Learning at Warp Speed

AI is the best teacher in history. Here is how to use it.

---

Why Traditional Learning Fails (And How AI Fixes It)

Let's be honest: most learning is inefficient. You read textbooks that put you to sleep. You watch lectures that could have been emails. You memorize stuff you forget by next week.

The problem isn't that learning is hard. It's that traditional learning methods don't match how human memory actually works.

The Forgetting Curve Hermann Ebbinghaus discovered in the 1880s that we forget information exponentially without reinforcement. Within 24 hours, we lose about 50% of what we learned. Within a week, we're down to 25%.

Traditional education doesn't account for this. You sit through a 2-hour lecture once and are expected to remember it 3 weeks later for an exam.

The Problem with Passive Learning Reading is passive. Listening is passive. You can zone out and still appear to be learning. Your brain isn't consolidating the information into long-term memory.

AI changes this by making learning active. You interact, question, apply, and get instant feedback. Each interaction strengthens the memory.

The Access Problem Historically, learning required finding experts—tutors, professors, mentors. They were expensive and unavailable on demand. AI provides on-demand expertise for any topic.

---

The Feynman-AI Method: Understanding Through Teaching

Richard Feynman famously said: "If you can't explain it simply, you don't understand it well enough."

The Feynman Method is the most powerful learning technique ever discovered. AI supercharges it.

The Classic Feynman Technique

1. Learn it: Read, watch, or study a topic
2. Explain it: Teach it to someone else (or pretend you're teaching)
3. Identify gaps: Where did you get confused? What couldn't you explain?
4. Go deeper: Return to the source material to fill those gaps
5. Repeat: Keep teaching until you can explain it simply

How AI Enhances the Feynman Method

Step 1: Get the Explanation

Ask AI to explain the topic in simple terms:

Explain [topic] as if you're teaching a smart 16-year-old who has no background in this field.

Use:
- Analogies from everyday life
- Concrete examples
- No jargon without explaining it

Stop when you've covered the fundamentals. Don't go deep yet.

Step 2: Explain It Back (The Critical Step)

This is where the magic happens. After getting the explanation, close the AI window and explain it OUT LOUD to yourself.

Use this exact prompt with AI:

I'm going to explain [topic] to you in my own words. After I finish each section, tell me:
1. What I got right
2. What I got wrong or confused
3. What I missed

Here's my explanation:
[paste your explanation or speak it]

Why this works:

• Speaking activates different neural pathways than reading
• The act of retrieving information (even to explain it) strengthens memory
• AI catches misconceptions before they become embedded

Step 3: The Pedantic Review

Once you've explained it, use AI to find gaps:

Review my explanation of [topic]. Be pedantic—find every simplification that might be technically wrong or missing important nuance.

For each issue I identified:
- Quote the exact claim I made
- Explain why it's wrong or incomplete
- Give me the correct understanding in simple terms

Then ask yourself: Did I really understand this, or did I just get lucky with vague descriptions?

Step 4: The Edge Case Challenge

To truly understand something, you need to stress-test it:

My understanding of [topic] is: [your simplified explanation]

Test this understanding with these challenges:
1. What's an edge case where this wouldn't hold true?
2. What would someone who disagrees with this theory say?
3. What's the most common misconception about this?

If you can't answer these, your understanding has gaps.

Real-World Feynman-AI Example

Topic: How blockchain consensus mechanisms work

Step 1: Initial explanation from AI AI explains: "Blockchain uses proof-of-work where miners solve cryptographic puzzles to validate transactions and add blocks to the chain..."

Step 2: You explain it back "Its like a competition where people try to solve really hard math problems, and whoever wins gets to add the next page to the notebook and they get rewarded with cryptocurrency. The puzzles are so hard that you need lots of computer power, which means you cant fake being multiple people."

Step 3: AI catches gaps "You described proof-of-work correctly, but you missed: The puzzles serve a specific purpose (preventing spam attacks), not just a general 'difficulty.' You also didn't mention that this process is energy-intensive because the puzzles are intentionally designed to be hard. A common misconception is that the 'math problems' are useful—they're not, they're intentionally wasteful."

Step 4: The insight "OH—so it's not that the puzzles are hard FOR NO REASON. They're hard because making it costly to add blocks prevents people from making fake versions of the ledger. The energy cost is a FEATURE not a bug. It makes the system expensive to attack."

That's the insight you wouldn't get from just reading or watching.

---

NotebookLM: The Research Weapon of Choice

Google's NotebookLM is underutilized and underappreciated. It's the most powerful research tool available for anyone learning complex topics.

What Makes NotebookLM Different

Unlike ChatGPT or Claude (which can say anything about any topic), NotebookLM becomes an expert on ONLY the sources you upload. This is critical:

The hallucination problem: When you ask ChatGPT about quantum physics, it can confidently make up stuff that sounds right. It has no way to verify if what it's saying is actually in its training data.

NotebookLM's solution: You upload 50 PDFs about a topic. When you ask a question, NotebookLM answers using ONLY those documents. If the documents don't contain the information, it says so. No hallucinations.

