Using AI Effectively in 2025
summary
Shows how to stream structured LLM output to React by sending token chunks from the server and parsing them as NDJSON newline-delimited JSON so partial JSON no longer breaks the client; demonstrates buffering and newline-splitting on the server, client ReadableStream + TextDecoder reads, and incremental UI updates for 20 items (cakes/movies). Replaces hand-rolled parsing with Vercel AI SDK streamObject + Zod validation, yields server-component UI fragments, and fits the core flow in under 50 lines.
AI-written summary · about
chapters (7)
- 0:00Introduction and setup
- 1:23Streaming UI workloads in React
- 10:07Streaming JSON responses
- 22:28Using AI SDK for streaming
- 28:49Generative UI demo
- 38:25Future of AI with MCP
- 43:42Closing and resources
related talks
38mUI Engineering for the AI Age
28mCreating fluent Frontend Experiences in an age of AI
17mReact Server Components in AI Applications
29mMaking AI better with React feat
14mStreaming UI and Scaling Full-Stack AI Like a Boss to Millions with Next.js
27mAI-Generated React Server Components
09mAI for React Developers: Opportunities, Learning, and Innovation
24mFull Stack Web Development with AI