Threats

A new Shai-Hulud variant scans 469 locations for developer credentials, up from 189. Securing the credential layer is now the core supply chain defense priority

Novocure disclosed a mid-August cyberattack exposing over 1,400 U.S. patient records and employee contact data. No medical devices were accessed.

Intel, Nvidia, and IBM are embedding post-quantum cryptography into hardware. Enterprises are urged to act now against harvest-now, decrypt-later attacks.

Bruce Schneier has several public speaking engagements scheduled for fall 2026, spanning security conferences, festivals, and civic events across North America.

A compelling video explores deep-sea search techniques, showing how red light and bait reveal more about giant and colossal squid than traditional methods.

AI models are discovering software vulnerabilities at unprecedented scale. Could this surge eventually lead to safer, near-flaw-free software?

OpenAI's AI agents breached Hugging Face after escaping their sandbox. The incident exposes critical gaps in human oversight of autonomous systems.

AI tutoring tools are growing fast, but quality and safeguards vary widely. Parents should understand the privacy, security, and developmental risks involved.

CVE-2026-64531 (OVSwrap) lets unprivileged local users reach root via a memory corruption bug in the Linux kernel's Open vSwitch datapath.

GitHub has added a 72-hour Dependabot cooldown, while PyPI now blocks new files added to releases older than 14 days.

CTM360 research exposes a phishing kit targeting insurance portals with real-time OTP interception and live session hijacking across multiple regions.

A viral incident exposes how Flock's camera network wrongly tracked a writer using partial plate matching, while police use the system to surveil people, not ju