OpenAI Dots: What the Always-On Agent Can Access, Do, and Cost
OpenAI’s Dots agent can work around the clock across apps and computers. Here is what it can access, which actions need approval, how memory works, and who can use it.
NVIDIA’s Open Agent Safety Platform Moves Guardrails Outside the Model
NVIDIA’s OpenShell runtime restricts AI-agent files, networks, processes and credentials, while Sentry adds a BlueField-4 hardware watchdog. Here is how the stack works and where its claims still need proof.
Apple’s iOS 26 Zero-Day Fix: Update to 26.7.1 Now
Apple says CVE-2026-86950 may have been exploited in targeted attacks. Here are the affected iPhone, iPad and Mac versions, fixed releases, and checks to make now.
AMD’s $8.2B World Labs Deal Is a Bet on AI Beyond Language
AMD is buying Fei-Fei Li’s World Labs for $8.2 billion. The strategic prize is a model lab that can shape future chips, robotics simulation, and the physical AI stack.
Citrix NetScaler Zero-Days: Patch Now, Then Hunt for Web Shells
Citrix patched two actively exploited NetScaler zero-days that allow unauthenticated remote code execution. Here are the fixed builds, attack path, log hunts and response order.
Alibaba’s Zhenwu V900 Turns China’s AI Push Into a Full-Stack Scale Test
Alibaba’s Zhenwu V900 pairs 216 GB of memory with a Qwen 4 roadmap and a 20 GW cloud target. Here is what is shipping, what remains a claim, and what buyers should verify.
EvilTokens Disruption Exposes the Device-Code Phishing Gap
Microsoft and partners disrupted EvilTokens after more than 12,000 inbox compromises. Here is how device-code phishing bypasses normal expectations and what Microsoft 365 teams should change now.
Check Point Zero-Day Puts Firewall Management Servers on a Three-Day Patch Clock
Check Point patched an exploited management-server zero-day and confirmed Spark VPN attacks. Here are the fixed hotfix takes, hunt commands, and September 25 deadline.
Meta Patches Muse Mac Zero-Day: Update Before Using Voice Input
Meta hot-fixed a Muse Mac flaw that let local code redirect dictation, steal an agent token, and inherit connected-service access. Here is how it worked and what users should check now.
Palo Alto Launches Always-On AI Pentesting With Mythos 5 and GPT-5.6-Cyber
Palo Alto Networks' Unit 42 now uses Mythos 5, GPT-5.6-Cyber, and open models for continuous attack-path testing. Its own results show why buyers should judge the harness, controls, and remediation workflow, not just the models.