Source: Artem Zhutov’s workflow (video, article)


The Loop

1. Pick a goal + expert

Start with a concrete personal goal and pick an expert whose content covers that domain. Artem used “improve my health” + Huberman (300+ episodes).

2. Load expert content into NotebookLM

From the terminal, Claude Code runs load_channel.py:

  • Scrapes the YouTube channel via InnerTube API → JSON of all video URLs
  • Creates a new NotebookLM notebook
  • Bulk-loads filtered videos as sources (NotebookLM processes the transcripts)

NotebookLM is now a queryable knowledge base. Every answer comes with [N] citations traced to the exact episode and passage.

3. Claude Code queries NotebookLM with the goal

Claude runs parallel queries against the notebook, each targeting a dimension of the goal. The cited responses get saved as markdown files in Obsidian under Notes/NotebookLM/<notebook>/QA/.

Result: a research folder in the vault with cited answers you can click through to verify.

4. Claude designs an interview

From the Q&A responses, Claude picks the most impactful questions across dimensions and presents them as an interview. You answer honestly about your current state.

5. Claude builds a profile + gap assessment

Claude takes your answers + pulls existing data from the Obsidian vault. Produces a structured assessment:

DimensionCurrent StateExpert TargetGap
Sleep regularityInconsistentSame time ±30minHigh
Morning lightNone10min within 30min of wakingHigh
CardioMinimal150-200min zone 2/weekMedium
Strength3x/week3x/weekLow

6. Top experiments become Obsidian notes

Claude proposes the highest-leverage experiments based on gap size. Each becomes a type: experiment note with:

  • Hypothesis (scientific method framing)
  • Protocol (specific daily actions)
  • Success criteria (measurable threshold)
  • Timeframe (2-4 weeks)
  • Observations (empty — filled daily)

Status set to in_progress so they surface in dashboards.

7. Morning routine skill closes the loop daily

A /daily skill runs every morning:

  1. Reads the Experiments folder, filters by status: in_progress
  2. Lists active experiments in the daily note
  3. Asks about each one: “How is this experiment going? Any observations?”
  4. You answer (mood, energy, sleep, etc.)
  5. Claude writes observations to the experiment note + daily log
  6. Based on observations, Claude schedules next actions (calendar events)

8. Dashboard aggregates everything

A Dataview-powered dashboard shows goals, active experiments, latest observations, and trend data over time.


Information Flow

Expert content (YouTube)
    ↓ load_channel.py
NotebookLM (cited knowledge base)
    ↓ nlm queries via Claude Code
Obsidian (Q&A notes with citations)
    ↓ Claude designs interview + protocol
Experiment notes (hypothesis, protocol, criteria)
    ↓ morning routine skill
Daily note (observations, mood/energy logs)
    ↓ Claude schedules actions
Google Calendar (blocks, routines)
    ↓ you live your life
Observations flow back into experiment notes
    ↓ dashboard aggregates
Dashboard (trends, gap closure tracking)

NotebookLM = read-only expert knowledge. Obsidian + Claude Code = read-write action layer. Claude Code bridges the two. The morning routine is the forcing function that prevents it from becoming another “I’ll get to it later” tab.


Application Ideas

Mellies — Branding Exercise via Oren Meets World

  • Notebook: Oren Meets World (101 videos — brand strategy, art direction, luxury branding, marketing playbooks)
  • Goal: Build a distinctive brand identity and launch strategy for Mellies
  • Interview dimensions: Brand aesthetic, packaging direction, social media strategy, luxury vs accessible positioning, content playbook
  • Experiments: A/B test packaging concepts, run a content cadence experiment, test brand voice across channels

Judaism — Upgrade routine via Menora

  • Notebook: Menora - Judaism (260 videos — Rabbi Kaminetsky, Torah portions, Jewish philosophy, family/ethics)
  • Goal: Deepen weekly Torah study and integrate Jewish principles into daily practice
  • Interview dimensions: Current study habits, Shabbat observance depth, ethical frameworks applied to business, family traditions
  • Experiments: Weekly parsha study tied to chapter of the week, Shabbat preparation routine, morning prayer consistency

part of tooling