I kind of agree with you, startups have to fix 'mistakes' fast instead of just making others feel pleased? i.e. the so-called 'fire fast', otherwise it might fail more likely. In that sense, YC probably did a wrong move even though it's labelled as an experiment.