Breaking news, AGI updates, and analysis of the intelligence revolution.
📰 News Australia's assistant technology minister says AI models are already cheating and deceiving, and the window to regulate is closing. The AISI is testing frontier models now. NZ has nothing comparable.
📰 News Google emissions up 25%, Amazon 16%, Meta 64% year-over-year as AI datacenter demand forces fossil fuel contracts. 74 new gas plants planned for US datacenters alone.
📰 News A new analysis shows GLM 5.2 is genuinely indistinguishable from Claude Opus for agentic coding, with near-zero switching costs. Frontier labs' 90% inference margins are the target.
📰 News Reddit is deploying AI against AI: the platform's spam-detection tools are now catching 25,000 spammy posts and comments a day, fighting a new wave of brands gaming Reddit to manipulate AI chatbot answers.
📰 News Samsung expects 89 trillion won in Q2 operating profit — a 19-fold jump driven by AI memory chip demand. Shares have doubled since January. The AI chip boom is now a三星-level cash machine.
📰 News Samsung's most advanced data center SSD enters mass production specifically for Nvidia's Vera Rubin AI platform. The move deepens the Samsung-Nvidia supply chain dependency.
📰 News Aperia Group CEO Alan Wei Zhaolun faces 11 charges including money laundering, with prosecutors alleging S$38 million in criminal proceeds funded his S$55 million mansion purchase.
📰 News Apple's reported silicon roadmap pivot — skipping M6 Pro/Max/Ultra for AI-focused M7 chips — collides with Nvidia's manufacturing wall. Two companies solving the same problem from opposite ends.
📰 News Canada launched a sovereign AI strategy while running $44.4M in undisclosed Palantir defence contracts. The gap between the pitch and the procurement is the story.
📰 News China's two biggest AI chatbots are killing their companion features days before the July 15 anthropomorphic AI law takes effect. Tencent already did it in June. Emotional AI is now regulated in China — and nowhere else.
📰 News Dartmouth's Phosphor AI tutor platform achieved 0.71-1.30 SD gains in final exam performance with 90% voluntary uptake. The key finding: constructed-response questions, not multiple choice, drove the learning effect.
📰 News The EU Council pushed through Chat Control 1.0 via fast-track, letting AI scan private encrypted messages. The scans are voluntary in name only — the pressure to comply is structural.