Field notes from the edge.
What our engineers learned this week. Hands-on technical deep-dives, postmortems, and strategy frameworks.
AIAI for Military Support
A study of 2,015 Israeli military personnel using a replica AI decision-support system for military targeting reveals that soldiers exhibit algorithmic aversion rather than automation bias, particularly in high-collateral-damage scenarios. The research demonstrates that explainable AI features can reduce this aversion and improve decision-making quality, while trust in military AI systems varies b
AIKimsuky Builds Offline AI Stack to Boost Phishing and Automate Malware Development
North Korea's Kimsuky APT group has evolved beyond using public AI chatbots to deploying private, offline AI infrastructure on dedicated servers. The threat actor is leveraging AI for enhanced document analysis of stolen data and developing capabilities to integrate AI directly into malware creation processes, according to research from South Korean security firm Genians.
AINEC tests parking tech that only starts charging once you exit your car
NEC is testing AI-powered parking technology that charges drivers only after they exit their vehicles, using cameras and real-time video analysis to detect when drivers leave their cars. The system aims to eliminate charges for time spent searching for spots or exiting facilities, while also enabling surveillance capabilities for monitoring visitor trends and detecting potential driver emergencies
AIMeasuring LLMs’ Ability to Perform Cryptanalysis
A new benchmark called CryptanalysisBench tests LLMs' ability to discover mathematical cryptanalytic attacks against 191 historical cryptographic primitives. Frontier models from Anthropic, OpenAI, and others successfully broke 65-86% of known-vulnerable schemes and discovered novel attacks, including previously unknown vulnerabilities in SpoC AEAD and KINDI's security proof. This represents an em
AIClaude AI Just Cracked a Post-Quantum Test Scheme and Found a Faster 7-Round AES Attack
Anthropic's Claude Mythos Preview AI has demonstrated significant cryptographic analysis capabilities by successfully deriving a key-recovery attack against the post-quantum HAWK-256 signature scheme and accelerating attacks on seven-round AES-128 by 200-800x. The HAWK attack leverages previously unexploited lattice symmetries and can execute in approximately 3 hours 42 minutes on a 96-core server
AICollege prof hides prompt to catch AI cheaters, finds human nature is pretty much as we thought
A college professor at Alcorn State University embedded hidden white text in an essay prompt instructing AI tools to insert the word 'Madagascar' nonsensically, catching 32 of 35 students using AI to complete assignments without proofreading. The incident highlights growing concerns about AI's impact on academic integrity and critical thinking skills, with research showing reduced brain activity a
AIMicrosoft Says New Cybersecurity AI Model Helps MDASH Hit 95.95% at Half the Cost
Microsoft has introduced its first cybersecurity-specific AI model, MAI-Cyber-1-Flash, integrated into its MDASH vulnerability identification and remediation platform. The new model combination achieved a 95.95% score on CyberGym benchmarks while reducing operational costs by 50% compared to Microsoft's previous best configuration. Access to the enhanced MDASH system is currently restricted to app
AIPhishing Attack Volume Down 20%, but Risk Still Rising
Despite a 20% decrease in overall phishing attack volume, the threat landscape remains increasingly dangerous as cybercriminals shift their strategy from mass campaigns to more sophisticated, targeted attacks. Hackers are leveraging AI technologies to enhance the quality and effectiveness of their phishing attempts, making them harder to detect and more likely to succeed. This evolution represents
AIBlame AI: Patch Tuesday Hits Record 206 CVEs
Microsoft's latest Patch Tuesday released a record-breaking 206 CVEs, signaling a new era where AI-driven vulnerability discovery is dramatically increasing the volume of security patches. This trend suggests that enterprise IT teams should prepare for significantly larger and more frequent patch cycles as AI tools accelerate the identification of software vulnerabilities at unprecedented scale.
AIResearchers Build Self-Replicating AI Worm That Operates Entirely on Local, Open-Weight Models
University of Toronto researchers have developed a proof-of-concept AI worm that autonomously navigates networks, generates customized attack strategies, and self-replicates using only locally hosted open-weight language models. The system operates independently without human intervention or reliance on commercial AI services, demonstrating significant cybersecurity implications for enterprise env
AIAI Worm
Researchers have developed a prototype AI-powered internet worm that carries its own large language model and executes it on compromised systems. This implementation closely mirrors John Brunner's original 1975 conceptualization of computer worms, representing a significant evolution in autonomous malware capabilities.
AIZoom CISO: AI as Security Enabler, Not Role-Replacer
Zoom CISO Sandra McLeod emphasizes that AI should augment rather than replace cybersecurity professionals, discussing how AI-driven security workflows can enhance protection for global communication platforms. She addresses the unique challenges of securing enterprise collaboration tools at scale while offering guidance for emerging cybersecurity leaders.
AIThe Intersection of Encryption and AI
Bruce Schneier reflects on his 2010 Dark Reading essay about cryptography's limitations in securing modern networks, emphasizing that while cryptography has strong mathematical properties favoring defenders, it cannot address most contemporary cybersecurity challenges like DDoS attacks, malware, and network penetration. He traces this argument from his 2000 book 'Secrets and Lies' through today, n
AIAs Global Powers Explore Humanoid Robots, Cyber-Risk Looms
Nation states are competing for leadership in the emerging humanoid robotics and embodied AI market, creating new cybersecurity challenges. As these technologies advance and supply chains develop, organizations face evolving cyber-risks that require proactive security strategies. The convergence of physical robotics and AI systems introduces unprecedented vulnerabilities that enterprise security t
AIMeet Our Newest AWS Heroes – May 2026
AWS has announced four new AWS Heroes for May 2026, recognizing community leaders from Italy, Canada, and Argentina who demonstrate exceptional commitment to knowledge sharing and community building. These individuals specialize in AI, serverless architecture, and cloud solutions, contributing through user group leadership, content creation, conference speaking, and developing tools that benefit t