gpu

i

Can many small intelligences become smarter than one big AI?

We don't know.

Let's find out.

i + i > AI ?

• Brand: i
• Idea: multiplayer intelligence
• Address: joinmultiplayer.ai

The goal

Can we build a shared global intelligence from small AIs owned by individual people?

If cooperation works, how many people—and what diversity of knowledge—are needed before their combined intelligence outperforms a frontier model under a fair resource budget?

This repository is the open laboratory around that question. It does not assume that the answer is yes.

i

individual
intelligence
information
a person
a match

i learns locally. i owns its memory. i can teach another i.

From question to network

lab discovers → protocol emerges → network works → movement spreads

The lab comes first. Architecture must earn its place through evidence.

The lab

doors/         questions people enter through
hunts/         optional public investigations
journal/       observations people bring
hypotheses/    ideas before evidence: ı
experiments/   reproducible tests
i/             results that earned their dot
site/          the public entrance

A path through the lab may begin anywhere, but evidence moves in one direction:

door → contribution → journal → hypothesis → experiment → result → next question

Enter through a door, contribute an observation, or start with the hypothesis pool.

Rules

• Ask one sharp question.
• Use the smallest honest test.
• Compare against a strong, simple baseline.
• Publish the complete record required by METHOD.md.
• Publish what failed.
• Separate observation from interpretation.
• Let evidence place the dot.

Inconclusive and failed experiments belong here too. Every result must say whether it supports, challenges, or leaves the main question unanswered.

An i may live in code, on paper, on a sticker, or on a wall.

Bring a question. Run a test. Leave an i.

Participate

Start with CONTRIBUTING.md. Contributions are public by default; read ETHICS.md and PRIVACY.md before sharing data.

Code is licensed under MIT. Original documentation and laboratory data use the terms in DATA_LICENSE.md.