Tuesday

August 11, 2026

Summer · 2 entries

The Read
phys.orgAI is making disinformation harder to spot—but we've found a new way to catch itResearchers developed a system to detect AI-generated disinformation by analyzing conversational flow rather than word choice. Instead of identifying whether text was written by AI, the system detects when comments derail discussions using red herrings, non sequiturs, and topic shifts—achieving 77%✦ Read ad free and get the full MichaelFilter · $5.50
8:20 pm
💬 Comment
The Listen
On this day · Full timeline

Get MichaelFilter

Enter your email. We'll send a one-tap link to read today's edition free, in your browser — the link works for 24 hours. Membership unlocks every issue, PDFs, and transcripts.

Already a member of theyoga.club or ashtanga.tech? You're already on file — no re-registration. We won't add you to any list.