Do Son OpenAI has released gpt-oss-120b and gpt-oss-20b, its first open-weight models since GPT-2. The release, under an Apache 2.0 license, aims to democratize access to AI.
Google DeepMind’s Genie 3 brings real-time scene control and enhanced stability to AI training, marking a leap toward more adaptive, interactive world models.
CISA adds three D-Link vulnerabilities (CVE-2020-25078, -25079, -2022-40799) to its KEV Catalog, confirming active exploitation of EOL IP cameras.
A critical vulnerability (CVE-2025-54594, CVSS 9.1) in React Native Bottom Tabs’ GitHub Actions allowed RCE and secret exfiltration. The flaw was patched, and no packages were affected.
The DFIR Report details a destructive attack: Bumblebee malware, delivered via SEO poisoning in Bing search results, led to a full Akira ransomware deployment in under 44 hours.
Mustang Panda, a China-aligned APT, is using the ToneShell backdoor to target government and military sectors with phishing and DLL sideloading, masquerading as a legitimate Chrome component.
Adobe released urgent patches for two critical flaws (CVE-2025-54253, -54254) in AEM Forms on JEE, allowing unauthenticated RCE & file reads, with public PoCs available.
SentinelLABS exposes a crypto scam using AI-generated YouTube videos to trick users into deploying malicious smart contracts, siphoning off hundreds of thousands in ETH.
A critical server-side template injection flaw (CVE-2024-23692) in Rejetto HFS 2.x is actively exploited to deploy ransomware and trojans at scale. Update or retire the service immediately.
Rockwell Automation's Arena Simulation software has three high-severity memory abuse flaws that could allow RCE or data leaks when opening a crafted file. Update immediately.
Zscaler reveals an AI-powered phishing campaign targeting Brazilian users. DeepSite AI and BlackBox AI generated fake government websites to steal Pix payments and personal data.
Trend Micro warns of two critical RCE flaws (CVE-2025-54948, -54987) in Apex One (on-prem) actively exploited in the wild. A fix tool is available, but disables remote agent installs.
From SOC Security Services to Automated Visual Inspection, visionary companies are using powerful, AI-driven platforms to secure,...
Meta and Stanford unveiled a 3mm-thick holographic VR headset prototype, using laser projection to eliminate bulky optics and bringing VR glasses closer to consumer reality.
Proton has patched a critical flaw in its iOS Authenticator app (v1.1.1) that logged TOTP secrets in plaintext. Users are urged to update immediately to secure their accounts.
Cloudflare accuses Perplexity AI of ignoring robots.txt files and impersonating a browser to evade detection, and has removed it from its list of verified bots.
OpenAI adds new "take a break" reminders to ChatGPT and is refining responses to high-stakes decisions, aiming to combat misinformation and promote healthier AI usage.
A critical SQL injection flaw (CVE-2025-54119, CVSS 10.0) in ADOdb’s SQLite3 driver allows arbitrary SQL execution via improperly escaped table names.
Amazon is reorganizing its audio division, integrating Wondery's narrative content into Audible and creating a new "Creator Services" team for celebrity-led podcasts.
Broadcom has shipped Jericho4, a new 3nm Ethernet router designed for distributed AI. It connects over a million XPUs across data centers with lossless, secure 3.2 Tbps HyperPorts.
Meta is reportedly in talks to acquire AI video startup Pika Labs, a move that would accelerate its push into generative AI and short-form video creation for its platforms.
Elon Musk's xAI has filed a trademark for “MacroHard,” a provocative jab at Microsoft. The move hints at a new multi-agent AI venture and reignites a simmering rivalry with Bill Gates.