This is absolute chaos...

Summary

The main topic is the high token consumption of GBT 56, contrasted with older models like GBT 55. The transcript highlights the struggle to stay within rate limits despite adjusting settings. The practical takeaway is that users need to optimize their prompts and usage to conserve tokens and maximize their subscription benefits, as OpenAI's current implementation is inefficient.