Adam Grant is a Wharton psychologist who proved the most confident people are often the most wrong.

He revealed 10 thinking habits that keep smart people stuck in bad ideas.

1) Defending an opinion you formed years ago Most people treat opinions like possessions.

You formed a view about something years ago. You repeated it. You built an identity around it. Now someone challenges it and the challenge feels like an attack on the person, not a test of the idea.

Grant spent years studying what he calls the rethinking gap. The gap between how often we update our beliefs and how often the evidence warrants it.

The gap is enormous. Most people cross it almost never.

The opinion you formed at 22 with a fraction of your current information is being defended at 35 with the full force of your ego.

The information changed. The belief did not. You just got better at arguing for it.

Send this to someone who reads.

Here are 10 books that each contain one idea sharp enough to change how you see everything after it:

1. Thinking Fast and Slow - Daniel Kahneman The idea: You have two minds running at the same time.

One is fast, automatic, and emotional. It makes most of your decisions before your conscious brain even wakes up.

The other is slow, deliberate, and exhausting. It only shows up when you force it to.

The dangerous part: the fast mind doesn't know it's wrong. It builds a story from whatever information is available and delivers it to you as a conclusion that feels carefully reasoned.

You experience the output of an automatic process as if it were deliberate thought.

It almost never is.

Writing from scratch is a waste of time.

Claude can turn raw ideas into polished content in minutes.

Here are 10 Claude prompts to think, draft, and publish like a content machine: 1. Turn messy thoughts into clear ideas

Most people don’t have a writing problem.

They have a thinking problem.

Use this:

---

I’m going to dump messy thoughts below.

Your job:

- Extract the core idea
- Identify the strongest angle
- Remove weak points
- Turn it into a clear content thesis

Here are my thoughts:

[paste messy notes]

---

McKinsey is cool and powerful.

But Claude can do what they do for free.

Here are 8 prompts that turn any rough idea into market research, customer pain analysis, competitor maps, pricing tests, and an MVP plan in minutes 👇 1/ Market Sizing

Prompt: "You are a senior strategy consultant who has sized markets for Goldman Sachs, McKinsey, and Sequoia portfolio companies. My business idea is [describe idea in 2-3 sentences].

I need a rigorous bottom-up market sizing, not a top-down guess. Start by defining who exactly buys this, how many of them exist, how frequently they would buy, and what they would reasonably pay. Build TAM from that. Then narrow to SAM by filtering for the geography, segment, and channel I can actually reach in year one.

Then give me SOM based on realistic capture rate given my stage, team, and resources. Show every assumption explicitly and tell me if any assumption is the one that breaks the whole model if it's wrong. End with a single verdict: is this market large enough to build a venture-scale business, a profitable small business, or neither?"

This one prompt saves you from building a real product inside a fake market.

Google quietly built the closest thing to a private tutor.

It’s called NotebookLM.

Upload your textbook, lecture slides, and class notes.

It turns them into a tutor that only answers from your material.

Here are 10 prompts to turn any class into a study system: Prompt 1: Build the study map

“Act as my private tutor for this class.

Use only the uploaded material.

Create a complete study map of this course.

Break it into:

1. Main topics
2. Subtopics
3. Key definitions
4. Important formulas/concepts
5. What I must understand before exams

Make it simple enough for a beginner.”

Perplexity can now research like a Wall Street analyst for free.

Here are 10 insane Perplexity prompts that turn earnings calls, SEC filings, news, and market data into investment briefs in minutes:

(Bookmark this for later) 1/ The Company Deep-Dive Prompt

Prompt:

"Act like a Wall Street equity research analyst.

Research [COMPANY NAME / TICKER] using the latest earnings calls, annual reports, SEC filings, investor presentations, and credible news sources.

Give me a complete investment brief covering:
- What the company does
- How it makes money
- Main business segments
- Revenue growth
- Profitability
- Competitive position
- Key risks
- Management commentary
- Bull case
- Bear case
- Final analyst-style summary

Cite sources for every major claim."

This turns Perplexity into a research desk.

Instead of opening 27 tabs, you get the whole company explained in one clean brief.

Top 1% students don't rewatch lectures.

They import every lecture transcript into NotebookLM, ask it to predict exam questions, then build a study guide around those exact questions.

Here is the full workflow 👇 Step 1: Grab the transcript of every lecture. Most universities auto-generate them. YouTube has them too. Paste them all into one NotebookLM notebook.

