OpenPipe has successfully applied Reinforcement Learning with Human Feedback (RLHF) to analyze Hacker News (HN) posts. The model used a reward signal to predict the success of HN stories, achieving an impressive 3/10 score. By analyzing the results, OpenPipe found that service complaints and indie app-related content were reliable indicators of front-page success. Additionally, the model identified "diamonds in the rough" stories that deserved more discussion but didn't receive upvotes.

Source: https://openpipe.ai/blog/hacker-news-rlhf-part-1