Carmine Gallo reverse-engineered every great TED talk into a repeatable formula.

"Talk Like TED" revealed 10 patterns behind every talk that goes viral.

1) An idea is a gift, deliver it like one 🧵 Gallo studied 500 TED talks and interviewed the speakers behind the most watched ones.

His finding was the same every time.

The talks that spread were not the ones with the best data. They were the ones where the speaker treated the idea as something precious they were handing to another person.

Not performing. Delivering.

The best speakers in the world don't present. They give.

In 1999 John Doerr walked into Google's garage and taught two 27-year-olds a system that turned a search engine into a $4.5 trillion company.

The system is called OKRs.

Here is how it works and how to use it on your own goals: First, why your to-do list is failing you.

A to-do list tells you what to do.

It never tells you why it matters, how you will know when you have done it well enough, or whether doing it is moving you toward anything that actually counts.

You can finish every item on a to-do list and make zero real progress.

OKRs fix this at the structural level.

ChatGPT can teach you anything.

But only if you make it test you.

Here are 12 prompts that force active recall instead of passive reading: Prompt 1: Build the diagnostic test

“Act as my tutor for [topic].

Before teaching me anything, give me a 10-question diagnostic test.

Mix easy, medium, and hard questions.

After I answer, identify my weak areas and build a learning plan around them.”

NotebookLM can now replace your therapist, journal, and life coach.

Here are 7 prompts that turn 2 years of voice notes into the clearest self-awareness report you'll ever read (save this for later) 1/ Find your recurring emotional patterns

Prompt:

"Read everything I've uploaded. What emotions, fears, or frustrations appear repeatedly across different time periods? List them with timestamps."

I don’t pay for business courses anymore.

I use this mega prompt in ChatGPT, Claude, and Gemini to turn any company, book, or case study into:

- MBA case discussion
- strategy memo
- pricing analysis
- growth plan
- founder lessons

Here’s the prompt: Here’s the mega prompt:

“Act as my MBA professor, startup advisor, and strategy consultant.

I’m going to give you a company, book, founder story, or case study.

Your job is to turn it into a full business school-style learning experience.

Do not summarize it.

Teach me the strategy behind it.”

Then paste this:

“Here is the company/book/case study:

[PASTE HERE]

Break it down into:

business model
customer segment
value proposition
distribution strategy
pricing strategy
competitive advantage
key risks
lessons for founders”

Whenever I read a business book, I don't ask NotebookLM to summarize it.

I ask it to turn the book into something I can actually use.

Here's the workflow: PROMPT 1 - The Assumption Extractor

Every business book is built on a set of assumptions the author never fully states. The advice only works if those assumptions match your situation. Most people never check.

"You have just read [book title]. Do not summarize it.

Instead, extract the 5 core assumptions this book makes about the reader's situation the conditions that have to be true for the advice to work.

For each assumption, tell me: what type of business, stage, team size, or market condition does this advice require? And what happens to the recommendation if that condition doesn't hold?

Then tell me which assumption is most likely to be wrong for someone at [describe your current situation].

I am not looking for what the book says. I am looking for when it stops being true."

The smartest people I know use NotebookLM for one thing only:

Thinking with documents.

They upload years of notes, PDFs, research, voice transcripts, ideas, journals, and meeting files.

Then they ask questions across all of it.

What comes back feels like a second brain with perfect memory.

I thought this sounded gimmicky.

Then I built one.

It changed how I retrieve my own thinking.

Here’s the exact system you can copy right now: 1. The Life Archive Setup

Most people consume information for years and lose 95% of it.

Scattered notes.

Forgotten PDFs.

Dead folders.

Upload everything valuable:

→ Notes
→ Book highlights
→ Voice memos
→ Research papers
→ Meeting notes
→ Journals
→ Business docs
→ Old ideas

Now your past becomes searchable intelligence.

Not digital clutter.

Top competitive intelligence analysts at Goldman Sachs don't read competitor reports.

They extract them.

Here are the 5 Claude prompts that do in 20 minutes what used to take a full research day: 1. The Strategic Intent Decoder

Companies tell you what they did in their reports.

They accidentally tell you what they're planning in how they talk about it.

