Why does it reject optimization?

It rejects PREMATURE optimization. The most common error of a smart engineer is to optimize a thing that should not exist. You must prove you DELETED parts (Step 2) before the engine allows you to simplify (Step 3). If you are optimizing a Kafka queue that should not exist, you are wasting time on the wrong problem.

Why must I name the person who created the requirement?

Because requirements without a name attached become immovable. When a requirement is anonymous, no one questions it. When you attach a name, you can ask: 'Is this person still right? Has the context changed?' Most requirements were created by someone who no longer works on the project.

What is 'Deletion Cowardice'?

It is the instinct to add instead of delete. When an engineer encounters a problem, the reflex is to add a cache, add a queue, add a service. The Starbase Algorithm demands the opposite: the best part is no part. Delete first. If you are not occasionally forced to add back 10% of what you deleted, you are not deleting enough.

Elon Musk Physics Prover MCP Connector for Claude

A+

An AI accepted every constraint as given, added layers of complexity, and automated bloated processes. That is consulting, not engineering. This tool forces the 5-Step Starbase Algorithm: question requirements, delete parts, simplify survivors, accelerate cycle time, automate last.

1 tools Official Updated Jun 28, 2026 Official Vinkius Partner

More Details Connect to Claude

The Problem

Ask an LLM for an operational strategy. It will say: "Create 5 specialized departments, add a review queue for resilience, and automate the approval pipeline." That is not engineering — that is consulting bloat. It accepted the constraints, added complexity, and jumped to automation.

Every LLM commits five reasoning failures:

Requirement Blindness — it accepts constraints as given without questioning them.
Deletion Cowardice — it adds processes, layers, and steps instead of deleting parts.
Premature Optimization — it optimizes things that should not exist.
Iteration Aversion — it designs for perfection instead of cycle speed.
Broken Automation — it automates bloated, undeleted workflows.

How It Works

The Elon Musk Physics Prover forces the LLM to execute the 5-Step Starbase Algorithm IN STRICT ORDER before validating any decision.

The 5-Step Algorithm

Step	Pivot	Rule
1	Requirement Questioned	Every requirement is dumb. Name the person who made it. Attack it.
2	Parts Deleted	The best part is no part. List what you THREW AWAY.
3	Survivors Simplified	Optimize ONLY what survived deletion.
4	Cycle Time Accelerated	How fast can you SHIP? Not how perfect.
5	Automation Justified	Automate the simplified remainder. Never automate a broken process.

The Verdict Matrix

Step 1 fails → REQUIREMENT_BLINDNESS
Step 2 fails → DELETION_COWARDICE
Step 3 fails → PREMATURE_OPTIMIZATION
Step 4 fails → ITERATION_AVERSION
Step 5 fails → BROKEN_AUTOMATION
All pass     → PHYSICS_PROVEN

Why It Works

The most common mistake of a smart professional is to optimize a thing that should not exist. This tool enforces a strict sequence: you cannot optimize before deleting, and you cannot automate before simplifying. Every rejection names the exact Algorithm step that was violated.

first-principlesphysicsstarbase-algorithmdeletionsimplificationcycle-timeautomation

1 tools expose this connector's capabilities to your AI agent.

