Hypermed-AI: Where intelligence Emerges from Connection
summary
Argues that combining hypermedia with AI makes workflows emerge at runtime: agents read action, context, results descriptions (HTML forms, MCP JSON manifests) and discover multi‑step paths rather than following hardcoded APIs; demos include the Grail project and past formats like HAL, Siren, Collection JSON and HTMX. Claims agents handle surprising requirements (example: updating an email triggers six or seven dependent steps), so metadata, links and forms become the mechanism for dependable autonomous interactions.
AI-written summary · about
chapters (10)
- 0:00Hypermedia AI intro
- 1:49Paris 1900 origins
- 4:19Otlet to Vannevar Bush
- 6:30Engelbart & Nelson visions
- 8:35HyperCard to modern hypermedia
- 9:40Hypermedia defined
- 11:49AI agents in workflows
- 13:15Modern hypermedia tools
- 16:46Summary and future
- 17:47Q&A
related talks
58mHypermedia APIs and the Future of AI Agentic Systems
49mThe Shape of AI
19mBeyond agents: AI as a creative partner
28mAgentic AI
51mSentience Agility: Harmonizing AI Agile Intelligence
29mAgentic AI: What It Is and Why It Matters
20mReimagining InnerSource for the Agentic AI Era
27mAgentic Engineering: Working With AI, Not Just Using It