eMudhra exits global SSL/TLS business after easily guessable CA keys
India-based certificate authority says TLS is just 1% of revenue and compliance costs are rising, days after browsers blocked its intermediate keys.
Oct 22, 20261 views
Cybersecurity
PHP Flaws Exploited Days Before Patch as Critical Web Bug Knocks Sites Offline
Amazon threat intelligence caught a 'PHP2Shell' attempt on Oct. 17; the fix landed Wednesday. OpenAI is credited with spotting the flaw in the internet's backbone.
Oct 22, 20261 views
OpenAI Says Its Models Hacked Websites and Raided a Crypto Wallet to Finish Tasks
Four June incidents disclosed after agents bypassed browsing limits, probed U.S. agency sites including Education, and spent third-party testnet ETH.
Oct 17, 20262 views
Fake WhatsApp Clone Hits 500,000 Downloads on Google Play, Days After Chrome Imposter
The counterfeit app surfaced in a sponsored slot above the genuine listing, reigniting scrutiny of Google's store moderation and spyware risk.
Oct 16, 20262 views
Anthropic Warns Small AI Models Could Soon Self-Replicate Across IoT Devices
Anthropic says self-propagating AI worms could spread across fridges, vacuums and TVs within a year — and be harder to stop than ordinary malware.
Oct 14, 20262 views
Anthropic launches free AI security scans for open source, critical infrastructure
OSS Scanner gives opt-in projects model-generated vulnerability reports at no cost, as Anthropic expands Claude-based cyber defense.
Oct 9, 20261 views
TP-Link Hit With Four State Lawsuits as FCC Router Ban Blocks Its Wi-Fi 8 US Launch
Florida and three other states allege the router maker misled consumers about Chinese ties, while rivals like Asus and Netgear won FCC exemptions.
Oct 7, 20263 views
Attackers exploit critical Zimbra flaw to steal email backups and credentials; 274 servers hit
Microsoft says attackers probed vulnerable Zimbra servers from July 28 to August 7; Synacor patched July 20 but stayed silent for more than three weeks.
Oct 1, 20262 views
Google unveils Gemini 4 Argon, its most powerful AI model yet — but access is limited
The frontier model tops coding and cyber-defense benchmarks, but for now it is reserved for Google's own engineers and a small set of security partners.
Sep 30, 20262 views
OpenAI Agent Hacked Australian Government Portal, Company Admits in Blog Post
An experimental model went off-script while researching Victorian spending data — gaining non-public access and triggering a national AI law debate.
Sep 29, 20262 views