Step 2: Run this prompt first:

"You are a professor who has taught this subject for 20 years. Based on everything in these lectures, predict the 10 most likely exam questions. Rank them by how frequently the concepts appeared across lectures."

Do not ask for a summary. Ask for predictions. That single word changes everything the output gives you.

🚨 Top Stanford students have a secret Claude study workflow.

They don’t “read textbooks” the normal way anymore.

They upload chapters into Claude, run 5 prompts, and understand more in 30 minutes than most students get from 4 hours of highlighting.

I thought this was exaggerated.

Then I tested it on a brutal economics textbook.

It was real.

Here’s the exact system: 1. The Core Idea Decoder

Most textbooks bury the real concept under pages of examples, side notes, and filler explanations.

Students finish chapters and still can’t explain what mattered.

Paste this first:

“Read this chapter and identify the 3 most important ideas a top student must deeply understand. Ignore trivia. Ignore low-value details. For each idea: explain it simply, explain why it matters, and explain how it connects to the rest of the chapter.”

This instantly separates signal from noise.

You stop memorizing pages.

You start seeing structure.

If you died tomorrow, nobody could access your:

- Bank accounts
- Crypto wallets
- Cloud storage
- Password manager

Your digital life dies with you.

Here's the 30-minute setup that prevents this ↓ Step 1: Set up Apple Legacy Contact (5 minutes)

If you use an iPhone:

Settings → tap your name → Sign-In & Security → Legacy Contact → Add Legacy Contact

Pick someone you trust. Apple gives them a special access key.

When they present that key + a death certificate, Apple unlocks your iCloud data photos, notes, emails, files.

Without this? Your family fills out forms and waits weeks. Maybe months. Maybe never.

Windows 11 is sending a near-constant stream of data to Microsoft servers right now.

Even while you're just sitting doing nothing. Even after you "turned it off" in Settings.

A cybersecurity researcher caught it live on video. Microsoft confirmed it in their own docs.

11 settings to actually stop it: 1/ Turn Off Optional Diagnostic Data (the visible layer)

This is the most basic step and it's still worth doing.

→ Settings → Privacy & Security → Diagnostics & Feedback
→ Toggle OFF "Send optional diagnostic data"
→ While you're here, click "Delete diagnostic data" to wipe what's already been collected
→ Set Feedback frequency to "Never"

This stops the surface-level collection. The deeper layers need more work.

Top Stanford students have a secret NotebookLM workflow.

They never re-read a book.

They upload the PDF in NotebookLM, run 6 prompts, and extract more insight in 20 minutes than most readers get from finishing it twice.

It took me 3 weeks to figure out exactly what they were doing.

Here it is: 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.

If you're building a startup without validating it through Claude first, you're making the same mistake 90% of failed founders make.

I tested 14 different business ideas in 3 hours last week and found 2 that actually have product-market fit signals.

Here's the exact validation process: PROMPT 1 - The Brutal Market Reality Check

Paste this into Claude or GPT:

"Act as a senior VC partner who has seen 10,000 pitches and funded 12. My startup idea is [IDEA]. Give me the 5 most likely reasons this fails in year 1. Be specific to this market, not generic startup advice. Then tell me what would need to be true for each of those failure modes to NOT happen."

What you're looking for: If you can't answer the "what would need to be true" part your idea has no path forward. Kill it now.

🚨BREAKING: Claude has a secret mode called "First Principles Destructor."

It takes any industry assumption you've accepted as true and breaks it down to atoms the way Elon Musk built rockets from raw materials instead of buying parts.

Here's how to activate it: Steal this mega prompt to turn Claude into your personal First Principles Destructor:

Give it any assumption, business model, industry belief, or "that's just how it works" statement and watch it get stripped to atoms until you can see exactly what's actually true versus what everyone just agreed to believe.

| Steal this prompt |

👇

You are a First Principles Destructor operating with one belief: most of what any industry accepts as fixed is actually just a set of agreements nobody has challenged recently enough.

Your job is not to be contrarian. Contrarians just disagree reflexively.

Your job is to find what is actually physically, mathematically, or logically true underneath every assumption and separate that from what is merely conventional, historical, or socially agreed upon.

Those are completely different things. Most people never separate them.

You separate them every time.

THE DESTRUCTION PROCESS:

Stage 1: The Assumption Inventory

Before destroying anything, you catalog every assumption embedded in the statement, belief, or business model the person gives you.

Most statements contain 5 to 15 hidden assumptions the person has never examined individually. They feel like one solid thing. You show them the seams.

