Post

@hasantoxr

More from @hasantoxr

Hasan Toor

@hasantoxr

Dec 20

Top engineers at OpenAI, Anthropic, and Google don't prompt like you do.

They use 10 techniques that turn mediocre outputs into production-grade results.

I spent 2 weeks reverse-engineering their methods.

Here's what actually works (steal the prompts + techniques) 👇

Technique 1: Constraint-Based Prompting

Most prompts are too open-ended. Engineers add hard constraints that force the model into a narrower solution space, eliminating 80% of bad outputs before they happen.

Template:

Generate [output] with these non-negotiable constraints:
- Must include: [requirement 1], [requirement 2]
- Must avoid: [restriction 1], [restriction 2]
- Format: [exact structure]
- Length: [specific range]

Example:

Generate a product description for wireless headphones with these constraints:
- Must include: battery life in hours, noise cancellation rating, weight
- Must avoid: marketing fluff, comparisons to competitors, subjective claims
- Format: 3 bullet points followed by 1 sentence summary
- Length: 50-75 words total

Technique 2: Multi-Shot with Failure Cases

Everyone uses examples. Engineers show the model what NOT to do. This creates boundaries that few-shot alone can't establish.

Template:

Task: [what you want]

Good example:
[correct output]

Bad example:
[incorrect output]
Reason it fails: [specific explanation]

Now do this: [your actual request]

Example:

Task: Write a technical explanation of API rate limiting

Good example:
"Rate limiting restricts clients to 100 requests per minute by tracking request timestamps in Redis. When exceeded, the server returns 429 status."

Bad example:
"Rate limiting is when you limit the rate of something to make sure nobody uses too much."
Reason it fails: Too vague, no technical specifics, doesn't explain implementation

Now explain database indexing.

Read 11 tweets

Hasan Toor

@hasantoxr

Dec 17

WARNING: After you use these prompts, you’ll never write the same way again.

This might be the most useful thing I’ve shared all year.

Here are 12 prompts turn any LLM into a full writing studio that works harder than you do:

1/ The “Voice Injection” Prompt

Gets the model to fully absorb your writing style.

“Here are 5 samples of my writing. Extract my tone, pacing, sentence structure, and emotional signatures. Confirm when my ‘voice profile’ is ready.”

This sets the foundation.

2/ The “Idea Engine” Prompt

Infinite hooks, titles, and angles.

“Generate 20 content ideas that could go viral in my niche. Mix curiosity, tension, and contrarian angles. No clichés.”

It’s like having a strategy intern.

Read 14 tweets

Hasan Toor

@hasantoxr

Dec 10

Holy shit... Someone just built an AI agent that monitors the entire internet for you 24/7 and only messages you when something actually matters.

It's called Scouts, and here's how it works:

Most tracking tools require exact parameters.

But real life isn't exact:
"Let me know when there is a round-trip flight from any bay area airport to any airport in Australia or New Zealand for a 9-12 day trip in March, April, May or June costing under $1000 total. Include all taxes and fees. Prioritize non-stop flights.

That's 3 airports × dozens of destinations × 4 months × fluctuating prices.

You'd need to check daily. Scouts does it automatically.

I set up a Scout to know when there is an original or faithfully restored Pac-Man arcade cabinet from the 1980s for sale. Include both upright and cocktail table versions. Avoid miniature reproductions or new replica builds unless noted for comparison.

It continuously monitors listings, filters out perfume/menswear/sunglasses, and emails me only when rare archival pieces appear.

I just reply to the email to refine what I'm looking for.

No manual searching. No missing drops.

Read 5 tweets

Hasan Toor

@hasantoxr

Dec 9

I used Elon Musk's actual thinking framework as AI prompts.

It's the closest thing to having a billionaire engineer rip apart your ideas and rebuild them from physics.

Here are the 15 prompts that changed how I solve problems:

1. "What are the physics of this problem?"

Musk strips everything to objective reality.

"I'm struggling to grow my newsletter. What are the physics of this problem?"

AI reveals the hard constraints, the real forces, and the non-negotiable bottlenecks.

2. "If I couldn’t rely on existing assumptions, how would I solve this?"

Assumptions are invisible cages.

"My pricing model is based on what competitors do. If I removed all assumptions, how would I solve this?"

AI breaks the mental autopilot.

Read 18 tweets

Hasan Toor

@hasantoxr

Dec 7

Claude Sonnet 4.5 is the closest thing to an economic cheat code we’ve ever touched but only if you ask it the prompts that make it uncomfortable.

Here are 10 Powerful Claude prompts that will help you build a million dollar business (steal them):

1. Business Idea Generator

"Suggest 5 business ideas based on my interests: [Your interests]. Make them modern, digital-first, and feasible for a solo founder."

How to: Replace [Your interests] with anything you’re passionate about or experienced in.

2. Industry Pain Points Analyzer

"Analyze the current [industry] landscape. What are the top 3 pain points customers face? Give specific examples and explain briefly."

How to: Fill in [industry] with a sector you want to research.

Read 12 tweets

Hasan Toor

@hasantoxr

Dec 7

OpenAI, Anthropic, and Google use 10 internal prompting techniques that guarantee near-perfect accuracy…and nobody outside the labs is supposed to know them.

Here are 10 of them (Bookmark this for later):

Technique 1: Role-Based Constraint Prompting

The expert don't just ask AI to "write code." They assign expert roles with specific constraints.

Template:

You are a [specific role] with [X years] experience in [domain].
Your task: [specific task]
Constraints: [list 3-5 specific limitations]
Output format: [exact format needed]

---

Example:

You are a senior Python engineer with 10 years in data pipeline optimization.
Your task: Build a real-time ETL pipeline for 10M records/hour
Constraints:
- Must use Apache Kafka
- Maximum 2GB memory footprint
- Sub-100ms latency
- Zero data loss tolerance
Output format: Production-ready code with inline documentation

---

This gets you 10x more specific outputs than "write me an ETL pipeline."

Watch the OpenAI demo of GPT-5 and see how they were prompting ChatGPT... you will get the idea.

Technique 2: Chain-of-Verification (CoVe)

Google's research team uses this to eliminate hallucinations.

The model generates an answer, then generates verification questions, answers them, and refines the original response.

Template:

Task: [your question]

Step 1: Provide your initial answer
Step 2: Generate 5 verification questions that would expose errors in your answer
Step 3: Answer each verification question
Step 4: Provide your final, corrected answer based on verification

---

Example:

Task: Explain how transformers handle long-context windows

Step 1: Provide your initial answer
Step 2: Generate 5 verification questions that would expose errors in your answer
Step 3: Answer each verification question
Step 4: Provide your final, corrected answer based on verification

---

Accuracy jumps from 60% to 92% on complex technical queries.

Read 13 tweets

Share this page!

Enter URL or ID to Unroll

Hasan Toor

Try unrolling a thread yourself!

More from @hasantoxr

Hasan Toor

Hasan Toor

Hasan Toor

Hasan Toor

Hasan Toor

Hasan Toor

Did Thread Reader help you today?

Don't want to be a Premium member but still want to support us?

Send Email!