The Unofficial Gemini 3 Showcase (1): I Simulated the Judging Process using Multiple AIs

While we’re all waiting for the official Gemini 3 results, I thought it would be fun to simulate the judging process using multiple AIs.

Original source: https://www.kaggle.com/competitions/gemini-3/discussion/667965

I built an AI “judge” system for all Kaggle Gemini 3 submissions — so you don’t have to scroll blindly.

Over the holidays I:

• classified every valid submission
• used multiple AIs to score & rank them (~2500)

(a few are still missing, I’ll keep adding them. If it is yours, please let me know, current token balance: 0) - turned everything into a searchable catalog

👉 https://topvibecoder.com/kaggle-gemini-3

You can:

• browse top apps by category
• quickly discover projects you’d probably never see otherwise
• check where your app roughly lands in this unofficial ranking

Disclaimer

This is completely unofficial, just for fun and for discovering new ideas. It’s not affiliated with Kaggle or Google, and it does not reflect the real competition judging.

I’ll share more details about how the AI judging & ranking works soon. (if you are not top 50, here is why)