logoalt Hacker News

mk_chantoday at 12:15 AM0 repliesview on HN

Just shows the offset openai feels like it has to add to ‘equalize’ the average discourse of its training material