3 ms·
Yikes. Unfortunately this is why (IMO) I see a lot of mediocre ML articles upvoted very highly on HN that have either very little discussion, or generally poor
by argonaut 10y ago
Yikes. Unfortunately this is why (IMO) I see a lot of mediocre ML articles upvoted very highly on HN that have either very little discussion, or generally poor quality discussion, especially around hot topics like neural networks. The discussion I've seen here is honestly the exception to the norm (probably due to the statistical nature of the topic). Personally this is one of the few areas where Reddit does better.
- TDL 10y agoAny suggestions which subreddits I should be looking at?
- argonaut 10y ago/r/machinelearning is pretty high quality, very technical. You'll see a lot of well-known researchers posting there sometimes (whereas here you only see karpathy every once in a blue moon). I'm only a random guy who did ml research as an undergrad.
- mindcrime 10y agoUnfortunately this is why (IMO) I see a lot of mediocre ML articles upvoted very highly on HN that have either very little discussion, or generally poor quality discussion, especially around hot topics like neural networks I think you're asking too much of upvotes. Since "upvote" and "save" are the same thing on HN, a lot of people upvote anything that meets the bar of "based on the title, it's plausible this is something I'd like to read eventually". That's really about all you can expect of upvotes.
- argonaut 10y agoI realize that's what many people do, but I don't agree with it because it's an implicit signal of interest/worthiness, due to affecting rankings. It gives people on HN the wrong perception of what is popular/worth learning and risks leading autodidacts astray.
- mindcrime 10y agoI actually agree and I would love to see HN adopt a separate "save" flag aside from "upvote" (like Reddit does), but... it is what is, for the moment anyway.