The apparatus has been running since 1994 and has never detected anything. This is its principal achievement. The instrument was built to look for a class of event that theory permitted and nobody had observed. Such experiments have an awkward property: a positive result is a discovery, and a negative result is, in the ordinary sense, nothing at all. Funding bodies are not constituted to reward nothing, and the team spent much of the first decade explaining why an absence was worth paying for. The argument, once made carefully, is not difficult. A theory that permits an event also predicts, at least loosely, how often it should occur. If the instrument is sensitive enough to have seen the event and did not, the theory's parameters are constrained. Run it long enough and the space of allowed values shrinks. Nothing is observed, and yet something is learned; the learning consists precisely in the shrinking. Over thirty years the experiment has excluded a range of parameter values that once occupied most of the theoretical literature. Several models that were mainstream in 1994 are now dead, and they were killed not by a contradictory observation but by the steady accumulation of no observation at all. This is a slower and less celebrated way of making progress, and it is at least as reliable. The psychological difficulty is considerable and the team is unusually frank about it. Graduate students arrive, work for four years on an instrument that will almost certainly show them nothing, and leave with a thesis whose central result is a limit. Several have described the experience as difficult to explain to their families. The principal investigator has said that the hardest part of the job is not the physics but maintaining morale in the absence of any event to celebrate. There is a further hazard, and it has caught comparable experiments elsewhere. An instrument that expects to see nothing will, eventually, see something spurious, and the temptation to treat it as real is powerful in proportion to how long the wait has been. The team's response is a protocol under which the analysis is performed on deliberately falsified data until every choice has been fixed, and only then is the seal broken on the real record. Nobody on the team can tell you what the instrument has recorded until the analysis is already committed. This procedure has been invoked eleven times. On each occasion the apparent signal disappeared when the real data were examined. The eleventh was in 2022 and was, briefly, the most exciting week in the laboratory's history. The experiment is funded to 2031. The current limits are quoted in essentially every paper in the field. The instrument itself sits in a decommissioned mine, monitored remotely, and continues to record, with great precision, the absence of anything at all.