3 ms·All models are wrong, some models are useful.by pixl97 3y agoAll models are wrong, some models are useful.godelski 3y agoThere's a corollary I use with respect to ML You don't need math to train good models. You do need math to know why your models are wrong.hansvm 3y agoIf you have any other favorite snarky tidbits I'd love to hear them. Do you blog?