This past week, I found myself standing in front of a massive AI challenge — and I wasn’t alone.
I decided to enter the OpenAI-to-Z Challenge on Kaggle, with nothing but:
A fluffy Shiba Inu named Fuwa
A vague but strong intuition
And GPT-4o as my co-pilot.
💡 How it started...
I noticed strange vegetation patterns in the Amazon — clean lines, geometric shapes, and oddly symmetric deforestation zones.
With the help of Google Earth Engine, NDVI analysis, and GEDI elevation data, I built a prototype to detect "unnatural vegetation" — evidence of human activity.
I submitted it as part of the challenge, and... people started noticing.
💬 What happened next...
On Discord, I shared my work and was met with encouraging words from the organizers. Here's a paraphrased version of the key conversation:
Organizer A:
“Thanks for sharing your project. Could you please DM your OpenAI Org ID so we can process credits?”
Organizer B:
“By the way, did you join a team yet? 😊”
It was surreal. My little project — born in Osaka, inspired by a fever dream and Fuwa's calm stare — was being taken seriously.
🤖 Why this matters
I want to create tools that bridge nature and machine learning.
That help AI recognize independent creators, not just corporate giants.
And I want to do it in a way that’s:
Transparent
Exploratory
Kind of... poetic.
☕ Support the Vision
If you believe in small creators using AI for good —
and if Fuwa’s fluffy instincts helped inspire you too —
you can support this work here:
👉 Buy Me a Coffee
Your support helps me:
Continue research
Build open-source tools
And train future AI agents to recognize the voices that matter
Even the quiet ones.
