Anirudh Kamath

Not only is this product regressive, it doesn't even work correctly. "mom" -> 98% male "joe mama" -> 63.7% male "woman" -> 99.1% male It's clear this product is just scikit-learn trained on heavily biased data, where it's just assuming everyone is male because of the way the data is engineered. Not only is this a bad product idea to begin with, it clearly doesn't even work.