The waiting problem in AI products (8 minute read) (opens in new tab)
Modern AI products and agents have largely ignored decades-old usability principles showing that users lose focus when systems respond too slowly, relying instead on vague spinners and minimal feedback that force people to invent coping behaviors like repeatedly checking tabs, refreshing pages, or monitoring long-running tasks themselves. AI tools should adopt the same progress indicators, ETAs, notifications, and detailed logs that operating systems and professional software have used for ye...
Read the original article