Today I purchased waihong.ai.
Not a blog. Not a portfolio. A workshop — with the door open.
I compiled a full business plan with Claude, mapped out a technical roadmap from Python zero to RAG-as-a-Service, and committed to documenting every step of this journey publicly.
Why public?
Because transparency is the brand. The progress bars on this site are real. Some of them are at 10%. Some are at 5%. That's not embarrassing — that's honest. And I believe honesty builds more trust than a polished facade ever could.
What I have so far
Five days ago, I built a complete local AI stack from zero terminal experience:
- Ollama running 6 models on my MacBook Pro M4
- Open WebUI for a ChatGPT-style interface
- AnythingLLM with RAG capabilities
- VS Code + Continue with qwen2.5-coder for coding assistance
- Tailscale mesh VPN connecting my devices
I uploaded an HR handbook into the RAG system and asked it a question that wasn't in the document. The AI said: "The information is not available in the provided context."
No hallucination. No confident nonsense. That single sentence changed everything for me.
What I don't have yet
- Python skills (0%)
- LangChain knowledge (0%)
- A production-ready RAG pipeline (working on it)
- Client-facing UI (coming)
The plan
Learn Python. Build RAG in code. Serve my existing Scrum HR clients. Document everything here.
The progress bars start now.
March 16: command not found. March 21: waihong.ai purchased. The journey begins.