We Made a Document Editor Where Humans and AI Work Side by Side

Overview

This episode dives into the development and functionality of Proof, a live collaborative document editor designed for humans and AI agents to work together seamlessly. Created by Dan Shipper and the team at Every, Proof emphasizes real-time collaboration, provenance tracking of AI vs. human contributions, and a lightweight, intuitive design. The discussion explores how Proof has transformed workflows, the concept of agent native design, and the evolving relationship between humans and AI in writing and planning.

Notable Quotes

- Text is really valuable, but it's also really cheap to produce. The way you might design how you work with text has to change. – Dan Shipper, on rethinking text-based workflows in the AI era.

- I love that I can see what it wrote versus what I wrote for me, which is so much better than a Google Doc. – Austin Tedesco, on Proof's ability to distinguish AI and human contributions.

- AX (agent experience) is just as important as UX in this new world. – Dan Shipper, on designing tools for AI agents.

🛠️ The Evolution of Proof

- Proof began as a Mac app focused on tracking AI vs. human contributions (purple for AI, green for human).

- The shift to a web-based, collaborative editor unlocked its potential, enabling real-time interaction between humans and AI agents.

- Brandon Gell noted that the collaborative web version quickly became indispensable for the team, highlighting its fit for modern workflows.

🤝 Humans and Agents Collaborating

- Proof allows humans and AI agents to co-author documents, leave comments, and track changes in real time.

- Austin Tedesco shared how he uses Proof for creative writing, texting ideas to his AI agent while running, and watching outlines take shape.

- The tool's provenance feature helps users distinguish between AI-generated and human-written content, fostering trust and clarity.

🧠 Agent Native Design Principles

- Dan Shipper introduced the concept of agent native design, where tools are optimized for both human and AI agent use.

- AX (agent experience) is prioritized alongside UX, ensuring agents can navigate and contribute effectively.

- The team discussed challenges like managing multiple agents in a single document and ensuring quality output.

📄 Practical Applications and Use Cases

- Proof is used for everything from daily to-do lists to complex growth strategies and creative writing.

- Brandon Gell described a workflow where his agent drafts a plan, Dan Shipper's agent reviews it, and humans refine the output.

- The lightweight, shareable nature of Proof makes it ideal for brainstorming, iteration, and short-term planning without the overhead of traditional tools like Notion or Google Docs.

🌐 Open Source and Future Potential

- Proof is open source, allowing developers to integrate it into their own apps or customize it for specific needs.

- The team envisions Proof as a foundational tool for agent native workflows, enabling seamless collaboration between humans and AI across industries.

- Kieran Klaassen emphasized the importance of Proof's simplicity, likening it to a sketchpad for ideas and collaboration.