Ask HN: ML Practictioners, how do you serve “large” models in production?4 points·by dankle·3 năm trước·0 comments
dankle·năm ngoái·discussBro releasing software under MIT. Others picking it up and use it under terms of MIT license. Bro gets upset.Can someone please explain why?
dankle·2 năm trước·discussSo many misconceptions and misunderstandings in this article. One has to wonder who are paying them. Maybe just the collecting societies at it again.
dankle·2 năm trước·discussJust use k8s for prod. Docker compose is great for local dev, i would never dream of deploying to prod with compose.
dankle·2 năm trước·discussThats exactly how it works. See russa, turkey, hungary now. Germany in 1930s.
dankle·2 năm trước·discussSo you copy pasted the first paragraph of text from the article. Well done!! /s
dankle·2 năm trước·discussNot a fan of the ”I care therefore I get to throw tantrums” rhetoric. But yes, I also care and expect my colleagues to do so as well. But being mad doesn't strike me as a good way of facilitating change.