I’m writing to share some frustrating experiences I’ve had with ChatGPT, particularly with paraphrasing large documents. I provided a 40-page document in a straightforward, PDF format, expecting ChatGPT to paraphrase it while maintaining an academic tone. However, the results have been consistently disappointing.
Despite detailed instructions, ChatGPT often misses entire sections, sometimes starting as late as page 30, completely skipping earlier content. This is perplexing because when I input text directly into the chat, ChatGPT handles the task admirably. But when it comes to processing the same content from a document, the performance significantly drops.
As a ChatGPT 4 Turbo (Plus) user and someone well-versed in AI, I’m baffled and quite frustrated by this inconsistency. The quality seems to be deteriorating instead of improving. Can anyone shed light on why ChatGPT struggles with document-based tasks, despite excelling with direct text inputs? Any insights or similar experiences would be greatly appreciated.