for me its not that ML is uniquely inherently bad, its more that I was really excited about its applications but realized that, in practice, it's usually pretty simple algorithms, and the more complicated applications dont have much direct application to socially positive stuff. prove me wrong!