Episode
Peter van Hardenberg - Ink and Switch, Automerge
- Published
- Jan 12, 2026
- Duration seconds
- 3032
- Processing state
processed
Actions
POST https://stenobird.com/v1/public/podcasts/devtools-fm/episodes/peter-van-hardenberg-ink-and-switch-automerge/transcription-requests
Idempotently request low-priority transcript generation for this episode.GET https://stenobird.com/podcast/devtools-fm/peter-van-hardenberg-ink-and-switch-automerge.md
Read the agent-friendly Markdown representation of this episode resource.
Summary
Explore the shift from cloud-centric computing to local-first software, where data ownership and offline capability are fundamental. Peter van Hardenberg explains how CRDTs and engines like Automerge enable seamless, real-time collaboration without sacrificing user control.
Topics
- Local-first software
- CRDTs
- Automerge
- Distributed systems
- Data ownership
- Ink and Switch
- Software architecture
- Decentralized computing
Highlights
- Main idea: Local-first software prioritizes data ownership and availability, ensuring users can work even when disconnected from the cloud
- Technical insight: CRDTs (Conflict-Free Replicated Data Types) allow for seamless merging of concurrent edits without central authority
- Failure mode: Using CRDTs for systems requiring strict scarcity, such as banking or limited-seat ticketing, is a poor architectural fit
- Practical takeaway: Columnar encoding can significantly optimize the performance of sync engines by reducing the metadata overhead of every keystroke
- Vision: The future of computing should move power back to individuals, enabling a culture of software craft and interoperability
Chapters
1:00The Journey to Research: Peter discusses his diverse background in software engineering, from physics engines to his time at Heroku.12:25Defining Local-First Software: An exploration of the core principles behind the local-first movement and the vision of the Ink and Switch lab.20:05The Mechanics of CRDTs: A deep dive into Conflict-Free Replicated Data Types and how they handle distributed state synchronization.27:30Optimizing Sync Performance: How Automerge moved from naive JSON encoding to columnar storage to handle high-frequency updates efficiently.35:15Decentralized Data Access: Discussing the design of systems that allow for encrypted, distributed data sharing and history traversal.46:30The Future of Software Craft: A vision for a computing ecosystem where software is customizable, interoperable, and owned by the user.