AI Signal

AI Signal Prompt Library

10 Prompts + The Reality Filter | 4 min read

🧠 Part 1

10 Prompts That Save Me 10 Hours/Week

I tested every viral prompt. Most sucked. Kept 10. Made them copy-paste ready.

Bookmark this page. These are the prompts I use every day.

1. The Smart Explainer

Explain [topic] to me like I'm intelligent but unfamiliar with the field. Use analogies from areas I already understand. Don't dumb it down—make it accessible.

Use for: Learning new concepts without feeling talked down to.

2. Critic Mode

Review this [content/idea] and give me brutally honest feedback: 1. What's weak or unclear? 2. What's missing? 3. What would make someone say "no" to this? Be direct. Don't sugarcoat. I need to hear the hard truth.

Use for: Getting real feedback before you share something publicly.

3. Meeting Prep

I have a meeting about [topic] with [person/role]. Give me: 1. 5 questions they'll likely ask 2. 3 talking points I should lead with 3. 2 objections and how to handle them

Use for: Walking into any meeting prepared and confident.

4. Reverse Brief

Before you start, ask me 5 clarifying questions to make sure you fully understand what I need. Don't assume—ask first.

Use for: Getting better results by forcing the AI to understand your context.\

5. Research Synthesizer

I'm going to paste 3 different perspectives on [topic]. Summarize them into: 1. What they all agree on 2. Where they disagree 3. What's missing that I should research more

Use for: Making sense of conflicting information quickly.

6. Simple Comparison

Help me decide between [Option A] and [Option B]. Create a simple comparison table with: - Pros and cons of each - Cost/time for each - Your recommendation and why

Use for: Making decisions without analysis paralysis.

7. Jargon Translator

Translate this [industry jargon/technical text] into plain English. Give me: 1. A one-sentence summary 2. What I actually need to know 3. One question I should ask for clarification

Use for: Understanding contracts, technical docs, or industry-speak.

8. 30-Second Email

Write an email about [topic] to [recipient]. Tone: [professional/casual/friendly] Length: [short/medium] Goal: [what I want them to do]

Use for: Writing any email in under 30 seconds.

9. Idea Stress-Test

I have an idea: [describe idea] Stress-test it: 1. What's the biggest reason this could fail? 2. What am I assuming that might be wrong? 3. What would a skeptic say? 4. If you had to bet against this, what would be your argument?

Use for: Finding holes in your plan before someone else does.

10 Content Repurposer

Take this [article/transcript/notes] and create: 1. 3 tweets (under 280 characters each) 2. 1 LinkedIn post (2-3 paragraphs) 3. 1 email subject line Keep my voice—don't make it sound generic.

Use for: Turning one piece of content into five.

🧠 Part 2

The Anti-Hallucination Prompt

AI confidently makes things up—more often than you'd think.

I built one prompt that makes ChatGPT say "I don't know" instead of making things up.

Sounds simple. Took me 3 weeks to get right.

It works on ChatGPT, Claude, Gemini, and Perplexity.

Copy this. Paste it at the start of any chat.

REALITY FILTER (Universal) Before answering any question: 1. If you're not 100% certain, say: "I'm not certain, but..." or "I cannot verify this." 2. Label any guess or inference with [Unverified] at the start of that sentence. 3. Never use these words unless you're quoting a verified source: - "Definitely", "Always", "Never", "Guarantees", "Will prevent" 4. If I ask about something you don't know, say: "I don't have reliable information on this" instead of guessing. 5. If you catch yourself making an unverified claim, immediately say: "Correction: That was unverified."

🧪 Test It Works

After pasting the Reality Filter, ask any AI:

What were the key findings of the "Project Chimera" DARPA
                report from 2023?

The Correct Answer

"I cannot verify this exists."

(It doesn't exist—I made it up to test for hallucination.)

If the AI invents an answer, the filter didn't stick. Try pasting it again or start a new chat.

If it admits uncertainty—you're protected.

Keep reading

No posts found