Been thinking about Bitchat, the Bluetooth mesh messenger. Could someone send a malicious message that corrupts the phones it passes through? In theory, yes. Every phone that relays a message has to read part of it, even if it can't decrypt the content. If there's a bug in that code, a crafted packet could crash the relaying phone or worse. And since every phone also passes messages on, a bad enough bug could spread from phone to phone like a worm. That's the downside of a mesh: you're processing data from strangers you never chose to talk to. It's not just hypothetical either. A possible buffer overflow was reported shortly after launch, researchers showed you could impersonate other users, and the app itself warns it hasn't had a proper security review. To be fair, it's open source, written in safer languages than old school C, and phones sandbox apps. The more realistic attacks are spam, impersonation and tracking who's nearby. What about AI? It can't do anything magical in the relay chain. It still needs a real bug to exploit. But AI is getting very good at finding bugs, so the cost of discovering one drops. Defenders get the same tool though. The bigger AI risk is if someone hooks an AI agent up to Bitchat. Then the message itself is the attack. No bug needed, just text written to trick the AI into doing something it shouldn't. What do you think? Are you using Bitchat?