You are viewing a single comment's thread from:
RE: Tags of Broken Dreams :: A Curation Observation
I found myself wondering if it would be worth building some simple filters and maybe even AI to build a better client that screens some of the "crap". It'd be easy enough to screen for most memes. There are tutorials for training AIs with Tensorflow to summarize long text and also tutorials that train AI on the intent of the text. I wonder how hard it would be to train something on the "quality" of a post. Given a large enough sample size, you might be able to build something that separates the wheat from the chaff (oh crap - did I use an old man saying there... yup. I sure did).
Man, things like that have been clamored for by the curating community for a long time (and regular folks as well). I can't say how easy or hard it would be to train something to 'find' good content, but as a curator I know in the first 10 seconds if something is going to be worth reading or not. At the very least, it'd be nice to have the chaff filtered out, leaving you with things that at least look like wheat.
Build it, and let me know when you have it done :)