R to @m_morzywolek: Should be a tad faster as hnswlib is a tad faster than the faiss hnsw implementation

R to @spakhm: I guess they just grew up.

R to @jobergum: I am very curious how Voyager compares to Faiss in terms of search time/quality. Can't wait to try it!

RT by @evil_genius_v12: It would be beneficial for RAG to enable the model to recognize previously asked questions and responses without including them in the prompt as chat history. It would help save tokens for extended context.

R to @m_morzywolek: Oh yes, God PLEASE!!!!

R to @m_morzywolek: What would you like them to store?

R to @GregKamradt: Long term memory is my personal wish 😁

R to @m_morzywolek: Yes https://www.browserbear.com/help/articles/383-can-my-task-automatically-on-schedule/

R to @yongfook: Can I run these on schedule?

R to @morsegamma1: @m_morzywolek I Dey among first 5 people wey comment You no even look my side oga o🥺💔

R to @DmytroKrasun: I'm always amazed at how these MRR charts are so flat at the beginning, and then they suddenly start to rise. In your case, what caused this breakthrough that made your MRR start to grow 📈?

R to @hwchase17: Can I do it using @weaviate_io vector db? 😁

R to @m_morzywolek: I know that I'm still early in the process so I’m gonna stay positive and keep working 🚀💪

http://Knowbase.ai is growing! I'm a week from a modest app launch and have over 200 users 🎉 . I must admit it's a great feeling when your creation is useful to others. Lessons learned: 1. Start planning marketing long before the launch 2. Try to talk to customers as much as possible 3. Distribution is king - invariably

R to @m_morzywolek: Useful innovation

R to @m_morzywolek: so cool