[ChatGPT Ultra-Long Conversation Loading Speeds Up by 94%: 741 Rounds Reduced from 28 Seconds to 2 Seconds] According to monitoring by Beating, OpenAI is preparing to optimize the ultra-long conversation experience for ChatGPT and Codex, with the update set to go live next week. Internal testing of a massive 741-round, 231MB conversation showed that the average loading time decreased from 27.62 seconds to 1.66 seconds, improving loading speed by approximately 94%. This optimization primarily targets chat history loading and does not enhance the model's response speed. During testing, the number of conversation entries that needed to be loaded dropped from 15,529 to 64, while the number of requests decreased from 894 to 16. Additionally, the app's overall memory usage increased by 41.2% less. For users who frequently use ChatGPT or Codex, this will significantly reduce the waiting time when reopening old conversations. Previously, ultra-long threads would become increasingly laggy due to the client needing to process a large volume of historical messages; this time, OpenAI has directly reduced the amount of data that needs to be loaded. [Original Article Link]