[R] I used to work “part time” (time doesn’t exist when you have a deadline) on Machine Learning research, and a lot of that area is just completely useless. A good portion of the articles pumped out are just “in this heavily synthetic framework, my model beats the state of the art by 0.1% accuracy”, and the only project I contributed to which was actually worthwhile never got any attention because it “was not novel” but most likely also because we didn’t have some big name researcher as a co-author. For a field about automating society I sure threw a lot of labour away there, and that’s the norm.