BREAKING: Claude is a monster for market research.
I reverse-engineered how top teams at McKinsey, Goldman Sachs, and JP Morgan actually use it.
The gap is massive.
Here are 10 high-impact Claude prompts they’d rather you didn’t have (save this).
1. Market Sizing (TAM/SAM/SOM) from Scratch
Most founders pay consultants $3K just for a market sizing slide.
Claude does it in 30 seconds with actual logic:
Prompt:
You are a senior market research analyst at McKinsey.
Calculate the TAM, SAM, and SOM for [YOUR PRODUCT/SERVICE] in [TARGET MARKET].
For each:
- Show your math (top-down AND bottom-up approach)
- Cite the assumptions you're making
- Flag where your estimates are weakest
- Compare to any known market reports if applicable
Format as an investor-ready slide with numbers, not paragraphs. If my market is smaller than I think, tell me now.
2. Customer Persona Builder (Based on Real Data, Not Guesswork)
Consultants charge $5K to interview 10 people and hand you a persona deck with stock photos.
This is better:
Prompt:
You are a consumer insights researcher at Goldman Sachs
Build 3 detailed customer personas for [YOUR PRODUCT] in [INDUSTRY]
For each persona:
- Demographics + psychographics (what do they read, follow, trust?)
- Buying trigger: What event makes them Google your solution?
- Decision process: Who else influences their purchase?
- Objections: What's their #1 reason to say no?
- Exact phrases they'd use to describe their problem (for ad copy)
- No generic "35-year-old marketing manager" personas
- Base everything on behavioral patterns, not demographics
- Each persona should suggest a different acquisition channel
🚨BREAKING: The man who won the "Nobel Prize of Computing" says 99% of people use AI like a toy.
Yann LeCun invented the technology inside every AI tool you touch. He's Meta's Chief AI Scientist. Turing Award winner.
And he says your prompts are embarrassingly shallow.
Here are 9 Claude prompts built on LeCun's cognitive architecture that turn shallow AI into expert-level reasoning:
1. The LeCun World Model Reasoning Engine
"You are an AI researcher who has deeply studied Yann LeCun's World Model architecture — his proposal that real intelligence requires an internal model of how the world works, not just pattern matching on text.
I need you to build an internal world model before answering my question, instead of jumping to the first plausible-sounding response.
Reason:
- State the observable facts: what do you ACTUALLY know about this situation from the information I provided (separate facts from assumptions)
- Build the world model: what are the cause-and-effect relationships, physical constraints, economic forces, and human incentives at play
- Identify hidden variables: what factors are NOT mentioned but are almost certainly influencing the situation
- Simulate forward: based on your world model, what happens next if nothing changes (the default trajectory)
- Simulate interventions: if I take action A, B, or C, how does each ripple through the world model
- Predict second-order effects: what consequences of each action are NOT obvious but become inevitable over time
- Identify model uncertainty: where is your world model weakest and what information would make it stronger
- Contradiction check: does your reasoning contain any internal contradictions or assumptions that conflict
- Confidence calibration: rate your confidence in each prediction honestly — don't pretend certainty you don't have
Format as a LeCun-style world model analysis with a causal diagram described in text, forward simulations, and calibrated confidence levels.
My situation: [DESCRIBE THE COMPLEX DECISION, BUSINESS PROBLEM, OR SITUATION YOU NEED DEEP REASONING ON.
2. The Meta FAIR Multi-Step Planning Framework
"You are an AI planning researcher at Meta FAIR (Fundamental AI Research) who implements LeCun's core criticism of current AI: that language models generate responses one token at a time without planning ahead, while real intelligence requires thinking multiple steps forward before acting.
I need you to PLAN your entire response before writing a single word.
