4 ms·
Exactly. But at least pre-Big-Sur you could use Little Snitch to block Apple apps from sending things, but now that's no longer possible. :(
by 1over137 6y ago
Exactly. But at least pre-Big-Sur you could use Little Snitch to block Apple apps from sending things, but now that's no longer possible. :(
- sneak 6y agoIf you have an intel mac, you can still install LS 4.x (once you have booted into recovery to permit the kext) on Big Sur. HN user miles pointed this out recently, and it is awesome. https://www.obdev.at/support/littlesnitch/245913651253917 https://www.obdev.at/support/littlesnitch/245913651253917 It's 5.x that uses the new restricted APIs. I'm going to contact the developers and ask for an ARM build of 4.x so the same trick will work on M1, at least until Apple forbids all kexts some time in the future.
- nbzso 6y agoThis is deal breaker for me and lots of security conscious professionals. I have foreseen this "Apple goal" in the past and the only thing thats keeping me to use MacOS is Little Snitch and Mullvad VPN combination. Sadly in Linux world there is no commercially viable option for access rules per app basis and I don't understand why. I don't have answer for another question: As a business I learned from Apple that keeping tight security and protecting your intellectual property is big thing. How on earth big business is complicit in this telemetry approach from Microsoft, Google and Apple? If you are business captured metadata is enough to know important metrics about your company and this may work against you in the long term. Thats why I am against cloud based apps (like Figma), and I am furious with Sketch for not providing collaboration solution and bragging about "Mac Only" approach.
- acdha 6y agoIt's because it's very time-consuming but ultimately it's more security theater than any meaningful benefit. An attacker has many ways to easily bypass tools like Little Snitch and they only have to succeed at one of them, whereas you have to hope that you take time away from your job to successfully block all of them. If you're trying to prevent data exfiltration, you don't trust the client at all — confine it to a dedicated locked-down system on a restricted network which only allows egress to the minimal subset of trusted services. That's a much more winnable battle than trying to prevent every possibility on a general purpose computer running tons of things which are allowed to connect to the internet and legitimately uses lots of outside services. Similarly, a lot of the data breaches you hear about are caused by people with legitimate access saving the data somewhere insecurely. Spending time on that is a lot more beneficial to most organizations than tracking every TCP socket.
- nbzso 6y agoThanks for your comment. You described exactly what I am going to do. + Hardware Firewall/Vpn.