I think the essay Physiognomy’s New Clothes is a necessary read for anyone who is currently working with machine learning or algorithmic technologies. I want to say that like physiognomy being a pseduo-science, machine learning, when it comes to scientific analysis, should be considered pseudo-technology. Its interesting to make connections between astrology as predictive and a psuedo-science like physiognomy as predictive. I think machine learning waffles between these two. On one end it could be used to gather existing information and formulate some kind of gesture towards meaning. On the other end it could be considered a kind of science in which variables and controls are monitored and results are tested. The latter end is obviously problematic and brings into question who are the creators and who are the subjects and for what are the results being used. Astrology wants to gesture at a person. Machine learning wants to tell about a person. What if we treated machine learning like astrology as opposed to a kind of science? What if prediction markets were based on astrological models?