23/23 were included in model training data though, so maybe the bus factor is a wash after all.