Matt Palmer

Matt Palmer is a developer-experience leader, technical educator, and creator of software for AI-assisted production. He works at SpaceXAI through Cursor, after leading developer relations at Replit and developer experience at Conductor.

Palmer began in product analytics and data engineering, working primarily with Python. He wrote Understanding ETL, an O’Reilly guide covering data ingestion, transformation, orchestration, troubleshooting, and scalability, before expanding into developer education and AI-assisted software development.

At Replit, he helped introduce Replit Learn, a free educational platform that teaches beginners to build applications by describing an idea, evaluating an agent’s output, and iterating. His guidance on secure AI-assisted development emphasizes encrypted credential storage, input validation, authentication, version control, and keeping sensitive operations outside browser-side code.

At Conductor, Palmer focused on developer experience and the emerging discipline of content engineering: producing documentation, changelogs, product tours, websites, and video from the same structured foundation as the software itself. His Content Is Code demonstration reconstructed a Conductor interface using React and Remotion to create a programmable product walkthrough.

• Structure determines AI output. Palmer argues that reliable automation depends on clean pull requests, descriptive change histories, maintained documentation, consistent design tokens, and clear boundaries between application and presentation code. Strong models cannot recover missing organizational context.

• Video becomes programmable infrastructure. His project mcut is an open-source video editing SDK for TypeScript applications and agents, with tools for timelines, rendering, transcription, React interfaces, command-line editing, and MCP integration.

• Creator tools should solve practical production problems. Palmer built Shotwell, an iPhone screenshot editor, and Shades, a browser extension that masks sensitive information during demonstrations and recordings.

• Productivity is not progress. Palmer warns that greater AI-driven output can weaken critical thinking. His interest in computer-using personal agents centers on expanding what people can accomplish without surrendering judgment about what deserves to be built.