# Crossroads — Qoder’s AI Coding Journey
**Date:** 2025-10-29 · **Location:** Beijing
---
> **“Absolutely right, unquestionably right.”**
---
## 🎙 Podcast Episode Overview
**Guests:** Koji, Ronghui
**Editing & Organization:** Starry
**Layout Design:** NCon
This episode features an in-depth conversation with **Shutang**, founder of Alibaba **Qoder** — an "Agentic Coding" product — reflecting on 60 days of competition in the global AI Coding **red ocean**.
---
### 📈 The Numbers Behind Shutang’s Confidence
- **100,000 users** in the first 5 days after launch
- **500,000 active developers** in just 60 days
---
## 🔥 The Global AI Coding Landscape
AI coding is arguably the hottest field in 2025:
- **Cursor**: $500M annual revenue, $10B valuation — a prized Silicon Valley unicorn
- **Anthropic’s Claude Code**: growth even steeper than OpenAI
> **Question to consider:**
> Where can Alibaba find space in such a competitive battlefield?
---
## 👤 About Our Guest
Shutang:
- Former technical lead for **Alibaba Double 11**
- Led China’s earliest **cloud-native transformation**
- Now spearheading Qoder’s rapid rise by combining **product thinking**, **technical strategies**, and **organizational capabilities**.
---
🎥 **Recorded at:** AI Hacker House
🎬 **Video available on:** Koji Yang Yuancheng's Xiaohongshu, Bilibili, WeChat Video, YouTube
---
## 🎧 Listen Now
- **WeChat Podcast**
- **Xiaoyuzhou Podcast**
---
## 📚 Interview Outline (Summary)
### 1. Quick Q&A
- Age, alma mater, MBTI, zodiac sign
- One-sentence pitch
- Revenue & profit figures
- Team size
- Pre-Qoder work
---
### 2. **AI Coding Panorama: From 0–1 vs 1–100**
A typical user:
1. Starts with AI generating a prototype site.
2. Switches to specialized tools for long-term maintenance.
**Key Points:**
- Three main categories:
1. Tools for creators to build instantly from ideas.
2. Tools boosting professional developer efficiency.
3. “Digital employees” replacing human labor.
- Trend toward **uniform integration**.
- Without self-developed models, tools risk being “wrappers” for providers.
- Tech giants hold advantage via cost control + model-product optimization.
---
### 3. **Breaking Out of the Red Ocean with Qoder**
> *“While others went for low-hanging fruit, we went straight to high-value peaks.”*
- Skipped flashy “0–1” generation
- Targeted **real-world software development** (95% of dev work)
- Positioned Qoder as **Agentic Programming Platform**
- Advocated paradigm shift: **“I don’t do it — the agent does.”**
---
### 4. **Why Requirements Documentation Matters**
**Spec-Driven Process:**
1. User gives one-line requirement
2. Documentation Agent auto-generates detailed spec
3. After confirmation, Code Agent executes tasks over time
Analogy: *Boss → Product Manager (PRD) → Engineers → Delivery*
---
### 5. **Product Philosophy: From “Flow” to “Control”**
- In AI pair programming, uninterrupted flow is rare
- Qoder emphasizes **user control over agent behavior**
- Automatic **machine model selection** to prevent decision fatigue
- Balances **performance–efficiency–cost triangle** via context engineering
---
### 6. **Organizational Strategy for Speed**
- Built on teams & assets from Alibaba’s Lingma
- Global-first approach:
Use best models now → Develop own models in parallel
- Independent BU structure boosting agility
---
### 7. **Repo Wiki: Taming Legacy Code**
Principle: *“Docs can lie — code is always current”*
- AI analyzes code + full commit history
- Produces a **real-time evolving “living document”**
- Deeply integrated into Git workflows
- Unique moat: customized models, agents, team flows
---
### 8. **Programmers in the AI Era**
Optimistic outlook:
- **Jevons paradox**: More automation → higher demand → possibly more programmers
- Future dev skillset shifts toward:
- Requirements insight
- Holistic design
- Results validation
---
## 🌐 Integrating Development with Content Ecosystems
Platforms like **[AiToEarn官网](https://aitoearn.ai/)** exemplify the “from creation to monetization” workflow:
- AI generation tools
- Cross-platform publishing (*Douyin, Kwai, WeChat, Bilibili, Xiaohongshu, Facebook, IG, LinkedIn, Threads, YouTube, Pinterest, X*)
- Analytics + AI model ranking
- Open-source for efficiency & multi-channel distribution
Docs: [AiToEarn文档](https://docs.aitoearn.ai) | Rankings: [AI模型排名](https://rank.aitoearn.ai)
---
## 📌 Takeaways from Shutang’s Insights
### Qoder Differentiators
- **Focus on real software** maintenance & evolution
- **Spec-driven agent workflows**
- **Machine-selected models**
- Multi-interface: IDE, CLI, plugins
- Repo Wiki for legacy code
- Cloud sandboxes for parallel agent execution
### Strategic Parallels
Both Qoder and AiToEarn integrate **creation**, **management**, and **distribution** into cohesive systems.
---
---
## 📎 Reference:
Qoder official site — [https://qoder.com/](https://qoder.com/)
---
