A flood-watch office waits eight seconds on a blank box while a long briefing generates. What should the interactive path do?
Select an answer to reveal the explanation.
Short Explanation
Eight seconds of blank box is a dispatcher staring at nothing. Stream tokens as they are produced so the first words show up. A night batch can wait for the full briefing. Mixing this one call into a huge job, or writing a kernel, is the wrong live path.
Full Explanation
Token streaming sends pieces as they are produced so the first words appear before the full briefing is done. That improves perceived latency on an interactive flood-watch path. Overnight or batch jobs can wait for the complete payload. Mixing one dispatcher into a huge batch, or replacing serving with a kernel, is the wrong live choice.