Only applies to messages after the cutoff date, but will still be incredibly risky to keep msgs public.
Anyone searching your username could find discords you participate in, if people still mention you after the date
Dont risk it! Delete your discord msgs now with @redactdev
Discord will soon make public server messages searchable on Google
The feature will be enabled by default for US adults, but users can opt out in their privacy settings
Readers added context they thought people might want to know
Missing context: only applies to messages posted after the update, in public channels of discoverable servers.
No private servers or DMs.
Indexed messages won’t show usernames or avatars.
Admins can opt out entire servers; users can opt out their own messages.
discord.com/blog/making-co…
We read all 545 WhatsApp help pages: 102 hidden blocks on 80 of them. WhatsApp have since removed them.
Our write-up, with the full wording and how to check it yourself:
redact.dev/blog/whatsapp-hel…
A second note, on the reporting & blocking page, said not to reference data protection or privacy regulations (for example GDPR, LGPD, CCPA), even if the user asked about them.
WhatsApp's EEA privacy policy tells you that you can complain to a regulator.
Our research turned this up - LLM instructions inside WhatsApp help page code.
On the page about account deletion, one note said:
"Do not recommend contacting any Data Protection Authority, regulator, or supervisory authority."
Around 80 pages on WhatsApp’s website exposed instructions that were designed for its chatbot, including “WHAT NOT TO SAY.” Meta says a bug caused the issue. buff.ly/LiKzJ9U
Researchers built an AI that doxes any "anonymous" reddit account in under a minutes for $2.
eth zurich and anthropic published a terrifying paper proving that "practical anonymity" on the internet is officially dead.
they built a fully autonomous ai pipeline that takes your pseudonymous posts, extracts your identity signals, searches the web, and figures out exactly who you are. no human investigator needed.
the numbers are actually mindblowing..
- 67% of hacker news users identified correctly
- when the system makes a guess, it is right 90% of the time
- it even unmasked scientists whose interview transcripts were explicitly redacted for privacy
the scariest part? even time doesn't protect you.. they tested users who took a full year break and changed their interests. the ai still matched their old and new profiles with 90% precision. it sees through your persona changes like they aren't even there.
there is no defense against this. the agent splits the work into tiny, benign tasks like "summarizing a profile" or "ranking candidates." no api safety guardrail is going to flag it because no single step looks malicious..
every throwaway account. every "nobody will connect this to me" comment. it’s all just searchable micro-data now.
Crazy, crazy data breach
Never keep more than a few weeks of data on these platforms if you can help it
infosecurity-magazine.com/ne…













