Leverage Your AI Unfair Advantages | Generated by AI
Question: Given my specific background and situation, what’s the most direct path to becoming financially wealthy in the next few years?
Answer:
You’re already positioned better than 99% of people. Let me be direct and specific to your actual situation.
Your Unfair Advantages (Don’t Waste Them)
- You train models on MI300X. Most “AI engineers” have never touched a GPU cluster.
- You consume 1B tokens/month and actually analyze what comes back.
- You build CLI agents and automation — you ship, not just talk.
- You have 12 years of full-stack depth under the AI layer.
This is rare. The question is how to monetize it at leverage.
The Three Paths, Ranked for Your Case
1. Get Paid More for What You Already Do (Fastest, 6–18 months)
You’re contract at a global bank doing AI engineering. That’s good but likely underpriced.
The move: Target US/EU remote AI engineering roles directly — not through agencies.
- FAANG AI teams, hedge funds (Jane Street, Two Sigma, Citadel all hire AI infra), AI labs (Mistral, Cohere, Together.ai), AI-native startups with Series B+.
- Your actual comp ceiling: $200K–$400K USD/year remote from China if you land the right role.
- Your current bottleneck is probably visibility, not skill.
What to do:
- Write 3–5 technical blog posts in English that demonstrate MI300X training, nanoGPT internals, or MoE implementation. Post on HN, Twitter/X. One viral post can change your job trajectory.
- Put measurable numbers in your resume: “Trained 760M param GPT-2 from scratch on AMD MI300X, 192GB HBM3. Achieved X tokens/sec throughput.”
- Cold apply directly to job pages. Bypass LinkedIn. Email the hiring manager directly.
2. Build a Product That Compounds (Medium-term, 18–36 months)
You built Fun Live to 30K users and 3M CNY. You know how to ship. The difference now is AI changes unit economics — one person can build what took a team of 10.
High-signal product bets given your skills:
| Product | Why it fits you | Revenue model |
|---|---|---|
| AI coding agent for a niche (banking, fintech compliance) | You know banking + AI + systems | B2B SaaS, $500–$5K/mo per seat |
CLI tooling for ML engineers (you already build ww, zz) |
You are the target user | Open-core / sponsorship / hosted |
| Fine-tuning / eval pipeline as a service | You train models, most can’t | API pricing |
| AI notes / knowledge base tool (you have 9,600 notes) | You live this problem | Subscription |
The banking/fintech AI niche is especially strong — compliance automation, document analysis, audit trail generation. You have domain knowledge most AI engineers don’t.
3. Build Audience → Monetize Knowledge (Parallel, starts now)
You already have a public knowledge base with 9,600 notes and 400 blog posts. That’s a distribution asset you’re not monetizing.
What Karpathy does: Explains deeply, builds in public, attracts both companies and individuals.
- Start a technical newsletter or YouTube channel focused on “training LLMs from scratch” — your MI300X work is genuinely interesting content.
- Sell a course: “Train GPT-2 from scratch on AMD GPUs” — no one else is doing this from the AMD angle. $199–$499 course, 500 sales = $100K–$250K.
- Consulting: $200–$500/hr for AI architecture reviews. Your token consumption and model training experience justifies this.
The Compounding Formula
(Higher salary) × (equity or savings rate) + (product revenue) + (audience flywheel)
= financial independence in 3–5 years
The mistake most engineers make: they optimize only one of these. You want all three running in parallel, with the job funding the product bets.
What Would Accelerate You Most Right Now
- Write one deep technical post per week — publish on your blog + cross-post to HN and X. This is your cheapest customer acquisition and talent signal.
- Apply to 5 high-comp remote AI roles this month — your skills warrant 2–4x your current comp if you’re in the right market.
- Pick one product idea and charge for it — even $50/mo from 20 users validates the direction. You already know how to ship.
- Save aggressively and invest in index funds — don’t let savings sit in cash. Your edge is earning power, not stock picking.
What NOT to Do
- Don’t do crypto speculation as a wealth strategy — you have a real edge in AI, use it.
- Don’t spend 2 years building a product in stealth before charging anyone.
- Don’t wait until you “feel ready” for senior AI roles — you’re ready now.
- Don’t optimize for Chinese domestic market comp — the USD/CNY arbitrage is your friend while it lasts.
The honest answer: your skills are worth significantly more than you’re likely being paid right now. The fastest path to wealth is closing that gap while building one product that can outlast any single employer.