Such a specific, real problem, and the diagnosis is spot on. Native deltas arrive bursty, and the UI is what turns that into something readable, so the smoothing belongs at the render layer. A small buffer that paces token display makes streaming feel dramatically calmer without touching the model.
Kartik N V J K
AI Developer | Making AI reliable, trustworthy & accessible to everyone | Active community contributor
Such a specific, real problem, and the diagnosis is spot on. Native deltas arrive bursty, and the UI is what turns that into something readable, so the smoothing belongs at the render layer. A small buffer that paces token display makes streaming feel dramatically calmer without touching the model.