logoalt Hacker News

BrokenCogs • yesterday at 6:38 PM • 2 replies • view on HN

Isn't this easy for frontier 'labs' to fix? If the user requests lots of writing and prose, then use an auxiliary model which has been trained on 'award-winning' writing instead of all the crap on the internet?


Replies

FeteCommuniste • yesterday at 7:06 PM

Award-winning writing has all kinds of different styles. The average of them will probably get you some boring median soup once again, though maybe a bit less bland than the LLMs trained on oceans of LinkedIn and Reddit slop.

➕ show 1 reply
krige • today at 6:59 AM

Oh boy, more IP theft and copyright infringement!