2025-12-16 AI4PKM x CMDS 행사

18 Dec 2025 · 28 min
28:15

The video records an offline AI4PKM × CMDS community gathering where 김진영 (Jin‑young Kim) shared his AI‑driven personal knowledge‑management (PKM) system and led a deep discussion with experts from various fields. The talk is organized into seven sections. First, 김진영 defines the essence of knowledge management as a tool for narrowing the gap between one’s current state and goals, distinguishing weak (raw) knowledge from strong (contextualized, verified) knowledge, and warning against “knowledge overload” that leads to burnout. Second, he examines AI’s double‑edged role: AI‑generated summaries can cause passive accumulation of information, so actionable items must be attached to convert knowledge into growth. Third, he traces his personal AI workflow evolution—from ChatGPT to prompt‑driven notebook LMs, cloud‑code agents, and finally Voice Mode—highlighting the shift from static tools to dynamic partners and the paradox of over‑automation (e.g., n8n’s convenience vs. loss of attention). Fourth, he details the concrete automation stack: EIC (enhance‑improve‑correct) prompts, an orchestrator with Task‑Worker architecture, desktop GUI wrappers for terminal‑averse users, and clipping pipelines that auto‑summarize YouTube content into Obsidian. Fifth, he showcases Voice Mode as a game‑changer, enabling hands‑off tasks, 24‑hour AI interaction, and a surge in “true‑positive” outputs, while also noting stress trade‑offs. Sixth, he explains how trial‑and‑error is captured as reusable Cloud Code “skills” (e.g., markdown‑to‑slide, video‑cleaning, Lego‑assembly verification), turning failures into compounded assets. Seventh, he outlines a future vision of a “brain network” where individual PKM agents share updates, persona‑based assistants adapt to context, and real‑time AI assistants intervene during work. The session concludes with community reflections on result‑oriented vs. growth‑oriented AI, natural selection of knowledge, experience gaps, and the importance of balancing convenience with active learning. Throughout, practical resources (GitHub repos, monthly challenges) and a “good‑enough” philosophy are emphasized.

Chapters

More from Jin

youtube.com

AI 위클리 인사이트 — 2026년 3월 첫째 주

7mo ago
youtube.com

데이터 지능 팟캐스트 -- 정부와 공공기관을 위한 데이터 사이언스 with UNC 김재연 교수님

8mo ago
youtube.com

AI4PKM W2 Tutorial Part 2 : 슬라이드 및 비디오 만들기 & 비디오 편집용 스킬들

9mo ago
youtube.com

AI4PKM W2 Tutorial Part 1 -- 컨텐츠 수집(Ingestion) & 정리(Organization)

9mo ago
youtube.com

AI시대의 지식관리 (1부) -- 왜 AI시대에 지식관리가 최선의 생존스킬인가?

10mo ago
youtube.com

1121 18년만에 찾아온 6개월 휴식 & AI 시대 인간으로 흥하는 방법 (11월 창발 세미나)

10mo ago
youtube.com

왜 ChatGPT보다 10배 나은 AI 활용법 (feat 클로드 코드)

11mo ago
youtube.com

Agentic Future for PKM (PKM 컨퍼런스 발표)

11mo ago
youtube.com

AI for Knowledge Work (한글; Narrated by AI)

1y ago
youtube.com

1002 창발 Product 그룹 발표

1y ago
youtube.com

지식 업무를 위한 AI -- ChatGPT를 넘어서 AI 에이전트로

1y ago
See all content