Challenge Summary
For this week’s creative sprint, you’re invited to become an AI-powered archivist — building a digital field guide for anything you care about.
Your challenge is to find or invent a fixed dataset — whether it’s every pasta shape, every Studio Ghibli film, or all the species on your imaginary planet — and turn it into something others can explore. This could be a Pokedex-style site, a searchable API, a minimalist visual encyclopedia, or a strange little tool for diving into weird data.
The core of the challenge is twofold: the corpus you choose, and the way you bring it to life. You can scrape data, generate it, remix public datasets, or compile your own from scratch. Then, use whatever tools you want — design, code, AI, no-code — to make the dataset feel alive.
Your index can be useful, beautiful, nostalgic, absurd, or all of the above. What matters is that you treat the data with care, and give people a reason to get lost in it.
Inspiration
There’s no one way to build an index. Yours can be functional, nostalgic, absurd, or all of the above. Below are some ideas to spark your thinking — feel free to remix or ignore them entirely.
You might create…
- A minimalist Pokedex-style site cataloging extinct animals, complete with AI-generated visuals
- A searchable API of every Studio Ghibli film, with filters for mood, snacks, and spirit animal companions
- A fictional explorer’s guide to the 88 moons of Jupiter, complete with climate zones and local myths
- A cursed terminal interface that lets users browse every late-night infomercial product from the ‘90s
- A field guide to made-up fungi, with Latin names, AI art, and a toxicity index
- A zine of classic pasta shapes, each one paired with a generative haiku and a historical footnote
- A museum-style digital wall that shows every NPC in a video game, with lore, relationships, and alignment
- An open-source index of AI agents built by the community, with prompts, use cases, and live demos
To build your AI-Dex, you can use:
- GPT-4, Claude, or Perplexity for content generation
- DALL·E, Midjourney, or Playground for visuals
- Supabase, Vercel, or Airtable for data storage and hosting
- LangChain, LlamaIndex, or Dust for retrieval-based interfaces
- Or just plain HTML, a spreadsheet, and a good idea
Use whatever tools help bring your dataset to life. Combine them however you want. The more it feels like a love letter to the data, the better.
