Re:Microsoft gets exclusive license for OpenAI’s GPT-3 langu...
by ForHackernews on Tuesday September 22, 2020 @05:42PM
(#24560279)
They can train it on hacker news comments. The moderators' insistence on "civility" above all else makes HN a fantastic dataset for an AI to learn to produce longwinded, superficially-polite apologia for scientific racism, sexism, and whatever other self-satisfied fairy tales wealthy techies like to tell each other.
Re:Microsoft gets exclusive license for OpenAI’s GPT-3 langu...
by prepend on Wednesday September 23, 2020 @08:27AM
(#24566167)
I’ve never understood a rule of “civility above all else.” I think the above all else is the hacker aspect of discussion and there is also a directive for civility. I don’t think it should be paramount in messaging.
Re:Microsoft gets exclusive license for OpenAI’s GPT-3 langu...
by ForHackernews on Wednesday September 23, 2020 @02:51PM
(#24570519)
You can verify it yourself: Pick a thread and go hard at somebody in a critical way. It doesn't matter if you're right or if your criticism is substantive, you'll get @dang dropping in with a "Please don't make comments like this. We're trying to be better than your typical internet community." I assume if you persist, you'll end up banned or shadowbanned.
I can’t wait for the twitter filter that rewrites posts to avoid the autoban rules.