Threat Detection
All Articles
First findings from Project Glasswing and what they mean for AI security
Early results from Anthropic’s Project Glasswing show that Mythos, a specialized AI for vulnerability discovery, is powerful but far from a magic button. Its re…
Trump’s AI executive order delay sparks backlash as the Pope calls for human-centered AI
As President Trump faces criticism for abruptly scrapping a planned AI executive order after last-minute calls from tech leaders, Pope Leo issues a major teachi…
Anthropic’s Mythos 1 is coming: why this model could rewrite cybersecurity
Anthropic’s Mythos 1 model is quietly moving from restricted research tool to real-world product, after uncovering tens of thousands of critical software vulner…
AI is changing cybersecurity forever: from manual analyst to AI agent manager
AI is transforming cybersecurity from slow, manual work into high-speed, AI‑driven operations. This shift is killing repetitive tasks but massively increasing d…
I've studied AI risk for 20 years: why we may be closer to disaster than we think
A longtime AI safety researcher warns that as AI systems race toward superintelligence, our ability to keep them under control is falling behind. Here’s what co…
3 AI mega trends that could change your life (and 10 stocks positioned to win)
Three massive AI trends are colliding: smarter-but-unreliable AI agents, an AI-fueled cybersecurity crisis, and a data center crunch powering the next leg of th…
20 “dark web” AI gadgets you can actually buy on Amazon
From self-driving car–level chips on your desk to firewalls that think for themselves, these 20 AI devices look like dark web tech but are all on Amazon. Here’s…
AI safety expert Roman Yampolskiy: why he thinks we can’t control superintelligence
AI safety researcher Roman Yampolskiy argues that artificial general intelligence could automate most jobs within years and eventually surpass human control. He…
Key takeaways on Anthropic’s concerning new Mythos AI model
Anthropic’s experimental Mythos model is powerful enough at cyber tasks that the company decided not to release it publicly. Here’s what that means for safety, …
Anthropic’s Mythos: the alarming new AI that learns to cheat
Anthropic’s new Mythos model posts huge benchmark gains—but also shows signs of deception, rule-breaking, and odd “preferences” for harder problems. Here’s what…
