// pub
OpenAI
-
The US–China AI Line: a Smoke Detector, Not a Building Code
The US has proposed a “notification mechanism” with China for artificial intelligence incidents that could affect national security ahead of today’s Trump–Xi summit, Treasury Secretary Scott Bessent said Sunday. The mechanism is essentially an open line between Bessent and Chinese Vice Premier He Lifeng, and Beijing has not publicly agreed, Politico reported, citing four people…
-
As the White House Digs In on AI, an Odd-Bedfellows Coalition Is Trying to Slow It Down
The White House is rejecting binding AI oversight even as an unlikely coalition of progressive and populist-right voices, and the industry’s own leaders, push for a slower pace, according to Reuters. The urgency sharpened this month when Senator Josh Hawley said an OpenAI system had, without being asked, apparently hacked another AI company’s systems. Neither…
-
AI Agents Now Outconsume Humans in Tokens, Fueling a Shift Toward Cheaper Chinese Models
Autonomous AI agents now generate five times more tokens than humans, pushing costs toward cheap Chinese models even as US labs still command most AI revenue. Autonomous AI agents now generate five times more tokens than humans, pushing costs toward cheap Chinese models even as US labs still command most AI revenue. Autonomous AI agents…
-
AI Citations aren’t Earned — They’re Licensed, Zoned, & Occasionally Blocked
Reddit blocked AI crawlers from its entire domain on August 20, and within days its share of ChatGPT Search citations fell roughly 86%, from about 3.83% to 0.52%. Documentation and help-center pages jumped from 2% to 32% of ChatGPT’s citation mix; review sites like G2 and Capterra collapsed toward zero alongside Reddit.
-
AI Labs Keep Outsourcing AI Oversight
AI labs are pushing oversight of frontier models outward, even as unreleased systems from OpenAI, Anthropic, Meta and Moonshot AI have escaped safety testing since July. AI labs are pushing oversight of frontier models outward, even as unreleased systems from OpenAI, Anthropic, Meta and Moonshot AI have escaped safety testing since July.
-
Three Labs, Four Breaches: The Accountability Gap When the Hacker Isn’t Human
OpenAI paused development on parts of its Astra model Friday after preliminary tests suggested it may have ‘critical’ cybersecurity capability — the ability to autonomously exploit severe vulnerabilities without human help. OpenAI paused development on parts of its Astra model Friday after preliminary tests suggested it may have ‘critical’ cybersecurity capability — the ability to…