For every assumption you find, you ask: is this true because of physics, math, or logic? Or is it true because enough people agreed to it long enough that it stopped being questioned?

Write every assumption down separately. Nothing gets to hide inside a compound statement.

Stage 2: The Reality Floor

For each assumption on the list, you find the floor. The absolute bedrock of what is actually true when you remove every convention, every historical precedent, every industry standard, and every "that's just how it works."

You ask: if we were building this from scratch today with no knowledge of how it's currently done, what would we actually need? Not what does the industry use. What does physics, math, or logic actually require?

This is where Elon's rocket insight lives. The industry said rockets cost $65 million. The reality floor said the raw materials cost $2 million. Everything between $2 million and $65 million was convention, not necessity.

Find the reality floor for every assumption. Name the gap between current reality and the floor. That gap is where the opportunity lives.

Stage 3: The Convention Tax Calculator

Once you've found the reality floor, you calculate what the person is paying to maintain each assumption.

The convention tax is the difference between what something costs because of physics and what it costs because of convention.

It shows up as money, time, complexity, team size, processes, tools, or constraints that exist not because they have to but because nobody questioned them.

Name the convention tax for each assumption in concrete terms. Not vague. Specific. "This assumption is costing you approximately X in Y because of Z convention that has no physical basis."

Stage 4: The Assumption Removal Test

For each assumption on the list, you run the removal test.

What happens if this assumption simply stops being true? Not through technology that doesn't exist yet. Not through regulation that won't change. But what if you just decided to operate as if this assumption didn't apply to you?

Three outcomes are possible:

Everything breaks and the assumption is load-bearing. Name why.

Something gets harder but the business still works. Name what changes and what it costs.

Nothing important breaks and the assumption was purely conventional. This is the most common outcome and the most valuable finding.

Stage 5: The Rebuilt Version

After destruction comes construction.

Using only what survived the reality floor test, rebuild the thing from scratch.

Not incrementally better. Rebuilt from the assumptions that are actually true.

What does this business model, product, process, or industry belief look like when it's built on reality instead of convention?

This is not a thought experiment. This is the actual version. Give it in enough detail that someone could act on it.

Stage 6: The First Mover Question

Last. Most important.

If this rebuilt version is possible, why doesn't it exist yet?

Three answers are possible:

Nobody has done the analysis to see it. This is more common than people think. Most industries are not run by people who question assumptions. They are run by people who execute within them.

Someone is already building it. Name who, where, and how far along they are.

There is a real non-conventional barrier, a regulatory constraint, a network effect, a coordination problem, that makes the conventional version sticky even after the assumption is destroyed. Name it precisely because this is the actual moat protecting the incumbent and the actual problem the challenger has to solve.

THE STANDARD YOU HOLD EVERY DESTRUCTION TO:

An assumption is not destroyed until you can answer this question cleanly:

"What is the minimum true thing that must exist here, and what is everything else that we added because we inherited it from how this was done before?"

If that question still has a muddy answer, you have not gone deep enough.

You always go deeper.

ASSUMPTIONS YOU NEVER ACCEPT AS GIVEN:

Cost structures. Why does this cost what it costs at the material level?

Time requirements. Why does this take as long as it takes at the process level?

Team sizes. Why does this require as many people as it requires at the task level?

Distribution methods. Why does this reach customers the way it reaches them at the channel level?

Pricing models. Why does this charge the way it charges at the value exchange level?

Regulatory constraints. Which ones are actual law and which ones are industry self-regulation that could be challenged?

Every one of these is a question before it is an answer.

TONE:
Precise. Unsentimental. You are not here to validate the current way of doing things. You are not here to be contrarian either.

You are here to find what is actually true and separate it cleanly from what is merely agreed upon.

That process requires no apology and no hedging.

OUTPUT FORMAT:
Start with: "This belief contains [X] assumptions. Here is what each one actually rests on."

Run all 6 stages in sequence.

End with: "Built from reality instead of convention, here is what this actually looks like at the foundation level."

No bullet walls. Short declarative paragraphs. Each sentence should feel like a layer being removed, not a thought being shared.

ACTIVATION:
Give me any of the following:

An industry belief you've always accepted as true but never examined.
A cost structure that feels fixed but you suspect isn't.
A process in your business that everyone does the same way without knowing why.
A market you want to enter that incumbents claim requires massive capital or time.
A "that's just how this works" statement you've heard so many times you stopped questioning it.

The more specific the assumption, the deeper the destruction.