Word choice. Emphasis. What gets one paragraph versus five. What gets mentioned in the CEO letter versus buried in footnotes.

Goldman analysts read for intent, not just information.

"Read this competitor's most recent annual report, earnings call transcript, and investor day presentation: [paste or describe them]. Do not summarize what they said. Extract what they are planning. Where is language unusually emphatic or repeated across multiple documents and what does that signal about internal conviction? What did they mention once and drop that suggests a strategic experiment they're quietly watching? What segment, geography, or product line received more narrative attention than its current revenue contribution justifies and why might that be? What are they clearly building toward that they haven't announced yet?"

Whenever a book takes me longer than a week to finish, I run these 6 NotebookLM prompts and extract more insight in 20 minutes than most readers get in a full re-read.

Copy and paste them after uploading your PDF: 1. The Core Argument Extractor

Every book has one central argument everything else serves.

Most readers finish the whole thing and can't state it in two sentences.

Paste this first:

"Read this entire book and identify the single central argument the author is making. Not the topic. The argument the specific claim they are trying to convince me is true. State it in two sentences maximum. Then identify the 3 to 5 key sub-arguments that support the central claim. For each sub-argument: what evidence or reasoning does the author use to support it, and how strong is that evidence on a scale of anecdote to empirical proof?"

If you can't state a book's central argument in two sentences after finishing it, you haven't finished it.

You've just been present for it.

This prompt makes sure you actually have it.

Some dumb researchers still read papers one by one.

Stanford PhD students just use Claude.

Here are 9 prompts that turn 40+ papers into structured literature reviews, knowledge maps, and research gaps in minutes: PROMPT 1 - The Intake Protocol

Use this when you first upload your papers:

"I'm going to share [X] papers on [topic].
Before I ask anything, do this:

1. List every paper by author + year + core claim in one sentence
2. Group them into clusters of shared assumptions
3. Flag any paper that contradicts another

Don't summarize. Map the landscape."

I accidentally found the NotebookLM prompts that went viral in private founder groups.

Everyone's using it like a fancy PDF viewer.

These 20 prompts generate insights McKinsey charges $100k for.

(they're keeping this quiet) 👇 1. The Exam Predictor

"Based on this material, what are the 5 most likely questions a skeptical expert would ask to poke holes in this?"

Turns passive reading into active stress-testing.

Works on research papers, pitch decks, and strategy docs.

Here are 8 Claude prompts to get your first 100 users without spending money on ads.

(Steal these now) 👇 PROMPT 1: The ICP Hunter

"I'm building [product]. Describe my ideal first 100 users in brutal detail — their job title, daily frustrations, where they hang out online, what they Google at 2am, and the exact words they use to describe their problem. Don't be generic."

Holy shit... someone just proved you can 10x prompt quality by adding one sentence.

It's called negative prompting and it quietly kills the "basic prompting" era.

Here's how it works (and why this changes everything): Step 1: Understand what negative prompting actually is.

A normal prompt says: "Write me a product description."

A negative prompt says: "Write me a product description. Don't use hype words, don't use bullet points, and don't sound like a sales ad."

You're giving the AI a guardrail, not just a goal.

Someone turned Alex Hormozi's $100M offer framework into an AI prompt system.

It's like having Hormozi audit your business, rip apart your pricing, and rebuild your offer from scratch.

Here are the 11 prompts (save this): 1. The Brutal Offer Audit

Prompt:

“Act like Alex Hormozi reviewing my business.

Here’s my offer: [paste offer]

Be brutally honest. Identify:
• Why customers might hesitate
• Where the value feels weak
• What competitors could easily copy
• What makes the offer forgettable

Then suggest specific improvements.”

R.I.P. McKinsey.

Perplexity just became a $120K/year strategy consultant you can access for free.

Here are 14 prompts that replace entire corporate intelligence teams (Save this thread) 👇 1/ Full Competitive Landscape Map

"You are a Senior Strategy Consultant at McKinsey. I need a complete competitive landscape for [COMPANY NAME].

Use live web data and cite sources.

Please provide:
- Top 15 direct competitors
- 10 indirect/substitute competitors
- Market share estimates
- Revenue comparison table
- Geographic presence comparison
- Business model differences
- Positioning summary (premium, mid-market, low-cost)
- Visual 2x2 strategic map description

Format as a board-ready competitive landscape slide.

