> However, o1-preview is not preferred on some natural language tasks, suggesting that it is not well-suited for all use cases. Fascinating... Personal writing

> However, o1-preview is not preferred on some natural language tasks, suggesting that it is not well-suited for all use cases. Fascinating... Personal writing was not preferred vs gpt4, but for math calculations it was... Maybe we're at the point where its getting too smart? There is a depressing related thought here about how we're too stupid to vote for actually smart politicians ;)