The GitHub Podcast

LIVE from GitHub Universe: Angie Jones on Goose, MCP, and the Real-World Future of AI Agents

Episode Summary

Abby sits down with Angie Jones, VP of Engineering at Block, live at GitHub Universe to talk about Goose, Block’s open source AI agent and reference implementation of the Model Context Protocol (MCP).

Episode Notes

Abby sits down with Angie Jones, VP of Engineering at Block, live at GitHub Universe to talk about Goose, Block’s open source AI agent and reference implementation of the Model Context Protocol (MCP). Angie shares how Goose went from an internal tool to an open source project that lets the community drive features like multimodel support, and how Block’s 12,000 employees across 15+ job functions (not just engineers) now use agents every day. They dig into practical, non-hype uses of AI agents: detecting when students are struggling, triaging open source issues, segmenting 80k+ sales leads, and even letting a salesperson “vibe code” a feature on the train. Angie also talks about trust and control when giving AI access to codebases, why developers are tired of flashy demos, and how her new AI Builder Fellowship is designed to support the next generation of native AI builders. 

Links mentioned in the episode: 

https://angiejones.tech

https://github.com/block/goose

https://github.com/block

https://github.com/modelcontextprotocol

https://github.com/features/copilot

https://testautomationu.applitools.com

https://www.selenium.dev

https://playwright.dev

https://www.cypress.io

https://code.visualstudio.com

https://www.salesforce.com

https://github.com/martinwoodward/pyfluff