You're tuned in to Tech Beat.
Microsoft's September security patches are causing fresh headaches for IT teams worldwide. The updates, meant to shore up Windows and Excel, have instead knocked out Remote Desktop Services for some users, silenced USB audio devices, and broken the basic paste function in Excel. It's a reminder that patching at scale is never as clean as it sounds.
Shifting to the bigger picture in artificial intelligence, Microsoft is out with a thirty-seven page humanist AI code of conduct, declaring that people matter more than AI. The announcement lands just as Anthropic's CEO Dario Amodei called for a coordinated slowdown in development, warning that model progress may be outpacing our ability to safely verify and control increasingly complex systems. Words and documents are easy. The harder question is whether either company slows down in practice.
Meanwhile, a quieter story is worth watching. An organization called METR is positioning itself as a key evaluator of AI capabilities, essentially helping determine what systems are safe to deploy. Critics argue that concentrating that gatekeeping role in a single private body creates a slow, largely invisible path toward centralized control over who gets to build what. The governance architecture of AI may matter as much as the technology itself.
Keep surfing. Tech Beat out.
