3 ms·
I think this is exactly right, and indeed this is a lot of the value. "Content-generation" is already a thing, and yes it doesn't need to make much sense. Appar
by rich_sasha 6y ago
I think this is exactly right, and indeed this is a lot of the value. "Content-generation" is already a thing, and yes it doesn't need to make much sense. Apparently people who read it don't mind.
- throwaway6e8f 6y agoPeople don’t read it, search engines do.
- masswerk 6y agoBTW, we should mandatorily tag generated content for search engines in order to exclude it from future training sets.
- hundchenkatze 6y agoApart from that, hopefully the people building training sets use gltr or something similar to prevent training on generated text. http://gltr.io/ http://gltr.io/
- a1369209993 6y agoWell, hopefully, someone will come up with a language model that picks words based on GLTR purpleness. Automated data collection deserves automated data.