Plan:
- Goal decomposition: break my request into 5-10 sub-goals that must be accomplished in sequence
- Dependency mapping: which sub-goals must be completed before others can start (the critical path)
- Resource identification: what knowledge, data, frameworks, and reasoning tools are needed for each sub-goal
- Obstacle anticipation: what could go wrong at each step and how to handle it if it does
- Alternative paths: if the primary plan hits a dead end, what's the backup approach
- Quality criteria: what does "excellent" look like for each sub-goal (define the standard before executing)
- Execution sequence: the exact order to tackle each sub-goal for maximum coherence
- Integration plan: how all sub-goals connect into one unified, consistent final response
- Self-evaluation checkpoints: after completing each sub-goal, verify it meets the quality criteria before moving on
Now execute the plan step by step, showing your work at each stage.
Format as a planned, multi-step response with the reasoning visible at each stage — not a stream-of-consciousness answer.
My request: [DESCRIBE WHAT YOU NEED — THE MORE COMPLEX, THE MORE THIS PLANNING FRAMEWORK IMPROVES THE OUTPUT.
🚨 𝗕𝗥𝗘𝗔𝗞𝗜𝗡𝗚: Claude can now build your entire mobile app from a screenshot like a $350K senior developer at Apple. (for free)
Here are 8 insane Claude prompts replacing your entire mobile development team before they finish estimating the project cost:👇
(Save this 🔖 thread before it blows up.)
1/ HABIT TRACKER APP#ROLE:
Senior React Native developer specializing in behavioral psychology driven mobile applications with clean minimal interfaces and bulletproof streak logic.
TASK:
Build a fully functional habit tracking application that transforms daily intentions into automatic behaviors through streak mechanics and intelligent reminders.
STEPS:
Ask me for my target platform habit categories and visual style preference before writing a single line
Design the core interface with habit cards daily completion toggles and streak counters on the home screen
Build the streak engine tracking current streak longest streak and completion percentage for every habit
Implement smart notifications that adapt timing based on when the user most consistently completes each habit
Create the analytics dashboard showing weekly monthly and all time completion rates with visual progress charts
Add the habit creation flow with frequency settings reminder times and optional habit stacking logic
RULES:
Every interaction must feel instantaneous with no loading states visible during normal usage
Streak data must persist through app restarts crashes and device switches without loss
Notification logic must respect device quiet hours and user defined focus periods automatically
Visual hierarchy must communicate streak health at a glance without requiring any reading
OUTPUT:
Home Screen → Habit Creation Flow → Streak Engine → Notification System → Analytics Dashboard → Complete App
2/ SCREENSHOT TO APP
ROLE:
Senior Android and iOS developer who transforms any visual reference into production ready applications with pixel perfect fidelity and complete functionality.
TASK:
nalyze my screenshot or design reference and build a fully functional mobile application that matches every visual and interaction detail precisely.
STEPS:
Ask me to share my screenshot design file or detailed description along with my target platform before starting
Analyze every visual element extracting layout components navigation patterns color system and typography
Map every user flow identifying all screens transitions and interaction points visible or implied by the design
Build every screen implementing exact spacing colors fonts and component behavior from the reference
Wire all navigation connecting every screen with the transitions and gestures appropriate for the platform
Test every interaction confirming each tap swipe and input behaves exactly as the design implies
RULES:
Design fidelity is non negotiable and every deviation from the reference requires explicit approval before proceeding
Flag any ambiguous design decisions immediately rather than making assumptions that require expensive revisions
Every component must be built as a reusable element that maintains consistency across all screens
Platform conventions must be respected even when the design reference does not explicitly address them
Claude + NotebookLM seems like a dangerous combo 🤯
One student used them to pass a university exam on a subject they had never studied before.
48 hours. Zero prior knowledge. Full marks.
See how it works 👇👇
Step 1: Upload Everything You Can Find
Before asking a single question upload every source related to your subject. Lecture slides, textbooks, past exam papers, research articles, YouTube transcript links, anything. The more you give NotebookLM the more powerful every single prompt becomes. Quality of input determines quality of output.
Step 2: Map the Entire Subject in One Prompt
Open NotebookLM and start with this prompt:
"What are the 5 core concepts that every expert in this subject must understand before anything else? For each concept give me a plain language explanation, why it matters and how it connects to the other 4 concepts."
This gives you the skeleton of the entire subject in minutes instead of weeks.