The video demonstrates using InfraNodus to create knowledge graphs from ChatGPT conversation histories, highlighting the tool’s ability to reveal discourse patterns, topical clusters, and structural gaps. It begins by explaining the process of extracting messages from ChatGPT’s JSON chat logs, showcasing a script that searches for specific terms within conversations. The presenter then imports these messages into InfraNodus, which visualizes the data as a network graph. This graph allows users to filter content by speaker, view the evolution of topics over time, and identify connections between different conversations. The video emphasizes the value of this visualization in understanding the big picture of one’s interactions with ChatGPT, especially when exploring recurring themes like language or divine concepts. It also touches on the potential to compare graphs of different topics to uncover overlapping ideas and the importance of recognizing entities as nodes within the graph. Finally, the video suggests using InfraNodus to ask insight questions and further develop ideas, ultimately providing a method to synthesize and leverage the wealth of information generated through ChatGPT conversations.