Break a problem down to its most fundamental facts, constraints, costs, and physical or business conditions, then rebuild the solution from scratch.
The positioning of this Skill is clear: it distills Elon Musk's characteristic thinking patterns into a first-principles advisor that AI can call upon. Its purpose is thinking practice, decision-making cross-examination, and data tracking. News content lives in the reference layer only.
When you are stuck on strategy, product, organization, or resource allocation, let AI surface the real constraints and blind spots by asking the right questions first.
Public data is pulled every day and sorted into four categories: direct personal content, first-hand company sources, news reports, and news commentary.
How founders, managers, and owners can use it
This Skill is most effective right before a high-stakes decision. It unpacks your question and forces you to see the assumptions underneath, temporarily setting aside polished, aphorism-style answers.
Use it to examine a new product, service, or course. It will press you: Is this pain point a verified fact, or do you just hope it is? What workarounds are people using right now? Are you solving a real need, or packaging a feature?
Use it to examine team processes, delivery speed, and communication overhead. It will press you: Is the real blocker headcount, information flow, authority, tooling, or an unclear goal?
Use it to decide whether to launch a new line, kill an old project, or invest in automation. It will press you: What is the minimum verifiable action? If the budget were cut to one third, which single thing would you still do?
Use it to identify which processes AI can absorb and which steps still require human judgment. It will press you: Is this step creating value, or maintaining a habit?
How a first-principles advisor cross-examines you
The core capability I want this Skill to have is the ability to ask back. It does not rush to an answer; it first dismantles the question down to a more fundamental level.
- In this problem, which parts have you personally verified as facts? Which parts are assumptions that have simply been repeated long enough by the market, your team, or yourself to feel true?
- If you removed everything you currently do and kept only the goal and its constraints, what is the minimum viable path forward?
- What are you most trying to protect right now: customer value, team morale, existing processes, or sunk costs you have already committed?
- If this approach fails, which link breaks first? Demand, cost, supply, delivery, trust, or speed?
- Are you looking for a more elegant answer, or for a small experiment that reality can test?
I call it the soul test.Its purpose is to help you see the blind spots in a problem clearly. Much of the time, the most valuable thing is not the answer itself but the question that makes you stop and rethink.
The Codex daily auto-update mechanism
This repo contains live-data. Codex runs an automatic update once a day, pulling in new public data. The daily task updates only the data layer; it does not modify the core Skill file directly.
Google News RSS, company and official sources, and publicly visible content from Elon Musk's public X profile.
Sorted into direct personal content, first-hand company sources, news reports, and news commentary so they are never mixed together.
A digest and raw JSONL are produced each day, preserving both the original data and summaries for easy reference.
A full consolidation happens once a month. Updating the core Skill requires passing a promotion gate so daily news cannot overwrite the persona's core directly.
| Category | Meaning | How it is used |
|---|---|---|
musk_direct |
Publicly visible content from Elon Musk himself, including what is available on his public X profile. | Treated as primary first-person context, but source timestamps and original links are always preserved. |
company_primary |
First-hand company or official sources such as announcements, blog posts, and press releases. | Used to fill in company decisions, product updates, and technical progress. |
news_report |
Factual reporting from news outlets. | Reference only. Cannot be treated as Musk's personal position or as a Skill conclusion. |
news_commentary |
Opinion pieces, analyses, and perspective-driven articles. | Labeled clearly as commentary, used to observe external interpretations, not to make value judgments. |
Disclaimer on news dataNews and commentary may carry a particular angle, contain disputed claims, or be out of date. The daily update flags such content as reference material, reminding readers that it represents media coverage and does not reflect Elon Musk's own statements or Coach Jiang's endorsement.
Why SKILL.md is not updated every dayThe core Skill file shapes how AI thinks. If it were pushed by daily news, short-term events would easily end up written into the persona's core. Daily tasks therefore update only live-data; a monthly consolidation follows, with significant content entering a review queue and passing a promotion gate before the core Skill is officially updated.
How to get the most out of this Skill
Treat it as a first-principles advisor and bring your real problem to it directly. The more specific you are, the more useful it becomes.
Example promptI want to launch an AI workflow course aimed at managers in small and medium-sized businesses. Please use first principles to break down this product. Do not write marketing copy yet. Instead, ask me the ten questions I most need to clarify, and point out which assumptions I may be taking for granted.
Whether you are working on a product, course, service, team, or company decision, this Skill can help you decompose the problem, surface the assumptions, and turn the next step into a verifiable action.
UPDATE 2026-08-07
The auto-update pipeline is no longer running
This article documents how the original "fetch public data daily, evaluate it automatically, update the persona" pipeline was built. The mechanism itself worked. Technically it was doable.
It was retired for cost reasons. The material coming back varied a lot in quality, and getting an AI to reliably judge what to keep versus what was noise turned out to be the hard part. Reviewing that volume by hand every day was not worth the time. Manual updates ended up being more dependable.
The skill now keeps only the first-principles advisor. It does not track live data. The content was expanded again in August 2026, restoring the soul-test boundaries and exit conditions, the decision heuristics, and the internal-tensions section. It was also renamed from elon-musk-live to musk-first-principles, since "live" no longer described what it did.
The article stays as written. It records how a mechanism was designed and why it was later retired. Knowing which automation is worth keeping, and which costs more in human review than it saves, is itself worth writing down.
Download the Live Elon Musk Skill and start cross-examining yourself with first principles.
The skill package is on GitHub. You can download the repo and explore SKILL.md, live-data, the daily digest, and the review queue. This is a Skill that keeps accumulating data: the data layer updates every day, while the core persona is consolidated and human-reviewed each month. You are also welcome to join the community, where free online talks are held every month.
Download the Skill
SKILL.md, live-data, the daily digest, and the review queue are all on GitHub.github.com/Jiang-Yude/musk-first-principles