A Bayesian filter for stupidity

stupidfilter.org

13 pointsdfranke6 comments

I wonder if they could sell the reversed results to teenagers as an non-teen filter. Something that would guarantee they only see their "stupid" peers comments. No more articles about boring stuff, no more long winded intelligent debate just pure unadulterated LOLs.

Seems like a reasonable goal to me. I already know that a number of people use POPFile to do things like sort "interesting" from "uninteresting" messages (e.g. one journalist uses POPFile to automatically filter out press releases that don't interest him).

However, one issue with comments is that they can be short leaving not a lot of text to work with.

John.

I would think short comments are a stupid-positive flag all by themselves, though. It's very rare for few words to carry great meaning online; in this case, a manual "this is wise" button may be appropriate.

If I used that heuristic on email, most of PG's mail would be in my spam folder.

I agree, although there's nothing wrong with using 'comment length' as an input to some learning system. In POPFile we use quite a lot of 'meta' information about email including length to make a decision.

John.

did you keep your comment short to make your point?

Reply on news.ycombinator.com