DIY ChatGPT: Enhancing RAVEN's long-term memories and starting to work on self-reflection | Artificial Intelligence Masterclass

All right, gang, here we are. So I guess we'll do a quick recap of where we left off.

It goes beyond the title with direct discussion of going, okay, memories, including: We're working on Raven, the long-term chatbot, cognitive architecture, etc., etc.

1. 01and then we do this recent chunk, which is just the foremost recent one, and we reconstitute that in our prompt file, which is prompt response
2. 02One is external sources and two is long memories There a lot that we can do with long memories So last time I took a couple of notes about like you know let do memory chunking let'
3. 03Because Raven could theoretically read a Unix timestamp, but unless the underlying model has been trained to translate a Unix timestamp, it's not going to make any sense