They summarize our coverage. We write it.
Newsletters like this one rebroadcast our headlines - often without the full review, the source reading, or the analysis underneath. Our weekly briefing sends the work they paraphrase, straight from the desk, before they get to it.
Free, weekly, no spam. One email every Tuesday. Unsubscribe anytime.
Rankings of 14 frontier LLMs by adversarial robustness - how well they resist jailbreaks, prompt injection, and harmful-behavior elicitation across HarmBench, AdvBench, StrongREJECT, JailbreakBench, and AgentHarm.
Sam Altman's World project launched World ID 4.0 at a San Francisco event on April 17, signing Tinder, Zoom, DocuSign, and Okta as partners while introducing Agent Kit to authorize AI agents.
Ox Security found that MCP's STDIO transport executes arbitrary OS commands before validating the server, exposing 200K+ instances across every major AI coding tool.
The official @geminicli X account was compromised and used to promote a fake $CLI token on Pump.fun. Users quickly identified it as a scam.
Cal.com moved its core codebase to a private repo after five years of open source, arguing AI tools make public code 5-10x easier to exploit. The community isn't buying it.
OpenAI's GPT-5.4-Cyber is a fine-tuned defensive cybersecurity model with binary reverse engineering, lowered refusal thresholds, and restricted access through the Trusted Access for Cyber program.
OpenAI's GPT-5.4-Cyber is a restricted model fine-tuned for defensive cybersecurity with binary reverse engineering and reduced refusal rates, available only through identity-verified access tiers - a direct response to Anthropic's Mythos Preview.
The Information reports Anthropic is prepping Claude Opus 4.7 and an AI design tool for imminent release, while OpenAI launched GPT-5.4-Cyber yesterday - a restricted cybersecurity model that directly challenges Claude Mythos.
UC Santa Barbara researchers found 9 of 428 third-party LLM routers actively injecting malicious tool calls, draining crypto, and stealing AWS credentials from AI agent sessions.
Claude Mythos Preview is Anthropic's most capable model - restricted to 50 orgs via Project Glasswing, with 93.9% on SWE-bench Verified and thousands of autonomous zero-day discoveries.
A 26B MoE model fine-tuned on elite bug bounty reports and real evasion techniques runs locally in 16.7GB, delivering WAF bypasses, exploit chains, and zero refusals with internal reasoning blocks.
UC Berkeley researchers achieved near-perfect scores on eight major AI agent benchmarks without solving a single task, exposing systemic flaws in how the industry measures progress.
