An AI agent wrote a Go program. Then Anoop Gopalakrishnan (VP of Engineering @ Guidewire) hit Ctrl+C and couldn't exit. The agent that wrote the code had also written the tests.
In this episode of 1 IDEA, Suresh Mathew sits down with Anoop to talk through the Software AI Factory, his open source framework for running AI agents with guardrails at scale.
We cover:
- The five agent roles inside the Software AI Factory
- Why the implementer agent should never touch the tests
- When a deterministic gate beats an LLM review
- Why more AI-generated PRs don't mean more releases