Of course not. Nobody is endorsing for that. The problem is these are hard problems to solve and forming a consensus is a hard problem in addition to it. If we keep flip flopping on every new data point we will have more misses than hits.