Amol Kapoor argues that coding agents can produce visual artifacts more effectively when they work in a format suited to structured text. He contrasts canvas-based editing and hand-authored SVG with HTML rendered by a browser.
Using a pelican drawing example and presentations, Kapoor describes HTML as an editable way to express layouts, typography and motion before rendering to formats such as PDF. He illustrates the approach with slides, documentation and video produced by his team.
Kapoor stresses that attractive formatting does not replace accurate content. He discusses supplying relevant source material to an agent while directing human attention to the story and the underlying information.
Watch on YouTube
