You turn on your VPN and your IP changes. It feels like you should be invisible, but some websites still know exactly where you are. Your IP is just one clue. Websites also look at your browser’s timezone (for example, "Asia/Kolkata" doesn’t match a German IP), your Accept-Language header (en-IN gives you away), and your DNS resolver, which is probably still set to your ISP’s servers back home instead of your VPN’s. One of the sneakiest trick is WebRTC. Your browser has a built-in feature for video calls that can fetch your real IP address at the operating system level, completely bypassing the VPN tunnel. A website can get it with just 10 lines of JavaScript. Some VPNs don’t block this by default. And if you’re logged in, it’s game over. Netflix, Google and Spotify don’t care about your IP address. They care about your account. Your registered country is stored in their database, not in your connection. Using a VPN while logged in means the VPN isn’t helping you at all. A VPN gives you a different postal address, but your timezone, language, DNS, WebRTC, browser fingerprint, and login details are still the same.

🦔 The IRS paid Palantir $1.8 million last year to build a tool called SNAP, the Selection and Analytic Platform, designed to identify the highest-value targets for audits, tax collection, and potential criminal investigations. The system pulls from more than 100 legacy IRS databases built over decades and is being piloted across three areas: disaster zone tax relief claims, residential clean energy credits, and gift tax returns. Palantir has received more than $200 million in IRS contracts since 2014. The agency has lost more than 25,000 staff since February 2025 through resignations and early retirement offers. My Take Using AI to modernize a tax agency running on 100 fragmented legacy systems built over decades is not inherently a bad idea. The IRS has failed to update its technology repeatedly since the 1960s and a smarter case selection system that catches genuine fraud more accurately than a decades-old scoring algorithm is a reasonable goal. The concern worth focusing on is what you get when you combine an agency that has shed 25,000 experienced staff with an AI system selecting audit targets from a process nobody fully understands. The existing DIF scoring system is already described as a black box by researchers. SNAP adds another layer of complexity on top of that. Who gets flagged and who doesn't is ultimately answered by the people configuring the system and the data they feed it, and with fewer experienced staff to apply judgment, the margin for error gets wider, not narrower. Hedgie🤗

After 12 years, Ubuntu MATE founder Martin Wimpress is stepping back, and the project is now seeking new maintainers to assume key responsibilities. linuxiac.com/ubuntu-mate-fou… #Linux #Ubuntu #OpenSource

One cable. One port. Zero room for error. This is the kind of task that looks simple — until a robot has to do it reliably in a tight space. In this demo, a RealMan robotic arm performs Ethernet cable insertion inside a server rack.