One big story this week was Michael Ulin leaving Paxton, the legal AI company he co-founded, to build AI agents that tackle paperwork and scheduling for office workers. Ulin said he loves building new things from scratch, explaining: "I’m happiest in the 0-to-1 trenches." His unnamed startup aims to create helpers that work quietly in the background so people can focus on harder tasks. This change comes after Paxton got $22 million earlier this year, proving investors still believe in AI for law even as leaders shift.

Mind Security from Seattle grabbed attention by raising $30 million to fight data leaks using self-operating AI. Backed by big-name investors, their system watches how companies use data and jumps in to block suspicious activity automatically. Unlike old security tools that just send alerts, Mind’s AI makes smart decisions on its own—like having a digital guard that learns while keeping watch.

While most action happened in America, the week showed how AI agents are branching out beyond simple tasks. From protecting secrets to handling busywork, startups are teaching AI to act more independently. These developments hint at a future where AI doesn’t just suggest solutions but takes smart actions on its own—changing how we work and stay safe online.