Latency, errors and accessibility of streaming
Comes up in: Frontend system design, AI system design
This concept is being written. It gets the full five steps — explanation, predict, interview angle, practice with tests, and review cards. Here is exactly what it will cover, so you can study it elsewhere in the meantime.
What it covers
- time-to-first-token UX (skeletons, "thinking" states, progress for long tasks)
- mid-stream errors and partial output
- retries
- rate-limit and quota messages
- screen readers and streaming (announce on completion or in chunks via polite live region, not per token)
- reduced motion for typing effects
Words you will know
What you will practise
- make the chat UI accessible with NVDA/VoiceOver test notes
Interview questions this answers
- How do you make streaming text accessible to screen-reader users?