digital-twin

Appearance

The "Elon Musk" Digital Clone ​

One of the hardest things to grasp about digital-twin is what we mean by "Operating Model." It's not about making an AI talk in a specific tone—it's about making an AI think and execute using a specific set of rules.

To prove this, we built a Digital Clone of Elon Musk.

Why Elon Musk? ​

Elon operates on very explicit, well-documented frameworks:

  • First Principles Thinking: Breaking problems down to physics.
  • The Idiot Index: Calculating raw material cost vs final cost.
  • Extreme Urgency: Deleting processes and moving faster.
  • Direct Communication: Banning corporate jargon and large meetings.

We took these public philosophies and loaded them into a digital-twin environment.

Try it Locally ​

You can find this demo in the examples/elon-musk folder of the repository.

The Setup ​

  1. AGENTS.md: We defined his core persona and intolerance for bureaucracy.
  2. wiki/management.md: We encoded his rules on meetings and the "Idiot Index".
  3. wiki/physics-principles.md: We encoded his 5-step engineering process (1. Make requirements less dumb, 2. Delete the part, etc.).

The Prompt ​

If you run the FIRST_PROMPT.md, you ask the twin to write a memo about a Starship heat shield issue.

A normal AI would write a polite, standard corporate email ("Dear Team, we need to synergize our efforts...").

The Digital Twin retrieves the wiki first. It realizes large meetings are banned, processes must be deleted, and the tone must be direct. It outputs a ruthless, physics-bound memo demanding the elimination of the QA step and a 10x speed increase, perfectly mirroring the intended operating model.