Don’t Make Me Prompt: Intelligent Interfaces Users Actually Like
summary
Contends that prompt‑first AI UIs create high cognitive load and mistrust and shows concrete alternatives: smart glasses (Meta Ray‑Ban 7 million sold and Xiaomi QON 12 million sold), voice plus pointing prototypes, a Norwegian bank agent tested over a 15‑turn scenario, Miro and Krea AI one‑click summaries, and a hardware shortcut saving ~4 hours/week. Gives numbers (70% intend to use AI vs 37% weekly users; average conversational session ~80s) and cites failures like automated car‑damage billing to argue for human‑in‑the‑loop designs.
AI-written summary · about
chapters (6)
- 0:00Introduction and history
- 10:17Smart wearables evolution
- 21:16AI interface challenges
- 30:30Natural conversation vs precision
- 38:16Six HAX principles for AI
- 49:29Q&A and closing
related talks
28mHuman Computer Interaction: The UX of AI
57mDesigning Human-AI Experience
46mDesign Patterns For AI Interfaces
27mMisplAIced: Good UI Needs Good UX
47mThe Art of Prompt Engineering: Best Practices for Effective AI Interactions
42mUX+AI: Mastering the UX of AI-Powered Products
45mKeynote: Conversations with Machines
34mWhen Software Understands the User