Since biology operates on video streams much more than still frames, I think the natural interpretation of the statement is “how much power would a computer dra

Since biology operates on video streams much more than still frames, I think the natural interpretation of the statement is “how much power would a computer draw when doing the same kind of (continuous) facial recognition that a brain does?” While I think the number is off by orders of magnitude with current algorithms, I think it’s a totally fair way to look at things.