MMarketing Against The Grain
← All frameworks
Self-Mastery

Learning-Surface Routing Map

Choose research, source-grounded dialogue, or guided learning by learner intent

Difficulty
Easy
Time to result
~days to results
Steps
6
Confidence
95%

This map routes learners according to their current knowledge state and intent. When someone needs an evidence-rich perspective on a substantive question, deep research gathers and synthesizes sources. When the learner already has documents or a research artifact and wants summaries, follow-up questions, or grounded dialogue, NotebookLM provides a bounded knowledge assistant. When the learner faces an empty-box problem and does not know the domain well enough to formulate useful questions, a guided learning surface structures the topic and prompts exploration. These modes can form a progression: guided orientation produces better questions, research develops perspective, and source-grounded dialogue deepens understanding of the collected material.

Origin

Extracted from Marketing Against The Grain

Core principles

  • 01Different learning states require different interfaces
  • 02Deep research builds evidence-rich perspective
  • 03Source-grounded assistants support interrogation of known material
  • 04Guided experiences help learners who do not yet know what to ask

How to run it

  1. 1

    Diagnose the learning state

    Determine whether the learner has a precise question, a source collection, or only a broad desire to understand a topic.

    Pro tip Ask what the learner could currently explain without assistance.

    Watch out Do not assume unfamiliar learners can specify a complete curriculum.

  2. 2

    Route empty-box learners

    Use a guided learning experience when the learner does not know the domain's structure or relevant questions.

    Pro tip Follow generated branches that reveal foundational concepts first.

    Watch out A generated course may omit prerequisites or overstate disputed claims.

  3. 3

    Route research questions

    Use deep research when the learner needs breadth, evidence, competing perspectives, or current information.

    Pro tip Frame the question around a decision or explanatory goal.

    Watch out Do not confuse a large report with genuine comprehension.

  4. 4

    Route bounded source work

    Use NotebookLM when the learner wants to interrogate, summarize, or repurpose a defined set of materials.

    Pro tip Curate sources before asking questions against them.

    Watch out Answers remain limited by the uploaded source set.

  5. 5

    Progress between modes

    Move to another surface as orientation creates questions or research creates a useful corpus.

    Pro tip Preserve outputs so each stage becomes input to the next.

    Watch out Switching tools without a changed learning need adds friction.

  6. 6

    Test understanding

    Explain the topic, answer retrieval questions, or apply it to a fresh problem without relying on fluent summaries.

    Pro tip Use active recall and practical exercises.

    Watch out Ease of consumption is not proof of learning.

In the wild

Learning pasta from zero

A beginner who does not know what to ask starts with a guided Learn About experience that introduces ingredients, dough preparation, shaping, and cooking. Once the beginner encounters a specific question about flour types, deep research gathers evidence, and selected recipes are loaded into NotebookLM for comparison.

→ The learner receives structure first, evidence second, and source-grounded dialogue third.

Common mistakes

Starting beginners with a blank chat box

Learners with no domain map often cannot generate the questions needed to obtain a coherent education.

Using deep research as a tutor

A comprehensive report can provide perspective without creating a paced or interactive learning journey.

Ignoring source boundaries

A source-grounded assistant cannot reliably answer beyond the material supplied to it.

Is it for you?

Best for

Self-directed learners moving from unfamiliarity to evidence-based understanding and follow-up inquiry.

Not ideal for

Situations where a qualified human instructor, supervised practice, or formal accreditation is required.

From the transcript

“where deep research is like a deep dive on a topic to gain perspective, and then you can kind of ask follow-up questions on Notebook…”

27:00

“it solves a lot of the empty box problem of like, hey, I don't know anything about this thing, I don't even know what to…”

27:30

From the episode

Google's Secret AI Advantage (Why DeepMind Will Dominate)

→