NSA
PulseAugur coverage of NSA — every cluster mentioning NSA across labs, papers, and developer communities, ranked by signal.
-
NVIDIA open-sources cuDNN kernels after 12 years, including MoE and sparse attention
NVIDIA has open-sourced parts of its cuDNN library, a significant move after 12 years of it being closed-source. This release includes over 20 Mixture-of-Experts (MoE) kernels and NSA sparse attention kernels. The codeb…
-
NSA tests Anthropic's Mythos AI for security flaws; Google, xAI join review
Anthropic is launching a $1.5 billion venture with Wall Street firms to integrate its Claude AI model across financial services. Separately, the NSA is reportedly testing Anthropic's Mythos AI model for identifying secu…
-
William Binney warns of "turnkey totalitarian state" enabled by surveillance
William Binney, a former NSA surveillance architect, has described the current state of surveillance as a "turnkey totalitarian state." He argues that whoever holds power gains access to an extensive surveillance system…
-
Five Eyes nations warn against using AI agents for sensitive tasks
Government agencies from the Five Eyes intelligence alliance have issued a joint advisory warning against the use of AI agents for sensitive tasks. The advisory, released by entities including the NSA and CISA, emphasiz…
-
Five Eyes agencies caution against rapid agentic AI adoption due to risks
Security agencies from the Five Eyes intelligence alliance have issued guidance cautioning against the rapid adoption of agentic AI. The report highlights that these systems can amplify existing organizational weaknesse…
-
Disneyland deploys facial recognition, NSA tests Anthropic's Mythos AI
Disneyland has begun employing facial recognition technology to monitor visitors, a move that raises privacy concerns. Concurrently, the NSA is evaluating Anthropic's Mythos Preview model to identify potential security …
-
CISA warns of data-theft bug in NSA-developed GrassMarlin tool
The Cybersecurity and Infrastructure Security Agency (CISA) has issued a warning about a data-theft vulnerability, CVE-2026-6807, affecting GrassMarlin, a network security tool developed by the NSA. This flaw, stemming …
-
Researchers uncover decades-old cyber-sabotage tool targeting critical infrastructure
Security researchers have uncovered a cyber-sabotage tool named fast16 that predates the Stuxnet worm by at least five years. This tool was designed to target software used in critical infrastructure projects like nucle…
-
Anthropic's highly secure AI model, Mythos, suffers embarrassing security breach
Anthropic experienced a significant security lapse with its highly capable cybersecurity AI model, Claude Mythos. Despite emphasizing the model's dangerous potential and restricting its access, unauthorized users gained…
-
CISA lacks access to Anthropic's powerful Mythos hacking model
The Cybersecurity and Infrastructure Security Agency (CISA) has been denied access to Anthropic's new AI model, Mythos Preview, despite its potential to identify and exploit security vulnerabilities. While over 40 compa…
-
Pentagon CTO: Anthropic barred despite NSA evaluating Mythos AI
The Pentagon's Chief Technology Officer, Emil Michael, has clarified that Anthropic remains a supply chain risk and is barred from Department of Defense systems. Despite this, some government agencies, including the NSA…
-
OpenAI inks deal with Department of War, adding domestic surveillance guardrails
OpenAI has finalized an agreement with the Department of War to deploy advanced AI systems in classified environments, incorporating specific safeguards against misuse. The updated contract explicitly prohibits the use …