Re:Reddit Is OpenAI’s Moat?
by ok123456 on Wednesday June 14, 2023 @12:10PM
(#36328078)
Any non-trival LLM that works by scraping the internet is already sufficiently advanced to be able to classify blogspam.
[ Parent
| on HN ]
Re:Reddit Is OpenAI’s Moat?
by giantrobot on Wednesday June 14, 2023 @02:00PM
(#36329895)
The issue Google has with blogspam is they don't want to filter it because it inevitably uses Google's advertising. So Google gets to self-deal traffic to blogspam which makes them money. They're not incentivized to actually eliminate blogspam.
LLMs aren't automatically disincentivized from training on blogspam so they're not going to avoid it either.