AI governance pushes forward as robots break speed records

Right now the tech map shows three big currents moving at once. In AI, OpenAI is in Washington arguing that the US needs mandatory safety standards because the systems might soon improve themselves faster than we can watch. At the same time the company's math breakthrough claim is being picked apart over who actually did the work. Microsoft and Sber are shipping new reasoning and app-building features, and Clearview is testing a police tool that links faces to social accounts using xAI's model. Over in robotics, Chinese humanoids just ran the hundred metres quicker than Usain Bolt at the World Robot Games, while a Moscow car wash network runs sixty-three robots across fifty-plus sites. Hardware is feeling the squeeze too: Huawei's top AI chip has jumped sixty percent to thirty-seven thousand three hundred dollars because high-bandwidth memory is scarce. On the security side, an Iranian crew called Mirage Kitten is pretending to be recruiters on LinkedIn, luring engineers with fake coding tests that drop malware. And up above, NASA's Swift Observatory is burning up after its final run. That's the picture today — safety debates, speed records, price spikes, and a hacking campaign all landing on the same day.