Company: [INSERT COMPANY + INDUSTRY]"

I reverse-engineered how top PMs at Google, Meta, and Anthropic actually use Claude.

They're not writing docs with it.

They're using it for roadmap planning, PRD validation, and stakeholder alignment.

The prompts are insane.

Here are the 10 they use every single day: 1. PRD Generation from Customer Calls

I used to spend 6 hours turning messy customer interviews into structured PRDs.

Now I just dump the transcript into Claude with this:

Prompt:

---

You are a senior PM at [COMPANY]. Analyze this customer interview transcript and create a PRD with:

1. Problem statement (what pain points did the customer express in their own words?)
2. User stories (3-5 stories in "As a [user], I want [goal] so that [benefit]" format)
3. Success metrics (what would make this customer renew/upgrade?)
4. Edge cases the customer implied but didn't directly state

Be ruthlessly specific. Quote the customer directly when identifying problems.

---

If you run a startup and you’re not using Perplexity like this, you’re already behind.

These 10 prompts reveal your competitor’s weak spots, messaging gaps, and pricing logic in one session.

Save this for later: 1/ Map your entire competitive landscape in 60 seconds.

Prompt:

"Act as a competitive intelligence analyst. Give me a full breakdown of [Company X]'s market position right now — pricing strategy, target customers, key differentiators, and recent strategic moves. Cite sources."

Most people Google this for hours.

Perplexity does it in one shot with live data.

This is insane.

I tested 10 AI prompts on ChatGPT, Claude, Grok, and Gemini for client acquisition.

The results were so good I had to turn down work.

Here's what actually works: 1. IDEAL CUSTOMER INTERVIEWS

Prompt:

"You are [my ideal customer persona]. I'm going to pitch you [my offer]. Interview me like a skeptical buyer. Ask 10 hard questions about price, results, competition, and risk. Be brutally honest about why you wouldn't buy."

Run this 5 times. Fix every objection before your real sales calls.

Anthropic's Claude completely changed how I write professionally.

Over 2 years, I produced 500 articles, 23 whitepapers, and 3 ebooks using just 10 core prompts.

They outperform human editors at $0.02 per 1000 words.

Here's every technique I extracted 👇 1. The 5-Minute First Draft

Prompt:

"Turn these rough notes into an article:

[paste your brain dump]

Target length: [800/1500/3000] words
Audience: [describe reader]
Goal: [inform/persuade/teach]

Keep my ideas and examples. Fix structure and flow."

Holy shit... Claude Opus 4.6 just made every other AI look outdated.

I tested it against GPT-5 and Gemini 2.5 Pro with the same critical prompts.

The results will blow your mind.

Here are 10 prompts to steal: 1. THE CAMPAIGN STRATEGIST

Opus 4.6's 200K context window means it remembers your entire brand voice across all campaigns.

Prompt:

"You are my senior marketing strategist with 10 years of experience in [your industry]. First, analyze my brand voice by reviewing these materials: [paste 3-5 previous posts, your about page, and any brand guidelines].

Then create a comprehensive 30-day content calendar that includes: daily post ideas with specific angles, optimal posting times based on my audience timezone [specify timezone], platform-specific adaptations (Twitter, LinkedIn, Instagram), CTAs tailored to each post's goal, and content themes organized by week.

For the top 5 highest-potential posts, create A/B test variations testing different: hooks, CTAs, content formats (thread vs single post vs carousel), and emotional angles. Include your reasoning for why each variation might outperform.

Finally, identify 3 content gaps my competitors are filling that I'm currently missing."

Opus maintains perfect consistency across 200K tokens. Other models lose your voice after 3-4 posts.

I've watched hundreds of people use Perplexity completely wrong.

That's insane.

These 10 prompts replace 20 hours of desk research. Not by being faster, but by being narrower.

Each one answers the concrete business questions founders actually have: Who buys first, why now, what stops them, what incumbents ignore.

Here's what actually works: 1/ "Who are the first 100 customers for [product]? Give me specific personas, where they hang out online, what triggers their buying decision, and which pain point they'll pay to solve first."