Beyond the Web of Today
summary
Intel lays out specific APIs and measured gains: WebNN enables on-device ML inference and demos show ~4× faster frames/sec vs WebAssembly for background replacement; SIMD yields ~3× typical speedups (one cited 160× case); WebGPU replaces WebGL for GPU compute; Compute Pressure exposes nominal→critical system states so apps can drop effects. Other claims: 65–70% of CPU time is spent on web workloads, 17.4M JavaScript developers, WebCodec cuts Zoom encode latency by ~75% and Meet sees ~20% battery improvements when using hardware codecs.
AI-written summary · about
chapters (9)
- 0:00Introduction and journey
- 3:51Web platform strengths
- 7:00Project Fugu and APIs
- 9:50Hardware acceleration focus
- 13:19Web machine learning
- 19:29Compute pressure API
- 22:00Media streaming improvements
- 26:15Foldable displays support
- 30:00Q&A and closing
related talks
39mThe AI PC and the Era of the Modern Web
14mWasm, WebGPU, & WebNN: How compute abstraction are enabling client side AI
15mWeb AI leaps forward on Intel AI PCs
18mWeb AI on next generation AI PCs
24mWeb Apps of the Future With Web AI
41mEscaping the Cloud: High-Performance AI in your Browser
16mAccelerating Web AI on
16mThe Web is AI Ready—maximize your AI web development with WebNN