Add redundancy

Where a failure would really hurt, add a second check that fails for different reasons than the first one.

You are four hours into a desert loop hike with two friends when your phone drops to 6% in the heat. The whole route lives in one app on that one phone. Your friends brought water and snacks, and both assumed you had navigation handled. You sit on a rock doing math about daylight. Nothing has gone wrong yet. But you can see, clearly and late, that every part of getting home safely runs through a single warm rectangle in your pocket.

Add redundancy is a repair move: where a failure would actually hurt, put a second, genuinely independent path behind the first. The cue is noticing one component whose failure takes everything with it. Do three things. Name the specific failure that would hurt, here losing the route. Ask what the backup depends on, and reject any backup that dies from the same cause, so a second navigation app on the same phone does not count. Then add a path with different failure reasons: a paper map and compass, plus the route downloaded on a friend's phone. What changes is that a dead battery becomes an annoyance instead of an emergency. This is wrong when stakes are low, since duplicating everything costs more attention than the failures ever would.