In the frenetic world of Web3, where agility and innovation reign, a darker side emerges with the rise of AI security threats. As developers integrate artificial intelligence into their workflows, they inadvertently invite a cavalcade of risks challenging our reliance on traditional cybersecurity measures. Thus, the necessity to grasp and counter these threats has reached...Read More

An AI model just cracked a 29-year-old bug in Squid, the proxy software running schools, ISPs, and corporate networks everywhere. Unit 42 turned two years of pen testing into three weeks using the same tech. Its verdict, from researcher Matt Brady: organizations are “somewhere between three and five months away from these frontier AI models...Read More

By: Vishnu Gatla Runtime has become the place where the real risk of AI-driven networks appears. Strategy documents, policy frameworks, and architecture diagrams are useful, but they do not decide whether an autonomous workflow should call an application programming interface, change a route, open a ticket, retrieve customer data, or trigger a production action. Those...Read More

CrowdStrike Holdings recently reported past second-quarter 2026 results showing revenue of US$1,470.9 million and net income of US$5.31 million, while also issuing higher full-year and third-quarter revenue guidance that now targets up to US$6.01 billion for the year ending January 31, 2027. Management linked this performance and outlook upgrade to rising demand for its AI-focused...Read More

Palo Alto Networks (PANW) has been back in focus after cybersecurity peers reported strong quarters and raised outlooks, lifting confidence in AI security spending and putting extra attention on Palo Alto’s upcoming earnings. After a sharp rally that included a 12% single day gain when peers like CrowdStrike and Okta reported strong results, Palo Alto...Read More

The evolution of artificial intelligence (AI) technology is upending the long-standing calculus of cybersecurity. The asymmetric dynamic—where attackers need to find just a single vulnerability while defenders must block every possibility—is poised to shift toward a defender advantage over the long term, as AI now enables defenders to actually examine all those possibilities. Taesoo Kim,...Read More

CrowdStrike Holdings Inc.’s most recent earnings print shows that Anthropic PBC’s Mythos model transformed artificial intelligence security from something chief information security officers needed to worry about down the road into an immediate buying event. The company’s entrenched position in endpoint  – along with its single lightweight agent, proprietary intelligence and a rapidly expanding portfolio...Read More

Introducing a New Angle on LLM Safety Our previous research on logit-gap steering demonstrated that the safety guardrails of an aligned LLM can be bypassed by closing a measurable gap in the model’s output scores. That work answered the question of how an attacker bypasses alignment. A natural follow-up question is where inside the model...Read More

Analysis CISA advisory AA26-231A is the first U.S. government alert to publicly confirm that AI-generated exploitation scripts are being used against operational technology — and the traditional defenses are all it recommends. The release of CISA advisory AA26-231A, titled Defending Against an Active Threat to Siemens S7 Series PLCs, marks a significant inflection point in...Read More

Artificial Intelligence & Machine Learning , Next-Generation Technologies & Secure Development , The Future of AI & Cybersecurity Unit 42 Says Human Experts Will Increasingly Oversee AI-Driven Defense Work Emilia David • August 28, 2026     Image: Shutterstock The job of security professionals will shift more toward administration as security moves to more autonomous...Read More