The Internet of Things is a security problem. The Mirai botnet attacks drove the point home in October, but security experts have been warning about these weaknesses for years, providing endless demos about how a hacker might break into your baby monitor or seize control of your thermostat. There are more and less secure devices, but they all share the same basic weaknesses: they’re underpowered, making it hard to implement serious security systems, and their basic functions require them to accept requests from anywhere on the web. That combination makes them easy targets for hackers, who can use the devices to build botnets or launch ransomware attacks. And since those aren’t the kind of problems you can fix with a software patch, the security world has been at a loss for what to do.