# How to Make Claude Code Your AI Engineering Team

- Channel: [Y Combinator](https://www.youtube.com/channel/UCcefcZRL2oaA_uBNeo5UOWg)
- Published: 2026-04-23
- Source video: [Watch on YouTube](https://youtube.com/watch?v=wkv2ifxPpF8)
- Read on ReadTube: https://www.read.tube/p/videos/wkv2ifxPpF8
- Language: Original (source language)

## Summary

**How to Build Software with AI Agents**

Gary Tan, CEO of Y Combinator, demonstrates how his GStack tool transforms Claude Code into an AI engineering team by implementing structured development processes like office hours, planning, and adversarial review. By simulating the rigors of YC-style startup mentorship and enabling parallel automated workflows, GStack allows individuals to rapidly build and maintain complex software projects.

Gary Tan, president of Y Combinator, introduces GStack, an open-source framework that transforms Claude Code into a comprehensive AI-powered engineering team. The core argument is that effective AI software development requires more than raw model intelligence; it requires a structured team-like process involving planning, adversarial review, and specialized skill sets. By mirroring the methodology used during YC office hours, GStack forces models to evaluate product feasibility, business viability, and design choices before writing code.

Key components of the GStack workflow include:
- Office Hours: Uses a refined questioning process to validate product ideas and refine business strategy.
- Adverarial Review: Automatically probes design documents for logic gaps, privacy risks, and technical flaws.
- Design Shotgun: Rapidly generates and compares multiple UI/UX directions for immediate feedback.
- GStack Browser: Provides robust browser automation for tasks like data scraping and end-to-end regression testing.
- Massive Parallelism: Enables engineers to manage dozens of concurrent sessions and PRs simultaneously, effectively turning the developer into a manager of multiple AI agents.