Beyond Conversation: Why Documents Transform Natural Language into Code
summary
Claim: structured, document-based natural language becomes a programming abstraction that forces clarity, versioning, and reusable protocols where chat produces context pollution and a problematic concurrency of one. Demonstrated tools and patterns: zapier-style handcrafted workflows, Gmail/Slack/meeting triggers, Notion as source, agents that open Jira tickets, and a human-in-loop approval loop. The prototype ecosystem—evoked by Wordware and criticized MCP—aims to run background agents that surface only failures for manual approval; reliability metrics and failure rates are not reported.
AI-written summary · about
chapters (12)
- 0:00Intro
- 0:19Problems with chatbased systems
- 0:52Context pollution
- 2:08Documents
- 2:55Concurrency
- 3:21Background
- 3:46Workflows
- 4:20Human in the loop
- 5:07Protocols
- 6:22Agent Economy
- 7:21Managing Agents
- 8:58Conclusion
related talks
16mBeyond the IDE: The Future of Developing Software With Agents
48mDialogues on Natural Code
25mBeyond the IDE: Toward Multi-Agent Orchestration
40mContext That Dreams
26mAgentic Fundamentals: How language models and code can communicate
18mHow structured thinking gives your AI superpowers
26mMaking Docs and AI Work Together: Agentic Coding in the IDE
29mFrom Copy-Paste to Composition: Building Agents Like Real Software