You say this as if shorts represent some systemic shift in attention spans or even anything more than a binning of the existing short-long distribution. But is

You say this as if shorts represent some systemic shift in attention spans or even anything more than a binning of the existing short-long distribution. But is there evidence for it? Has the distribution of video lengths ever been anything other than a possion random variable?