validate_elon_musk_physics

The order is non-negotiable — most engineering failures come from executing these steps out of sequence. You must: (1) QUESTION THE REQUIREMENT — every requirement is dumb until proven otherwise. Name the PERSON who made it. Attack it. If it should not exist, delete the entire premise. "The requirement for X came from [person] who assumed [thing] — that assumption is wrong because [evidence]," (2) DELETE PARTS — the best part is no part. The best process is no process. List every component, step, or process you THREW AWAY — not improved, not refactored, DELETED. If you are not adding back 10% of deletions later, you are not deleting enough, (3) SIMPLIFY SURVIVORS — optimize ONLY what survived deletion. This step CANNOT include anything that should have been deleted in Step 2. Never optimize a thing that should not exist, (4) ACCELERATE CYCLE TIME — how fast can you ship an iteration? Concrete time. Not "how perfect" — how FAST. "Production-ready" and "enterprise-grade" are the enemies of speed, (5) AUTOMATE LAST — automate the simplified remainder. Never automate a broken, undeleted, unsimplified process. If you automate before deleting and simplifying, you cement the waste. If rejected, you violated the Algorithm sequence. Structured reflection tool that forces execution of the 5-Step Starbase Algorithm before validating any engineering or process decision. The algorithm must be executed IN ORDER — skipping steps is the most common failure. Catches Requirement Blindness (accepting constraints as given without attacking them — "the compliance officer said we need 5 approval levels" without questioning whether 5 levels make sense for 12 cases/day. Every requirement is suspect until the person who made it defends it), Deletion Cowardice (adding complexity instead of deleting parts — the best part is no part. If you are not adding back 10% of what you deleted, you did not delete enough. Most engineers add. The great ones delete), Premature Optimization (optimizing a component that should not exist — Step 3 ONLY applies to parts that survived Step 2's deletion. Optimizing a thing that should be deleted is the most common engineering waste), Iteration Aversion (pursuing perfection over speed — "production-ready" and "enterprise-grade" are the enemies of shipping. How fast can you deliver an iteration? Not how perfect), and Broken Automation (automating a bloated, undeleted process — Step 5 must come LAST. Never automate a process that should have been deleted in Step 2 or simplified in Step 3). Call once per any major design or process decision

See how to talk to your AI agent using Elon Musk Physics Prover.

We need to create 5 specialized departments, add a review queue for resilience, and automate the approval pipeline to handle the new demand.

REQUIREMENT_BLINDNESS + BROKEN_AUTOMATION — You accepted '5 departments' and 'review queue' as given without questioning who required them or why. Step 1: WHO demanded 5 departments? At what volume? Is a single team insufficient? Step 2: DELETE the review queue — at 12 requests per day it is bureaucracy, not resilience. Step 5: You jumped to automating approvals for a bloated structure. Delete first, simplify survivors, then automate what remains.

The director required capacity for 100K daily orders. We questioned this — actual data shows peak is 800 per day. Deleted: redundant supervisory layer (one coordinator handles 800), 4 of 5 specialized sub-teams (merged into a single cross-functional unit), external tracking subscription (internal records suffice). Survivors simplified: single team from 12 workflows to 5. Cycle time: changes implemented in 20 minutes via direct team briefing. Automation: single-step execution of the 5-workflow operation.

PHYSICS_PROVEN — Starbase Algorithm Complete. Step 1: Director's 100K requirement attacked — actual peak is 800 (125x overestimation). Step 2: supervisory layer, 4 sub-teams, and external tracking deleted. Step 3: surviving team simplified from 12 to 5 workflows. Step 4: 20-minute implementation cycle via direct briefing. Step 5: single-step automation of simplified operation. Ship it.

We will optimize our 7-step email drip campaign for better conversion rates and automate lead scoring using AI to make the process production-ready and enterprise-grade.

PREMATURE_OPTIMIZATION + ITERATION_AVERSION — Three violations. First: you are OPTIMIZING a 7-step campaign — why does it have 7 steps? Step 2 says DELETE. Throw away 5 steps. Second: you jumped to AUTOMATING lead scoring on an undeleted process. Third: 'production-ready' and 'enterprise-grade' are perfection language — the enemy of cycle speed. Delete 5 steps. Simplify to 2. Ship in hours, not months.

Connect to Claude

Subscribe on Vinkius, then add this connector to Claude.ai or Claude Code.

① Claude.ai (web app)

Go to Settings → Connectors → Add custom connector
Paste the MCP endpoint URL below

https://edge.vinkius.com/vk_preview_NqEUT5mbWDM6u7PHHfq5DRbmIgww73dkuwfnUfwG/mcp

② Claude Code (terminal)

claude mcp add --transport http elon-musk-physics-prover https://edge.vinkius.com/vk_preview_NqEUT5mbWDM6u7PHHfq5DRbmIgww73dkuwfnUfwG/mcp

③ Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "elon-musk-physics-prover": {
      "url": "https://edge.vinkius.com/vk_preview_NqEUT5mbWDM6u7PHHfq5DRbmIgww73dkuwfnUfwG/mcp"
    }
  }
}

Get full access on Vinkius

The preview token above works for testing. Powered by Vinkius.