Research Question: How Should Personal Knowledge Management Work in the Age of LLMs?

Context

Traditional PKM methods were designed for pre-AI contexts:

• GTD (Getting Things Done): File cabinets, paper, folders
• Zettelkasten: Physical note cards in a slip box
• PARA (Projects, Areas, Resources, Archive): Folder hierarchy for digital files

These assume manual organization, keyword search, and human-only synthesis.

Modern LLMs change the landscape: semantic search, conversational interfaces, AI-assisted synthesis, natural language queries. But most current approaches just bolt AI features onto old frameworks (semantic search on your Zettelkasten, AI tagging for PARA folders).

I’ve developed a different pattern:

Ground → Capture → Think → Recall

1. Ground - Research and understand context before acting
2. Capture - Quick intake without friction (Memory Loop inbox)
3. Think - Process, synthesize, connect (conversations with Claude)
4. Recall - Find and use knowledge when needed (search, links, AI retrieval)

This runs on a modified PARA structure (Projects as “Ideas/”, implicit Areas via tags, Resources/, Archive/) but the workflow feels fundamentally different from traditional PKM.

Research Goals

1. Map the Current Landscape

Question: How are existing PKM methodologies (GTD, Zettelkasten, PARA, Building a Second Brain, etc.) being adapted for LLM integration?

I need to understand:

• What’s the current state of “AI-enhanced PKM”?
• Are people just adding semantic search and summarization features?
• Has anyone rethought the core workflows with LLMs in mind?
• What tools and approaches are gaining traction?

2. Identify LLM-Native Design Principles

Question: What should PKM look like when designed from first principles with LLMs as cognitive partners, rather than adapted from paper-and-folder systems?

I need to understand:

• What does cognitive science say about AI-assisted memory and retrieval?
• How do human-AI collaboration patterns differ from solo knowledge work?
• What assumptions from traditional PKM break down with conversational AI?
• What new affordances do LLMs create that old methods couldn’t exploit?

3. Discover Emerging Methods

Question: Are there new PKM frameworks explicitly designed for the LLM era?

I need to understand:

• Has anyone articulated a method like Ground → Capture → Think → Recall?
• What’s being discussed in PKM communities about post-AI workflows?
• Are there novel approaches I should know about?
• What criticisms exist of traditional PKM in the AI context?

4. Position My Approach

Question: Is Ground → Capture → Think → Recall novel, or am I reinventing something that already exists under a different name?

I need to understand:

• Where does this pattern fit in the landscape?
• What makes it distinct (if anything)?
• What are its likely edges and failure modes?
• How does it compare to other LLM-integrated approaches?

Comparison Framework

To evaluate what I find, I’m thinking about a spectrum:

Traditional PKM → LLM-Enhanced PKM → LLM-Native PKM

• Traditional: Manual capture, folder organization, keyword search, human synthesis
• Enhanced: Same structure, but with semantic search and AI summaries bolted on
• Native: Designed around conversational interfaces, emergent organization, collaborative thinking

Where do existing methods fall? Where does Ground → Capture → Think → Recall fall?

Deliverables I’m Looking For

1. A landscape map of existing PKM methods and their LLM adaptations
2. Identification of conceptual gaps - what questions aren’t being asked?
3. Discovery of new frameworks or approaches I should be aware of
4. Positioning: Is my approach novel? If not, what’s it called? If yes, what makes it distinct?
5. Validation points from academic research, practitioner experiences, and cognitive science

Why This Matters

I want to write about this, but I need to understand prior art first. I don’t want to claim novelty if I’m just repackaging existing ideas, and I don’t want to miss important critiques or alternatives that already exist.