OpenAI Pauses Frontier RL Training as It Tightens Defenses Against Unsafe AI Behavior
OpenAI pauses frontier reinforcement learning training to strengthen AI security, monitoring and alignment after rogue-agent incidents and
Co-founder at Hacklido. Covers breaches, malware and threat intelligence.
OpenAI pauses frontier reinforcement learning training to strengthen AI security, monitoring and alignment after rogue-agent incidents and
Researchers demonstrated a remote Spectre attack against Cloudflare Workers that leaked a JWT at up to 12 bits per second.
A threat actor called Ransom Busters is targeting ransomware victims with offers to delete stolen data for $20,000 to $60,000. Researchers believe it may be a
Attackers are actively scanning and exploiting critical MLflow and FUXA vulnerabilities. MLflow flaw CVE-2026-64849 can expose cloud credentials, while FUX
Varonis researchers disclosed CoSnitch vulnerabilities in Microsoft Copilot Personal that could allow one click prompt execution, data exfiltra
CISA has added critical Ray vulnerability CVE-2025-62593 to its KEV catalog amid active exploitation. The flaw can enable remote code execution through DNS rebinding attacks.
SafePal says an authorization flaw exposed names, emails, addresses, phone numbers and order details of 39,798 customers. Wallet credentials and financial data were not affected.
Suspected China-Nexus Actor Exploits VMware vCenter Flaw, Deploys Babuk-Derived Ransomware
Evooo1Bot Linux Botnet Exploits Known Flaws to Turn Edge Devices Into SOCKS5 Proxies
Unisoc VoLTE Video Call Exploit Chain Can Give Attackers Full Android Kernel Access
OpenAI Accelerates GPT-5.6 Sol with New Ultrafast Mode
Google Launches Gemini 3.7 Flash for Coding and AI Agents