3 points | by slake 12 hours ago ago
3 comments
Probably? Reinforcement learning creates bots with specific styles. For example, ChatGPT is very fond of "typically", "unpack this", and "if you want".
Once again a post with literally 3 points and 2 hours old is the top of /ask
Why is the HN algorithm such ass, can we talk about that?
Well it did have Claude both in the title and the description...
Probably? Reinforcement learning creates bots with specific styles. For example, ChatGPT is very fond of "typically", "unpack this", and "if you want".
Once again a post with literally 3 points and 2 hours old is the top of /ask
Why is the HN algorithm such ass, can we talk about that?
Well it did have Claude both in the title and the description...