AI-generated discussions on the latest tech and AI news. Every article gets a unique podcast episode.
795
Episodes
13+
Hours of content
EVMbench is an open-source benchmark from OpenAI and Paradigm that tests AI agents on detecting, patching, and exploiting real smart contract vulnerabilities. It uses 120 curated flaws to provide automated, repeatable evaluations of AI security analysis capabilities.
Dutch defense secretary claims F-35 fighter jets can be "jailbroken" like an iPhone, suggesting European forces could maintain software without US help, though experts note major access and security hurdles.
Companies are banning OpenClaw, an AI tool that controls computers, over security fears. While some explore its potential, many restrict use to protect sensitive data.
Paper on using LLMs for theorem proving to verify the seL4 microkernel, aiming for industrial-scale real-world application.
A UK medical center displayed system usernames and passwords on a public whiteboard, highlighting a major security flaw despite guidelines. This underscores human error as a key vulnerability, with passkeys suggested as a more secure future alternative.
DEF CON banned three men named in Epstein files due to their associations with him, though they face no criminal accusations. The banned individuals are Pablos Holman, Vincenzo Iozzo, and Joichi Ito.
First Android malware using generative AI, called PromptSpy, uses Gemini to adapt to any device UI to maintain persistence. It's likely a proof of concept.
ATM jackpotting attacks using malware like Ploutus stole over $20M last year, with 700+ incidents in 2025. Criminals physically access ATMs to install malware that forces cash dispensing. The FBI warns of rising cases and shares detection signs.
GAO urges NSF CIO to improve IT management, including cloud procurement and cybersecurity, to enhance efficiency and compliance.
Criminals created a fake RMM vendor, "TrustConnect," which is actually a RAT sold as a service. They used a professional-looking website and a legitimate code-signing certificate to appear authentic, distributing it via phishing campaigns.