Hacker News new | ask | show | jobs
by dumbfounder 1253 days ago
It isn't pictures, it is text only. I think there is a huge difference. I had to police content for Twicsy (a search engine with 10 billion Twitter images indexed) and I have seen some very bad stuff. There is a huge difference.
2 comments
realAzazello 1253 days ago
> It isn't pictures, it is text only

For ChatGPT related, yes/maybe. But for other content filetering work, it's primarily video & pictures: https://www.wired.com/story/social-enterprise-technology-afr...

link
absrd 1252 days ago
This is just false, if you read the article Sama also collected explicit, illegal images on behalf of OpenAI -- this was the reason the contract was cancelled.
link