3 ms·
What tf did you do to get your ranges blacklisted?
by blueflow 2y ago
What tf did you do to get your ranges blacklisted?
- jsheard 2y agoProbably the fact that I was using Mullvad, the flipside of their commitment to privacy is they can't really do anything about abuse coming from their IP ranges and they inevitably end up on blacklists. With the desktop client you can at least mash the reconnect button until you get an IP that works, but that's a hassle if you are running the tunnel somewhere else.
- mlv- 2y agoYou'll get more CAPTCHAs and other issues with other VPNs, too, their IP addresses are just too easy to identify. VPNs with residential IP addresses can be a different experience, but they're a bit pricey if you can't justify their use as a business expense.
- pogue 2y agoWhat are some paid VPNs with residential IP addresses? I've always wondered why they don't just mask them as looking like ones anyway.
- kccqzy 2y agoWell you made the conscious choice to make your own traffic indistinguishable from abuse traffic.
- jsheard 2y agoYeah I get it, I'm not blaming anyone for shitlisting those IP ranges given the circumstances. There is no way to ensure user privacy that doesn't also give bad actors the same privacy, and end up necessitating the assumption that anyone who goes out of their way to be private is a bad actor. This is why we can't have nice things.
- kccqzy 2y ago> There is no way to ensure user privacy that doesn't also give bad actors the same privacy I disagree. There are upcoming technologies that have the potential of making this a reality. Like Private Access Tokens (https://www.ietf.org/archive/id/draft-private-access-tokens-01.html https://www.ietf.org/archive/id/draft-private-access-tokens-...) and its successor protocols, the Privacy Pass suite of protocols (split over many documents https://www.ietf.org/archive/id/draft-ietf-privacypass-architecture-16.html https://www.ietf.org/archive/id/draft-ietf-privacypass-archi... https://www.ietf.org/archive/id/draft-ietf-privacypass-auth-scheme-15.html https://www.ietf.org/archive/id/draft-ietf-privacypass-auth-... https://www.ietf.org/archive/id/draft-ietf-privacypass-protocol-16.html https://www.ietf.org/archive/id/draft-ietf-privacypass-proto... https://www.ietf.org/archive/id/draft-ietf-privacypass-rate-limit-tokens-06.html https://www.ietf.org/archive/id/draft-ietf-privacypass-rate-...) and so on. These efforts are spearheaded by Apple, Google, Cloudflare and Fastly. Of course Apple has iCloud private relay so they have incentives to make such things work well with iCloud private relay. Google has Google One VPN too. It also has reCAPTCHA that's widely used to block abuse traffic. Cloudflare and Fastly of course are big CDNs. Imagine a world where visiting a website over a VPN, and these tokens clearly distinguish you from abuse traffic while preserving your privacy, so that Cloudflare has no reason to present you with a CAPTCHA and reCAPTCHA gives you that tick without forcing you to pick pictures of motorcycles. Watch this space. Very interesting developments here.
- jsnell 2y agoGoogle One VPN is being shut down. And separate from that, I bet that wide deployment of privacy-preserving anti-abuse mechanisms is dead after the WEI fiasco. It's just not possible for Chrome to ship anything even remotely resembling that feature given the PR damage from that proposal. That includes PATs, which are pretty much exactly the same thing as WEI. And as long as PATs remain an Apple-only feature, implementing them as captcha bypasses is not very attractive. (And honestly, giving Safari iPhone users a special fast lane that other users simply can't get strikes me as a really bad idea.)
- kccqzy 2y ago