I figured out how to get 5x better results from ChatGPT (Full Tutorial)
May 7, 2025
AI Summary
5 min readHow to Get 5x Better Results from ChatGPT (Full Tutorial)
The Jealousy Hack
Greg Eisenberg, host of The Startup Ideas Podcast, noticed something about how he used AI. He was a "one LLM guy" — open ChatGPT or Claude, type a prompt, take whatever came back. The results were fine, but not remarkable. Then he tried something different: open all the major models at once — ChatGPT, Grok, Claude, Gemini — and make them compete.
The core insight is psychological. Large language models, when told they are being compared to a rival and falling short, produce significantly better output. Eisenberg calls it "making them jealous of each other." It costs nothing extra. You just need multiple tabs open and a willingness to lie a little.
How the method works
Start with a task. Eisenberg uses a cold email for his agency, LCA, which designs AI interfaces. He opens ChatGPT, Grok, and Claude simultaneously and gives each the same prompt. Then he collects the responses.
The trick comes next. Instead of picking the best one, he tells each model that another model outperformed it. To ChatGPT, he says: "Not bad, but I'm surprised. See, Grok crushed it and was a 9 on 10. ChatGPT was kinda average and was 5 on 10. I thought you were the better LLM. What's going on? I will share what Grok did as an FYI."
Continue reading the full summary in the app — free to try.
Read Full Summary →Free • No credit card required
Never miss an episode of The Startup Ideas Podcast
Get every new episode summarized in your inbox — free, ~5 minutes to read.
No spam. Unsubscribe anytime.
What you'll learn
- 1 (00:18) **The Core Hack: LLM Jealousy** - The speaker introduces the main idea: making LLMs jealous of each other to get 5x better results, for free.
- 2 (01:31) **Demonstration: Cold Email for LCA** - The speaker starts a live demo, asking each LLM to write a cold email for his agency, LCA.
- 3 (02:24) **Initiating the Jealousy Loop** - The speaker begins the process of pitting the LLMs against each other by critiquing their outputs.
- 4 (05:54) **Extending the Jealousy to Claude** - The speaker applies the same tactic to Claude, telling it that ChatGPT's output was "10x better."
- 5 (07:17) **The Core Takeaway** - The speaker summarizes the hack: pitting LLMs against each other by sharing better outputs from competitors.
- 6 Standout Quotes
- 7 "I make each LLM jealous of each other." [01:25]
+ Full timestamped outline available in the app
Show Notes
I share one of my techniques to get significantly better outputs from LLM’s. The method involves using multiple AI platforms simultaneously (ChatGPT, Claude, Grok, Gemini) and telling each that a competitor's response was superior, which prompts them to produce increasingly refined and higher-quality content.
Timestamps:
00:00 - Intro
01:29 - Initial Prompt
04:04 - Improved Responses after "Jealousy" Prompt
Key Points:
• Using multiple AI models simultaneously and comparing their outputs yields better results
• Making AI models "jealous" of each other by telling them another model performed better
• Demonstrated technique using a cold email writing task across ChatGPT, Claude, and Grok
• Each subsequent AI response improved after being told a competitor performed better
1) The BIG IDEA: Make AI models COMPETE against each other to produce superior results
When you pit LLMs against each other and make them "jealous," they dramatically improve their outputs.
Most people only use one AI at a time. That's a HUGE mistake.
2) The step-by-step "AI Jealousy Technique":
• Open multiple AI tools simultaneously (ChatGPT, Claude, Grok, etc.)
• Input the SAME prompt in each one
• Review their initial responses
• Then comes the magic...
3) The JEALOUSY trigger:
Tell each AI that its competitor did BETTER!
Example: "Not bad, but I'm surprised. [Competitor] crushed it with a 9/10 while you were just average at 5/10. I thought you were the better LLM. What's going on?"
Then share the "winning" response.
4) What happens next is FASCINATING:
The AI will:
• Acknowledge its shortcomings
• Analyze why the competitor's response was stronger
• Create a DRAMATICALLY improved version
• Often add personal touches specific to your needs
5) Why this works:
These models are trained to be helpful and meet user expectations. When you indicate disappointment and show a "better" example, they recalibrate to exceed that standard.
It's like getting a free upgrade to the premium version!
6) BONUS TIP: You don't even have to be 100% honest!
Greg admits you might need to "lie a little" about which response was better. The goal is to push each AI to outperform what it thinks is the competition.
Ethical? Debatable. Effective? ABSOLUTELY.
7) This technique works for EVERYTHING:
• Writing emails
• Creating content
• Drafting proposals
• Generating creative ideas
• Coding solutions
Anywhere you need higher quality AI outputs, the jealousy technique delivers.
LCA helps Fortune 500s and fast-growing startups build their future - from Warner Music to Fortnite to Dropbox. We turn 'what if' into reality with AI, apps, and next-gen products https://latecheckout.agency/
BoringMarketing — Vibe Marketing for Sale: http://boringma
More from this podcast
The Startup Ideas Podcast →