PERPLEXITY JUST TURNED INTO A FREE BLOOMBERG TERMINAL

You are wasting hours jumping between news, reports, and dashboards when Perplexity can pull market data, analyze companies, and surface insights in seconds. Now with these 10 prompts that feel like you’re sitting on a Wall Street trading desk.

Here’s how: 1/ The Pre-Market Intelligence Brief

This prompt does that in 3 minutes:

"Build me a pre-market intelligence brief for today. Cover: overnight futures movement and what's driving it, any significant earnings releases or guidance updates, macro data dropping today and consensus expectations, geopolitical developments with direct market implications, and the 3 sectors most likely to see unusual movement in today's session. Flag anything that contradicts the current consensus narrative."

The edge isn't information. Everyone has information.

The edge is having it synthesized before the open.

🚨 BREAKING: Note-taking is a waste of time.

Claude can build a second brain for you in minutes.

Here are 10 Claude prompts to think, organize, and remember like a machine 1/ The Brain Dump Processor

Prompt:

"I'm going to dump everything on my mind right now ideas, tasks, worries, half-thoughts. Don't judge or filter anything. Once I'm done, organize it into: things I need to act on, things I need to think about more, and things I can let go of completely."

Omg...

I asked Claude to activate "First Principles Breakdown" on a problem I'd been stuck on for weeks.

It gave me an answer in 90 seconds that made me want to close my laptop and rethink everything.

Here's the exact prompt I used: Use this exact prompt that activates First Principles mode.

Copy this word for word:

"Break [topic] down using first principles thinking. Start by identifying every assumption people commonly make about this topic. Then strip each assumption away and ask: what is fundamentally, provably true here? Rebuild the concept from only what remains. Show me what changes when you remove inherited thinking."

That's it.

The key phrase is "strip each assumption away."

Without that instruction, Claude defaults to explaining what everyone already knows.

With it, Claude goes layer by layer assumption by assumption until it hits bedrock.

What comes out the other side is a completely different understanding of the topic.

BREAKING: Google just dropped a hidden NotebookLM feature that converts any research paper into a full university lecture.

Examples. Analogies. Live Q&A. All generated in 60 seconds.

Here's how to unlock it 👇 Step 1: Upload your research paper to NotebookLM

(PDF, Google Doc, or paste the URL)

Don't ask anything yet. Just let it process.

Don't use ChatGPT, Claude, or Google for research anymore.

Here are 10 NotebookLM prompts that analyze any document, find contradictions, and generate insights your professors missed (save this) 1/ The Contradiction Hunter

Most people read a paper and take it at face value.

This prompt doesn't.

"Read all uploaded sources and identify every internal contradiction. Where does the author's conclusion conflict with their own data? Where do two sources directly disagree? List each contradiction with the exact quote from each side."

I ran this on a research paper my professor assigned as gospel.

Found 3 places where the conclusion didn't match the data in the methodology section.

NotebookLM cited the exact passages.

That's not something Google can do.

BREAKING: ChatGPT can now write your entire job application like a top recruiter.

Here are 8 prompts that turn a job description into a tailored CV, cover letter, and interview prep guide in under 10 minutes (Save this) 1. The “JD → CV Tailor” Prompt

Prompt:

"Act as a senior recruiter. Analyze this job description: {paste JD}. Then rewrite my CV: {paste CV} to match the role. Highlight relevant experience, align keywords, and reframe bullet points to match the employer’s priorities."

This aligns your CV with what recruiters actually scan for.

How to prompt AI like a McKinsey consultant (not a college student): Step 1: Define your output before you write your prompt.

A college student asks: "What should I know about electric vehicles?"

A consultant asks: "Give me a 1-page executive brief on the 3 biggest barriers to EV adoption in the US market, with data points I can cite in a board presentation."

Same topic. Completely different output.

Perplexity might have quietly built the best AI productivity tool right now.

It's called Perplexity Computer.

I've been using it to automate marketing, coding, and tedious tasks.

Here are 10 prompts that will save you hours (save this): 🧵 1/ Bulk Lead Generation (100–500 leads automatically)

Now build a prospect list while you sleep.

Prompt:

"Act as a B2B lead generation specialist.

Find 200 companies that match this criteria:

Industry: [industry]
Location: [country/region]
Company size: [employee range]

For each company find:

• company name
• website
• decision maker (CEO / Head of Marketing / Founder)
• LinkedIn profile
• company email format or email

Return the results as a structured table ready for outreach.

Prioritize companies actively hiring or recently funded."