Security
A teardown of the Claude Code client found a function that edits the punctuation in your system prompt. The apostrophe in Today's turns into a four-state label the moment you point the tool at your own endpoint. Here is how it works and how to check your own.
AI
Last week OpenAI's best model launched to about twenty customers approved by name by the US government, and Anthropic's new flagship lived three days before a directive switched it off. The official story is cybersecurity. Here is my conspiracy theory about the second engine under the hood.
Security
A researcher just found 10,000 GitHub repos serving Trojan malware, cloned from real projects and wearing their real commit history as a disguise. Why 'it's on GitHub' stopped meaning 'it's fine', why vigilance is the wrong fix, and why my code lives on a box in my office.
AI
Morndur weaves whatever players write into one shared story, and its AI narrator is forbidden from inventing a chapter out of nothing. I said so on Lemmy and got called a slop merchant. What that taught me about the worth of human writing, and why every post here still carries a canary.
Privacy
Hide My Email kept you private by hiding in the crowd at icloud.com. This summer Apple moves every alias to its own subdomain and hands any service a one-line rule to block them all. What changes, and the version no vendor can switch off.
Security
My own Ghost blog has open signup, so I went looking for a CAPTCHA. Ghost built one and ripped it out. Then Cloudflare Turnstile started reading every visitor's GPU.
Fediverse
The euphoric post said it works. Half-truth. Federation works; every interaction (like, boost, reply, open article) silently dies. Here is the real fix: Apache AllowEncodedSlashes + nocanon, plus the hairpin-NAT gotcha.
Postfix was the easy afternoon. Spamhaus, PBL/SBL/XBL and the open-resolver trap were the week-long education nobody puts in the triumphant self-hosting posts.