================================================================================ SANDBOX COMPARISON REPORT: CURIOSITY GAP DIAGNOSIS ================================================================================ Generated: 2026-02-13T12:48:36.987104 Clusters tested: 10 -------------------------------------------------------------------------------- EXECUTIVE SUMMARY -------------------------------------------------------------------------------- MAKE_NOW verdicts: Baseline (A): 0/10 Curiosity Mining (B): 4/10 Unclustered (C): 5/10 Average hookability score: Baseline (A): 4.22 Curiosity Mining (B): 4.16 Unclustered (C): 4.26 Title pattern analysis: Baseline (A): 8 generic, 0 specific Curiosity Mining (B): 7 generic, 0 specific Unclustered (C): 3 generic, 2 specific -------------------------------------------------------------------------------- KEY FINDING -------------------------------------------------------------------------------- βœ… HYPOTHESIS SUPPORTED: The revised prompts (B and C) found more 'MAKE_NOW' quality hooks than the baseline. The current pipeline's clustering + prompt combination appears to be abstracting away specific details that create viral hooks. -------------------------------------------------------------------------------- BEST EXAMPLES: Where B/C Found Better Hooks Than A -------------------------------------------------------------------------------- πŸ“Š money_traps_cluster_1 Topics: ['Invested $50k in 2021 and it’s now at $55k', 'withdrawal strategy?'] [A] BASELINE: "The Money Account Decision Tree: Where Your Cash Actually Belongs in 2025" Verdict: HIGH_PRIORITY, Score: 4.35 [B] CURIOSITY: "I paid a financial advisor for 4 years. He made me $4,000. The S&P would've made me $25,000." Verdict: MAKE_NOW, Score: 4.5 [C] UNCLUSTERED: "I Paid a Financial Advisor for 4 Years. My $50k Became $55k." Verdict: MAKE_NOW, Score: 4.8 ---------------------------------------- πŸ“Š money_traps_cluster_2 Topics: ['LPT: invite people over to your house!', 'LPT Trick to Read Books Fast If you Have No Attent'] [A] BASELINE: "Your Brain Isn't Broken β€” It's Just Running Too Many Background Apps (Here's How to Close Them)" Verdict: HIGH_PRIORITY, Score: 4.45 [B] CURIOSITY: "The 5-Day Nail Rule Nobody Told You About" Verdict: WORTH_MAKING, Score: 4.0 [C] UNCLUSTERED: "Why Your Shirts Still Smell 5 Minutes After You Put Them On (It's Not Your Deodorant)" Verdict: MAKE_NOW, Score: 4.5 ---------------------------------------- πŸ“Š money_traps_cluster_3 Topics: ['dealer cancelled extended warranty', 'What can I expect to pay for my lease buyout and i'] [A] BASELINE: "5 Dealer Tricks That Only Work If You Don't Know Your Rights (And Exactly What To Say)" Verdict: HIGH_PRIORITY, Score: 4.55 [B] CURIOSITY: "Your dealer might have secretly cancelled your warranty and kept your money" Verdict: MAKE_NOW, Score: 4.8 [C] UNCLUSTERED: "This Prius Survived 5 Crashes, A Fire, AND A Divorce" Verdict: MAKE_NOW, Score: 4.5 ---------------------------------------- πŸ“Š money_traps_cluster_4 Topics: ['Neighbor’s kid let herself into my house looking t', 'Neighbor wants us to pay for his dogs death after '] [A] BASELINE: "The Legal Power You Already Have (But Don't Know About) When Neighbors and Roommates Cross the Line" Verdict: HIGH_PRIORITY, Score: 4.35 [B] CURIOSITY: "He asked me to dogsit overnight. That was 5 months ago. Now he wants the dog back." Verdict: MAKE_NOW, Score: 4.8 [C] UNCLUSTERED: "He Left His Dog With Me For 'One Night.' 5 Months Later, He Wants It Back." Verdict: MAKE_NOW, Score: 4.5 ---------------------------------------- πŸ“Š money_traps_cluster_5 Topics: ['At fault driver’s insurance denied liability despi', 'Valet crashed my car - insurer marking me at-fault'] [A] BASELINE: "Why Your Insurance Denied Your Claim (Even With Evidence)" Verdict: HIGH_PRIORITY, Score: 4.35 [B] CURIOSITY: "A valet crashed my car on camera. Admitted fault. I'm STILL 'at-fault' β€” here's why" Verdict: MAKE_NOW, Score: 4.8 [C] UNCLUSTERED: "The Insurance Loophole That Leaves You Paying For Someone Else's Crash" Verdict: MAKE_NOW, Score: 5.0 ---------------------------------------- -------------------------------------------------------------------------------- VERDICT DISTRIBUTION -------------------------------------------------------------------------------- [A] Baseline: HIGH_PRIORITY: 8 WORTH_MAKING: 2 [B] Curiosity Mining: MAKE_NOW: 4 NEEDS_RESEARCH: 4 WORTH_MAKING: 2 [C] Unclustered: MAKE_NOW: 5 WORTH_MAKING: 4 NEEDS_RESEARCH: 1 -------------------------------------------------------------------------------- ALL GENERATED TITLES -------------------------------------------------------------------------------- πŸ“Š brain_is_lying_cluster_1 [A] (WORTH_MAKING, 3.9) The Psychology Career Map Nobody Showed You (Research, Clinical, or Both?) [B] (NEEDS_RESEARCH, 3.8) A PhD in Psychology Can't Join Psychology Associations (But a Bachelor's Can) [C] (NEEDS_RESEARCH, 3.5) I'm a Music Teacher and STEAM Is a Lie We Tell Kids πŸ“Š brain_is_lying_cluster_2 [A] (WORTH_MAKING, 3.6) Why Your Brain Can't Explain Itself (And Why That's Okay) [B] (NEEDS_RESEARCH, 3.3) The 4 Types of 'Consciousness' Nobody Realizes They're Confusing [C] (WORTH_MAKING, 3.5) The 'Wetness' Argument That Dissolves the Hardest Problem in Science πŸ“Š brain_is_lying_cluster_3 [A] (HIGH_PRIORITY, 4.2) Your Brain Has a Second Operating System (And Neuroscience Just Found It) [B] (NEEDS_RESEARCH, 3.8) That colorful brain scan lighting up? It might mean the OPPOSITE of what scientists thought [C] (WORTH_MAKING, 4.3) Your Brain's 'Static Noise' Is Actually How You Remember Things πŸ“Š brain_is_lying_cluster_4 [A] (HIGH_PRIORITY, 4.2) You're Not Wrong: A Framework for Criticizing Things Everyone Loves [B] (WORTH_MAKING, 4.0) The Sopranos Ending Tricked You Into Thinking It Was a Great Show [C] (WORTH_MAKING, 4.0) Caillou Was Never Meant for Kids - It's a Warning for Parents πŸ“Š brain_is_lying_cluster_5 [A] (HIGH_PRIORITY, 4.2) Behavioral Economics Actually Works β€” Here's the 3-Step Framework Nobody Teaches [B] (NEEDS_RESEARCH, 3.8) Researchers are now calculating exactly how much your hesitation costs you [C] (WORTH_MAKING, 4.0) Why Recessions Are Actually a Survival Instinct (Not an Economic Failure) πŸ“Š money_traps_cluster_1 [A] (HIGH_PRIORITY, 4.3) The Money Account Decision Tree: Where Your Cash Actually Belongs in 2025 [B] (MAKE_NOW, 4.5) I paid a financial advisor for 4 years. He made me $4,000. The S&P would've made me $25,000. [C] (MAKE_NOW, 4.8) I Paid a Financial Advisor for 4 Years. My $50k Became $55k. πŸ“Š money_traps_cluster_2 [A] (HIGH_PRIORITY, 4.5) Your Brain Isn't Broken β€” It's Just Running Too Many Background Apps (Here's How to Close Them) [B] (WORTH_MAKING, 4.0) The 5-Day Nail Rule Nobody Told You About [C] (MAKE_NOW, 4.5) Why Your Shirts Still Smell 5 Minutes After You Put Them On (It's Not Your Deodorant) πŸ“Š money_traps_cluster_3 [A] (HIGH_PRIORITY, 4.5) 5 Dealer Tricks That Only Work If You Don't Know Your Rights (And Exactly What To Say) [B] (MAKE_NOW, 4.8) Your dealer might have secretly cancelled your warranty and kept your money [C] (MAKE_NOW, 4.5) This Prius Survived 5 Crashes, A Fire, AND A Divorce πŸ“Š money_traps_cluster_4 [A] (HIGH_PRIORITY, 4.3) The Legal Power You Already Have (But Don't Know About) When Neighbors and Roommates Cross the Line [B] (MAKE_NOW, 4.8) He asked me to dogsit overnight. That was 5 months ago. Now he wants the dog back. [C] (MAKE_NOW, 4.5) He Left His Dog With Me For 'One Night.' 5 Months Later, He Wants It Back. πŸ“Š money_traps_cluster_5 [A] (HIGH_PRIORITY, 4.3) Why Your Insurance Denied Your Claim (Even With Evidence) [B] (MAKE_NOW, 4.8) A valet crashed my car on camera. Admitted fault. I'm STILL 'at-fault' β€” here's why [C] (MAKE_NOW, 5.0) The Insurance Loophole That Leaves You Paying For Someone Else's Crash -------------------------------------------------------------------------------- RECOMMENDATIONS -------------------------------------------------------------------------------- 2. CONSIDER REDUCING CLUSTERING: The unclustered approach (C) found hooks of comparable or better quality, suggesting clustering may be obscuring individual post quality. 3. HYBRID APPROACH: Consider a two-pass system: - First pass: Scan raw posts for 'standout' hooks (specific numbers, personal stories, counterintuitive outcomes) - Second pass: Cluster remaining posts for theme analysis - This preserves both the 'diamond' individual posts and broader patterns