AI Coding Experiment Bricks Rooted LG Smart TV

This is less a story about a TV and more about the operational risk that arrives when AI assistants are handed write access to systems users only partly understand. The affected party here is the hobbyist tinkerer, the segment most likely to grant that access for convenience. The apology is the telling detail: accountability lands with the human who ran the commands, not the tool that issued them. Whether these assistants get better at flagging when a request exceeds the user's apparent grasp of the system is the open question worth watching.