The Complete NotebookLM Workflow

Step 1: Gathering Sources

Upload everything relevant to your topic:

• Academic papers (PDFs)
• Book chapters
• Research summaries
• Your own notes
• Industry reports
• Any documentation you're studying

Step 2: The Deep Dive Query

Once uploaded, use these high-value queries:

Based on the sources, what are the 5 most important concepts I need to understand first?

For each concept, give me:
- A one-sentence definition
- Why it matters
- Where in the sources I can learn more

Find all contradictions between these sources, especially regarding:
- Methodology
- Conclusions
- Assumptions

For each contradiction, explain what the disagreement is and why it matters.

What questions would a professor ask on an exam about this topic? Create 10 questions with varying difficulty.

Step 3: The Audio Overview (The Secret Weapon)

NotebookLM can generate a "podcast" where two AI hosts discuss your sources:

"This is wild. Upload 5 research papers on a topic, click 'Generate Audio Overview,' and two AI hosts spend 15 minutes discussing the key findings, debates, and implications."

Why this is so powerful:

• Listening engages different learning modalities
• The conversational format helps with retention
• It's like having two experts discuss your exact reading material
• You can listen during commutes, exercise, chores

Use it for:

• Commute learning (turn 30-minute drive into learning time)
• Pre-lecture prep (listen to the research before class)
• Review sessions (reinforce learning from different angles)
• Complex topics that are hard to learn from text

Step 4: The Source Comparison

When you have multiple sources on the same topic:

Compare how Source A (academic paper) and Source B (industry report) approach [topic].

Specifically:
- What do they agree on?
- What do they disagree on?
- What does the academic paper emphasize that the industry report ignores?
- What practical insights does the industry report include that the academic paper misses?

Synthesize these perspectives. When would I trust one over the other?

Advanced NotebookLM Techniques

For literature reviews: Upload 20 papers on your topic. Ask:

Group these papers by theme. For each group:
- What common finding or approach do they share?
- How do the papers differ in methodology or conclusions?
- What's the timeline of how understanding evolved?

Then identify: Where is there consensus? Where is there debate? What questions remain unanswered?

For book learning: Upload chapters from multiple books on the same topic. Ask:

Each author seems to have a different emphasis. Compare:
- Author A's perspective on [topic] vs Author B's
- What evidence does each use?
- Who would agree with whom? Who would debate?

Give me a framework for understanding where each author is coming from.

---

Elicit & Research Rabbit: Academic Research Tools

For academic work and rigorous research, Elicit and Research Rabbit are essential.

Elicit: The Research Assistant

Elicit is designed specifically for academic literature review:

What it does:

• Finds relevant papers (no hallucination—it's searching real databases)
• Summarizes papers in structured formats
• Extracts key findings, methodology, and limitations

Best for:

• Literature reviews
• Research gap identification
• Methodology comparison
• Evidence synthesis

How to use it:

1. Start with a question: "What are the effects of intermittent fasting on metabolic health in humans?"

2. Get structured summaries:

   • Each paper summarized with: Main finding, Method, Sample size, Limitations
   • You can compare across papers systematically

3. Find contradictions: "Which studies found different results from others, and why?"

4. Identify gaps: "What aspects of this question haven't been studied yet?"

Research Rabbit: The Discovery Engine

Research Rabbit is about finding the right papers:

What it does:

• Start with one good paper
• It shows you related papers, authors, and topics
• Build a collection of relevant literature
• Visualize connections between papers

Best for:

• Starting a literature search when you have one relevant paper
• Discovering newer papers that cite or are cited by papers you know
• Understanding the intellectual landscape of a field

How to use it:

1. Start by uploading or searching for one seminal paper in your field
2. Research Rabbit shows you the "paper galaxy"—related papers, authors who cite it, papers it cites
3. Add relevant papers to your collection
4. Use the collection to guide your reading order

Why These Beat Regular Search

The problem with Google Scholar:

• Returns thousands of papers
• You have to read titles and abstracts manually
• Easy to miss the most relevant papers
• No synthesis

The Elicit/Research Rabbit advantage:

• Structured summaries let you consume literature faster
• Connected paper networks show you the intellectual map
• They surface papers you wouldn't find through keyword search

---

The 3-Tier Learning Loop:深度 Learning System

The most effective AI learning system uses three tiers of interaction:

Tier 1: The Foundation Layer (What is it?)

Start with the basic concept:

Explain [topic] in simple terms.

Include:
- A one-paragraph summary
- The core idea in one sentence
- Why someone learning this should care

Then give me 3 real-world examples of this concept in action.

Goal: Build an accurate mental model of what the thing IS.

Tier 2: The Detail Layer (How does it work?)

Once you understand the basics, go deeper:

Now explain [topic] in more depth.

Focus on:
- How it actually works (the mechanism)
- The key components or steps
- What variables affect the outcome
- Common misconceptions about how it works

Use a structural format (numbered steps, diagrams in text, whatever makes it clear).

Goal: Understand the mechanics, not just the summary.

