#15: OpenAI fine-tuning on function calls, RekaAI introduces multimodal assistant Yasa-1, RAG vs Long Context LLM's, RA-DIT: Retrieval Augmented - Dual Instruction Tuning