In this monologue, Nate B Jones provides a detailed comparison of ChatGPT 5.1 and Gemini 3, focusing on how to prompt each model effectively. He highlights that Gemini 3 excels at processing messy, high-entropy context like logs, PDFs, and multimodal inputs, while ChatGPT 5.1 is better suited for clean, organized inputs and complex, multi-step tasks such as reasoning, coding, and narrative development. Jones outlines specific prompting strategies for each, including what to keep doing (e.g., defining role, audience, and tone), what to stop doing (e.g., dumping huge unfiltered context), and what to start doing (e.g., treating ChatGPT 5.1 as an internal function library and using Gemini 3 as an entropy eater). The key takeaway is to align the model with the type of entropy it handles best: context entropy for Gemini 3 and task entropy for ChatGPT 5.1, to maximize productivity and results.
Sign in to continue reading, translating and more.
Continue
