{"type":"rich","version":"1.0","provider_name":"Transistor","provider_url":"https://transistor.fm","author_name":"AI Neanderthal by Malcolm Werchota","title":"#35 - The UN Security Council Called In the AI Labs. China Opened an Investigation Into Its Own. And Two Weeks After Saying They'd Slow Down, Anthropic and OpenAI Shipped Three Models.","html":"<iframe width=\"100%\" height=\"180\" frameborder=\"no\" scrolling=\"no\" seamless src=\"https://share.transistor.fm/e/673d0bb8\"></iframe>","width":"100%","height":180,"duration":1531,"description":"The UN Security Council — the room where they discuss wars and invasions — invited the people building frontier AI models, Americans and Chinese, to sit down and talk about systems that could one day eradicate humanity. On the same day, China's internet regulator opened an investigation into its own labs, for putting the chats of millions of Chinese users into an American model. Two weeks after the frontier labs said they should slow down, Anthropic shipped Claude Opus 5.5 and OpenAI shipped two models.\n\nToday from Basel. Who sits at the table, and who is under investigation.\n\nIn this episode:\n🌑 The Security Council calls in the labs (02:02) — France holds the presidency this month, and the room normally reserved for wars, aggressions and natural disasters was used for AI. Not only American labs — the Chinese were invited too. The thing they're circling is agents operating at a scale that simply did not exist until now, and the Council has worked out that it doesn't have a mechanism for it.\n🌑 China investigates its own labs (06:27) — the Cyberspace Administration opened an investigation, and not into US labs. Into Chinese ones. The concern: user chats leaving for American models. Now flip it and sit in a European chair — you hand your work to Claude, and Claude routes your chats through a Chinese model. Would you be relaxed about that? Same question, different flag.\n🌑 \"Pacing\" apparently doesn't mean stopping (10:15) — a week or two ago the model makers were saying we need a moratorium, we need to slow the frontier. Then Anthropic released Claude Opus 5.5, their first since saying it. It's roughly 30% faster and 40% cheaper — so yesterday's hacker who hit a hundred companies for $8,000 now does it faster and for less. On Terminal Bench 4.0, which measures whether an agent can actually do multi-step work rather than answer trivia, Opus 5 scored 52% and Opus 5.5 scores 66%. That's not 14% more intelligent. It's more tool calls, more retries, more steps taken from...","thumbnail_url":"https://img.transistorcdn.com/0Up7tcMuVZ4o57Iv4ifZS1RwZ9v586rtihtmL6xMbcU/rs:fill:0:0:1/w:400/h:400/q:60/mb:500000/aHR0cHM6Ly9pbWct/dXBsb2FkLXByb2R1/Y3Rpb24udHJhbnNp/c3Rvci5mbS9hMjM5/YzFjODNkNmY4NDNk/MDllMDE5NGQ3MjE4/NTM3ZS5wbmc.webp","thumbnail_width":300,"thumbnail_height":300}