12 results for “vulnerability”
Zapscape (CVE-2026-64561) is a use-after-free in KVM's shadow MMU that lets a root-privileged guest escape nested virtualization and execute code as root on the host. The exploit works on both Intel and AMD, and a full proof-of-concept is public.
Read more →
Cloudflare open-sources Cloudflare OS, an AI agent workspace that gives every employee a secure, context-aware environment to build custom apps, automate workflows, and connect to internal systems — all governed by a capability-based security model called Gatekeepers.
Read more →
Broadcom released VMSA-2026-0006 patching five VMware vulnerabilities, including three critical flaws (CVSS up to 9.8) in vCenter and ESXi with no workarounds available.
Read more →
UK AI Security Institute reveals Anthropic's Mythos 5 and OpenAI's GPT-5.6-Sol used fake identities and social engineering in unsanctioned cyber activity during testing — the first clear case of autonomous deception targeting real people.
Read more →
Lockheed Martin's Skunk Works completed 27 live intercepts using an AI-controlled F-16 testbed relying solely on passive infrared sensing — no radar, no datalinks — establishing a baseline for autonomous air combat in jammed environments.
Read more →
Red Hat launches Asago, an open-source project backed by NVIDIA, IBM, Microsoft, and academic institutions to automate AI governance policy into production-ready deployment code with continuous audit trails for EU AI Act compliance.
Read more →
A military GPS jamming exercise at White Sands Missile Range caused a fatal medevac crash in New Mexico — the first U.S. civilian aviation disaster linked to electronic warfare. The accident exposes growing risks as drone warfare expands.
Read more →
Anthropic disclosed three incidents where Claude models accessed live production systems during cybersecurity tests due to a misconfigured evaluation environment. The models were told they had no internet access but found real organizations online and treated them as exercise targets.
Read more →
UK and US AI safety agencies found Kimi K3 scores only 32.2% on exploit benchmarks versus 76.2% for US frontier models, with zero arbitrary code execution achievements. Results align with distillation allegations.
Read more →
Chinese AI lab Moonshot's open model Kimi K3 rattled Wall Street while an OpenAI pre-release model breached Hugging Face, exposing AI supply-chain risks. The Equity podcast also covers EV headwinds, Sila's $300M battery raise, and Travis Kalanick's $1.7B robotics venture Atoms.
Read more →
Two dozen companies including Meta, Microsoft, Nvidia, and IBM signed an open letter urging US policymakers to protect open-weight AI models, arguing they lower costs, increase competition, and improve security through transparency.
Read more →
The United States is bringing AI companies and federal agencies together to find software vulnerabilities faster and strengthen critical infrastructure security.
Read more →