AI Agent Memory and PKM Workflows Mature

First-person walkthroughs of Knowledge Infrastructure (The Curator, 1,272 pages), Recall second brain test, AI note-taking apps decision framework, local LLM PKM with Blinko/Open Notebook, two-agent chief of staff in Obsidian, RAG pipeline with confidentiality. New: Frankie Brain second brain example from dinner recap — real-world PKM anecdote. New: first-person comparison of Antigravity vs local LLM for Obsidian file tasks — tests context-awareness and respect for existing structure; Antigravity outperformed in understanding folder conventions and weaving new notes into schema. Replicable benchmark for AI agents on PKM vaults. New: first-person walkthrough of local RAG research workflow using Ollama, embeddings, and AnythingLLM — covers full pipeline from collection to retrieval, integrates with voice assistant, emphasizes data ownership. Growing ecosystem of replicable setups.