Tier 3: The Edge Case Layer (Where does it break?)

To truly master something, understand its limits:

Now push on [topic]. Find:
1. When does this approach fail or underperform?
2. What are the most common mistakes people make with this?
3. What are the limitations of this approach?
4. What would someone who disagrees with this say?

Be specific and use evidence from the sources when available.

Goal: Understand the boundaries and develop critical thinking.

The Loop: Repeat Until Mastered

After going through all three tiers, you loop back:

1. Where did you get confused in Tier 2?
2. What edge cases did you not anticipate?
3. What questions do you still have?

I went through the 3-tier learning process on [topic]. Here are my remaining questions:

1. [Question about something I'm still confused about]
2. [Question about an edge case I don't understand]
3. [Question about how this connects to related concepts]

Answer these and point me to additional resources for the gaps.

---

The Spaced Repetition System

AI can help you remember what you learn. The key is spacing—reviewing material at increasing intervals.

How Spaced Repetition Works

1. Learn something new (Day 1)
2. Review it (Day 2) - if you remember it, the interval increases
3. Review again (Day 5) - if you remember, interval increases more
4. Review again (Day 14)
5. Review again (Day 30)

Each successful retrieval strengthens the memory. The brain learns "this is important enough to remember long-term."

AI-Powered Spaced Repetition

I just learned about [concept]. Create a set of 10 flashcards for spaced repetition.

For each flashcard:
- Front: Question that tests understanding (not just recall)
- Back: Answer with explanation
- Difficulty: Rate 1-3 (1 = easy, 3 = hard to remember)

Format as:
Q1: [question]
A1: [answer] (Difficulty: [1-3])

The Active Recall Method

Don't just re-read. Test yourself:

Here are 5 concepts I learned today:
1. [Concept A]
2. [Concept B]
3. [Concept C]
4. [Concept D]
5. [Concept E]

For each, without looking at my notes:
- Explain it in one sentence
- Give one real-world application
- Identify one thing you're not sure about

Then I'll tell you which ones you got right and which need more work.

---

Common Learning Mistakes (And How to Avoid Them)

Mistake #1: Passive Consumption

Reading and watching feel like learning. They're not.

What's happening: Your brain goes on autopilot. You see words and your brain says "yep, I know what those words are" without actually processing and storing the information.

The fix: After every piece of content, do something ACTIVE:

• Explain it out loud to yourself
• Write a 3-sentence summary
• Apply it to a real situation
• Teach it to someone (or pretend to)

Mistake #2: Assuming AI Knows Everything

AI can confidently make things up.

What's happening: LLMs have no internal fact-checker. When asked about obscure topics, they generate plausible-sounding text that might be completely wrong.

The fix: Always verify claims, especially:

• Statistics and numbers
• Academic citations
• Historical dates and events
• Technical specifications

Use Elicit or Perplexity (with web search) for factual questions.

Mistake #3: Not Building on Prior Knowledge

You forget things because you don't connect them to what you already know.

What's happening: Isolated facts don't stick. They're random data points your brain can't file away.

The fix: Every time you learn something new, explicitly connect it:

I just learned that [new concept] works by [mechanism].

This reminds me of [existing knowledge] because [how they're similar].
But it's different because [key difference].

This matters because [why this connection matters for understanding].

Mistake #4: Binge Learning Without Retention

You watch 5 hours of tutorials and feel like you've learned a lot. You have not.

What's happening: Passive consumption feels productive. But without active retrieval and spaced repetition, you're just building short-term familiarity, not long-term knowledge.

The fix: Never end a learning session without:

1. A 5-minute summary you write from memory
2. One real-world application you plan to try
3. A scheduled review reminder (set it in your calendar)

Mistake #5: Trying to Learn Everything

Information overload prevents any learning.

What's happening: You collect 50 bookmarks, save 100 articles, and follow 30 topics. You feel overwhelmed and learn nothing deeply.

The fix: The Pareto principle applies: 20% of any field contains 80% of the value.

I'm trying to learn about [broad topic]. Help me find the 20% that will give me 80% of the value.

What are the:
- 3 foundational concepts I must understand?
- 5 most important techniques or methods?
- 3 mistakes to avoid?

Ignore everything else until I've mastered these.

---

Building Your Personal Knowledge System

The AI Second Brain

Create a system where AI helps you learn, organize, and retrieve knowledge:

1. Capture: When you learn something important, write it down 2. Process: Use AI to extract key insights and connections 3. Organize: Store in a searchable system (Notion, Obsidian, etc.) 4. Review: Use spaced repetition to maintain knowledge

The Weekly Review Protocol

Every week, do this:

Review my learning from this week.

Topic 1: [What I learned]
- One-sentence summary:
- Why it matters:
- Where I'd use it:
- Gaps in my understanding:

Topic 2: [What I learned]
[Repeat]

What themes emerge across this week's learning?
What questions do I still have?
What should I focus on next week?

---

What's Next: You've mastered learning and creativity. Next week, we bring these skills into the workplace to save you 10+ hours a week.