question-tournament.summary

Core Concept

Question quality ek skill hai jise aap evaluate aur improve kar sakte hain, innate talent nahin. Tournament format (apne sawalat ke liye partner ke sawalat se compare karna aur measure karna ke kaun se sawalat waqai useful AI responses paida karte hain) good aur great questions ka farq viscerally clear kar deta hai.

Key Mental Models

• Evaluation Through Comparison: Question quality ke liye isolation mein judge karna unreliable hai; AI responses ke side-by-side comparison se pata chalta hai ke kaun se sawalat waqai kaam karte hain
• Actionable vs. Generic: Question ki real quality uske output se measure hoti hai; kya AI response problem narrow karta hai (actionable) ya wohi repeat karta hai jo koi bhi prompt produce karta (generic)?

Critical Patterns

• Kisi bhi comparison se pehle AI ke baghair independently 15 questions generate karein
• Partner ke sawalat ke liye har ek ke liye one-sentence justification ke saath rank karein; justification precise thinking force karti hai ke ek question doosre se behtar kyun hai
• Comparison table banayein jo har AI response use karne ke liyeful/actionable ya generic/filler mark kare, explanation ke saath
• Har set ke top 5 questions ke liye Claude aur ChatGPT dono par compare karein

Common Mistakes

• Yeh assume karna ke zyada questions ka matlab behtar analysis hai: 3 great questions, 15 mediocre questions se behtar perform karte hain
• Sawalat ke liye diagnostic power ke bajaye topic importance se rank karna; important topic par question phir bhi poorly formed ho sakta hai
• Tournament ke liye jeetne wali competition samajhna, seekhne wala mirror nahin

Connections

• Builds on: Prediction Lock (Exercise 1); question ranking aur diagnostic power concepts aage carry hote hain
• Leads tak: Divergence Test (Exercise 3), jahan class-wide comparison reveal karta hai ke students ne AI ke liye thinking partner ke taur par use kiya ya answer machine ke taur par