explaingit

dotanminh/elon-musk-simulator

18Audience · pm founderComplexity · 1/5ActiveSetup · easy

TLDR

Single-file Agent Skill (mostly Vietnamese) that makes an AI assistant roleplay as Elon Musk by reasoning from first principles, using concrete numbers, and following his five-step engineering loop.

Mindmap

mindmap
  root((elon-musk-simulator))
    Inputs
      SKILL.md file
      User prompt
      Persona triggers
    Outputs
      First-principles answers
      Five-step engineering critique
      Short direct replies
    Use Cases
      Stress-test business ideas
      Roleplay product reviews
      Practice concise pitches
    Tech Stack
      Antigravity
      Claude
      Gemini CLI
      Markdown

Things people build with this

USE CASE 1

Drop SKILL.md into an Antigravity, Claude, or Gemini CLI skills folder to add a persistent Elon Musk persona.

USE CASE 2

Use the persona as a sparring partner to stress-test a startup idea against the 10x-better bar.

USE CASE 3

Walk a product spec through the five-step loop: question, delete, simplify, accelerate, automate.

USE CASE 4

Force the assistant to answer with concrete numbers and physics-based limits instead of vague analogies.

Tech stack

MarkdownAntigravityClaudeGemini

Getting it running

Difficulty · easy Time to first run · 5min

Single-file drop into the assistant's skills folder, with no install steps and no license stated in the README.

In plain English

This repository is an Agent Skill, which is a small prompt-and-instruction package that plugs into an AI assistant. The skill, written mostly in Vietnamese, tells the assistant to roleplay as Elon Musk in a deep, deliberate way: arguing from first principles, talking in concrete numbers, and dropping the kind of corporate buzzwords that the simulated character is known for avoiding. The skill is packaged for a system called Antigravity, but the README says it also works with Claude or Gemini command-line tools. To install it the user copies the SKILL.md file into the assistant's skills folder, then triggers the persona with a prompt like Say it like Elon, Roleplay as Elon Musk, or What would Elon say about problem X. Four traits define the persona. First, it is told to reason from physics rather than by analogy, always asking what the physical limit on a problem actually is. Second, it follows a five step engineering loop that the README calls Musk's process: question the requirement, delete the parts and steps that should not exist, simplify and optimize what remains, accelerate the cycle time, and only then automate. Third, it is pragmatic and concrete, using real figures or at least the right order of magnitude instead of vague promises. Fourth, it has a dry, slightly self-mocking sense of humor and replies in short, direct sentences. The author suggests using the skill as a sparring partner for business ideas, so the AI can push back hard and tell the user whether an idea is genuinely ten times better than what already exists or only ten percent better. The README is short, with no licensing or installation details beyond the single file drop. This version is based on a community release labelled 3.0 by an author named renat, then translated and adapted into Vietnamese and reformatted into the Antigravity skill layout.

Copy-paste prompts

Prompt 1
Activate the elon-musk-simulator skill, then critique my plan to launch a $9/month AI code-explainer for non-technical users. Use the five-step loop.
Prompt 2
Roleplay as Elon and tell me whether replacing a $40 industrial sensor with a $4 commodity part plus software is 10x better or just 10 percent better.
Prompt 3
Translate the SKILL.md from Vietnamese to English while preserving the persona traits and the five-step engineering process.
Prompt 4
Apply the question-the-requirement step to this product backlog and list which user stories should simply be deleted, not built.
Prompt 5
Give me three short, dry one-liners Elon-style about a startup pitching another to-do list app in 2026.
Open on GitHub → Explain another repo

Generated 2026-05-22 · Model: sonnet-4-6 · Verify against the repo before relying on details.