@hrbrmstr Brazil is pretty wild. Tons of Mikrotik, even more pirated TV boxes, and extortion-driven DDoS (including by security providers).

@hrbrmstr doesn’t seem likely — sadly, all residential buildings.

https://www.scmp.com/news/hong-kong/society/article/3334292/44-dead-279-missing-hong-kong-fire-3-arrested-investigation-launched

@hrbrmstr too many residential proxies? 😬

Oh well.

@hrbrmstr The source port spread is pretty cool to look at too — very different distribution for CGNAT compared to typical single-subscriber use.

@hrbrmstr Note that with CGNAT the same public IP is shared across several (I want to say 50-100 but don’t quote me on this) subscribers. It can be tricky to detect, unlike VPN exit nodes (and it’s been a pain for us too).

@hrbrmstr I checked these against our own tags: about 103k of them are related to a major residential proxy wholesaler, and about half of that (that we know of) are related to Badbox 2.

In most cases, these IP addresses are from folks who bought an Android-based TV box from some random seller […]

@hrbrmstr hah, I should have known better 😅

@hrbrmstr Next, you’ll need to check out Karakeep! (I have also been really ramping up on self-hosting these past few months)

@hrbrmstr Been on the FreshRSS/Tailscale bandwagon for a while and it’s been great (with Reeder on Mac/iPhone, which I’ve basically been using for as long as it’s existed).

I’ll be looking forward to your update on scraping for non-RSS sites because it’s definitely something I need to explore […]

@resingm The privacy flag is pretty broad as it includes (most?) hosting: https://community.ipinfo.io/t/how-we-classify-ip-addresses-in-the-ip-to-privacy-detection-dataset/5573#hosting-data-center-cloud-5