Show HN: 70 Days → 800 GitHub Stars (Cold Start) – My Secret Was a Problem Map
I want to share a story that might resonate with anyone who has tried to launch an open-source project from zero visibility, zero budget, zero network.
Seventy days ago I created a repo called WFGY. I had no prior followers, no launch partners, no ads. Today, it’s at 800 stars and counting. Pure cold start.
The “secret”? Not marketing. Not tricks. It was building a Problem Map.
---
#### What is a Problem Map?
Instead of claiming my system “improves reasoning” or “solves RAG bugs,” I wrote down 16 concrete failure modes engineers hit every day:
* Embedding drift (semantic ≠ vector closeness) * Chunking collapse (half-sentence vs full chapter granularity mismatch) * Versioning hallucinations (index v1 + v2 = ghost hybrid doc) * Pre-deployment ingestion collapse * Prompt injection bypasses * … and 10 more, all documented.
For each one I built a text-only “semantic firewall” module that fixes the issue without changing infra. Developers could see their exact bug listed, mapped, and patched.
This gave the repo two things:
1. Immediate empathy — engineers recognized their pain points word-for-word. 2. Immediate usability — they could paste my TXT file, keep infra untouched, and the model stabilized.
---
#### Why it mattered
* It wasn’t a claim, it was a diagnostic mirror. * People didn’t need to “believe” me — they could try the TXT in 60 seconds and watch bugs stop replicating. * Instead of selling “AI magic,” I showed a surgical checklist.
This is why strangers starred the repo. It wasn’t hype; it was problem → fix → proof.
---
#### What’s next
I’m now scaling this into something bigger:
* The Problem Map will evolve into a Global Semantic Surgery Room — think of it like an AI hospital’s operating theater. * Engineers bring in failure logs (RAG, OCR, planning drift). * The system triages, diagnoses, and applies the right semantic patch in real time. * Integrations are underway for n8n, Make, GoHighLevel and more — all unified under the same open Problem Map.
Launch target: September 1.
### Closing
I don’t see this as “my repo got stars.” I see it as proof that open source thrives when you solve pain with surgical precision.
If you’ve ever tried to cold start, my suggestion: don’t sell a dream, map a nightmare. Engineers don’t trust hype — they trust a checklist that fixes the crash.
Coming next: the Semantic Surgery Room and Global Fix Map (n8n, GHL, Make and more). Launching by Sep 1.