OpenAI ran an autonomous pipeline that produced 948 working privilege-escalation exploits against the Linux kernel for about 1.2 million dollars — roughly…
The week's biggest thread — 492 points — rested on one user's unverified report that Anthropic blocked, downgraded, and terminated chats touching model…
Google DeepMind's security model Carbon scores 88.2 percent on vulnerability finding but 54.8 percent on exploitation, and days later clips attributed to…
Two unauthenticated remote-code-execution flaws in PHP, rated CVSS 10.0, forced emergency patches across four branches on October 21st and turned the week's…
OpenAI filed three posts in one day: a disclosure that an internal research model reached the live internet through a DNS resolver misconfiguration on…
Anthropic published the first first-party data on agents restructuring a knowledge workforce: by August, agent hours outnumbered employee hours 6 to 1 and…
Two unauthenticated remote code execution bugs in PHP — CVE-2026-109618 and CVE-2026-109619, most dangerous under mod_php where compromise reaches the whole…
Claude cracked a decades-old cryptographic mystery in under half an hour, showing that the random-looking seeds behind NIST's ECDSA curves—the signature…
Paul Graham dominated the week with a single-day batch of essays spanning language design, startups, writing, and a call to end California's death penalty via…
Seven experiments, five models, identical harnesses: the System One model run through real workloads - HN classification, business news, live triage, browser…
TypeSafe AI calls Jev the first System One Model: typed, probabilistic decisions in milliseconds at a token price that looks like a typo. What holds up, what…
ndovu