The assumptions behind the Shift Project analysis (largely based on a 2015 paper, whose assumptions have been significantly revised in 2019 and 2020) contain a series of flaws, which, taken together, seriously exaggerate the electricity consumed by streaming video.