Summary
A wave of security incidents this week has prompted Peter Yang to share six steps for locking down online accounts. He points to roughly a dozen security “fires” in a few days, including a government warning for organizations running PHP to update immediately or go offline, and a report that banks are finding unauthorized AI agents inside customer accounts.
The video explains one theory for the surge: much of the world’s infrastructure code was written before AI and contains many latent vulnerabilities, and AI now lets both attackers and defenders find them quickly. The presenter also describes an alert about an agent deleting emails from his own account.
The six steps are disconnecting unused apps, especially those tied to Google accounts; using a password manager such as 1Password and its Watchtower audit; enabling authenticator-app two-factor authentication instead of SMS; rotating or retiring API keys; having AI scan and update software and packages; and avoiding breach-related phishing emails and calls.
He also demonstrates a single prompt that works with ChatGPT desktop or Claude desktop to produce a personal security report covering connected apps, reused passwords, missing two-factor authentication, and more, with a rule that nothing is changed without approval.
📺 Source: Peter Yang · Published October 22, 2026
🏷️ Format: Tutorial Demo
