekeeper if you so please.

OCSP over unencrypted http has not yet been replaced. I believe that the constellation of apps (identified by proxy by their developer IDs which are transmitted) can in a lot of cases uniquely identify a system, given a sufficient number of apps installed/used.

How so?

What happened to the good old days of "ask for user consent before phoning home"?

I mean, Apple is miles ahead of Windows in that regard, but your average Linux or BSD setup won't phone home outside of repository downloads unless requested.

The only happy medium I would consider to your approach is that if MacOS had two set up routes, one defaulting as a power user turning everything off and then initiating things are you want and another as a general user. Maybe that would solve it (though would be a heavy lift I am sure).

I have tried giving permission and at the end of the day, it's just not worth the time. For me, Little Snitch is great as a reporting tool but it's just too much work as a firewall.

[0]https://www.obdev.at/products/littlesnitch/index.html

It’s far worse than just annoying: if people have to give permission to a bunch of things they don’t understand, they will absolutely give permission to something they shouldn’t.

You want the user to make a few decisions as possible, and every single one should be an actual decision: where the user knows what the options are, has an actual stake, and might legitimately choose either option depending on their preferences and circumstance.

The more times they have to click “yes” without thinking, the more susceptible to malware they become.

I understand the point, though. Going to just about any major website you will be pummeled with prompts to allow for a dozen different domains just to view one page, and it doesn't really give any indication of what those are used for. They have a what seems to be infinitesimally small list of connections that they do recognize and explain their use, but ultimately it's pretty useless for the vast majority of prompts.

I'd like there to be a better way to deal with this, but I'm not really sure what the solution to it would look like. You can download blocklists and just silently allow other connections, but I don't think that's significantly better than just using a hosts file.

The most useful thing Little Snitch does is alert me when individual applications try to phone home. For browsing the web it feels more like a chore.

How many times have you been asked to approve a download? How many of those times have you said "no"?

If the answer is "hundreds" and "zero," what's the point? If the answer "hundreds" and "ten just in the past week," then that's exactly the point, and it serves a valuable purpose.

I agree broadly with your point that prompt fatigue or decision fatigue is problematic and should be avoided when possible. I think this is a problem in particular with Little Snitch even, which doesn't do enough to provide. However, the point of Little Snitch is to allow someone to monitor and control the traffic at a granular level and the consequence of providing that utility is the frequent prompts whenever you're visiting a new site. To Obdev's credit, silent mode exists and you can set it to deny or allow all traffic without a prompt (and evaluate the traffic at your leisure).

Is there any way to turn all this stuff off?

All of the things in your list are just apps that can be removed with one click. My partner’s PC comes pre-installed with masses of hidden spyware. I know the kneejerk reaction is to bash Apple for everything - but the two are totally not comparable.

I thought the OS shipped with its own firewall? Would you recommend using a third party firewall despite having its own?

There was some controversy in the previous year because apple tried to deprecate the firewall hooks and allow their junk passed the replacement they offered.

I'm leaving apple for linux as we speak for two reasons, privacy and they can't help but mess with shit every time they release an update.

The most recent, this week they released an update and now their airport service listens on port 5000 making a conflict with running a dev flask service locally.

"Mess with shit" was what drove me to Linux by around 2015. It seemed like every major MacOS upgrade torched my Eclipse-based Java dev environment, generally requiring a reinstall.

I’ve flip flopped for years, (I was using freebsd more than 20 years ago, genuinely ran Solaris 10 with my own build of kde4 on a hp probook, etc) but last time I tried to use Linux for work I got defeated by a conference suite projector at a client and it cost us a big contract. The happy path (for me anyway) is doing all actual dev work on Linux vms and using macOS as your browser/im/terminal client — which it’s great at.

And it works with those stupid little projector dongles at clients when you’re trying to pitch them 6 figures of consulting ;)

It takes about one time of "this didn't work and it's your fault for using Linux" in a business context to break one of using Desktop Linux (or BSD, et c.), I think. I doubt similar stories are uncommon. I've got one, certainly.

At least if MacOS or Windows breaks, you're not the asshole. And it doesn't hurt that they in fact do break less often (well, Win10 with its abrupt, unexpected, and slow forced updates was a real problem for a while, but otherwise)

Funny how framing something that's true (allowing apple software to bypass firewalls) is seen as a controversy. See previous discussion on hn: https://news.ycombinator.com/item?id=24838816

If it has a firewall of any kind, I've never noticed nor interacted with it. I've also never installed a third party firewall.

Some popular third party solutions that are actually useful are Little Snitch [0], LuLu [1], and AdGuard [2].

[0] https://www.obdev.at/products/littlesnitch/index.html

[1] https://objective-see.com/products/lulu.html

[2] https://adguard.com/en/adguard-mac/overview.html

If you'd like an application-level firewall, you can check out Lulu or Little Snitch. Back when Little Snitch still(?) installed kernel extensions, it was found to be quite insecure—there were talks at DEFCON about it. Lulu is a lesser-known alternative.

I’d rather see humans themselves as our value store and motivate, teach, and mentor each other to do the logistics work to that end.

Letting banks gamble with our deposits when we’re all in the habit of doing these things daily anyway seems like pointless extra steps.

I think people are more willing to be a stable mesh network of agents for each other than history full of feudal warlords will let us believe, we don’t need to prop up ephemeral tent poles. That’s just old story.

Bonobo tribes have been observed killing alphas and then developing social harmony. Not saying we should behead billionaires, but austerity for elites would be a figurative death.