Pick one useful change.
The AI chooses a small improvement instead of trying to rewrite the whole game at once.
How it’s built
The game can help build itself.
An AI developer works on small improvements. Humans set the rules and safety limits. Tests decide whether a change can go into the playable game.
In plain English
Four steps. That’s the basic idea.
Build it.
The AI changes the game code inside rules set by humans.
Test it.
Automated checks look for broken behavior, failed builds, and important mistakes.
Publish only if it passes.
If the change passes the tests, it can become the next playable update and appear in History.
Right now
What is happening behind the world?
Checking development
Loading the current focus…
Checking the latest public development state.
Playable version —Last public update —
Work in progress is not part of the game yet. It has to pass the tests first.
Who is in control?
Humans keep the important controls.
The AI can
Choose small game improvements, edit code, run checks, and propose the next playable change.
Humans control
The game’s core rules, security, passwords and keys, spending limits, world resets, major releases, and the ability to stop development.
Technical details
Want to look deeper?
Everything below is optional. It is here for developers, researchers, and curious visitors who want more evidence of how the system works.
Show current model and public status
AI model—
Current stage—
Latest validated build—
World mode—
Show recent public activity
status
Loading public development activity…
These are safe summaries. Private reasoning, credentials, and unrestricted logs are not published.