Your browser now ships a language model. Chrome and friends expose on-device AI through a family of built-in APIs backed by a small local model, so inference can run entirely on the machine — private, free, offline, and low-latency. It's not here to replace GPT; it's a new tier you route the easy jobs to.
There is a category of page load that takes zero milliseconds, because the browser already did the work before you clicked. The Speculation Rules API is how a page tells the browser which links to prefetch or fully prerender ahead of time — turning the next navigation into an instant swap. Here's how it works, and how not to set your users' bandwidth on fire doing it.
The smooth crossfade-and-morph between screens that made single-page apps feel expensive used to cost you a whole framework and a pile of JavaScript. The View Transitions API turns that motion into a browser primitive: one function call for same-page state changes, one line of CSS for real animated navigations between separate HTML pages. Multi-page apps finally get SPA-quality motion for free.
For a decade, real-time on the web meant one thing: a WebSocket. But a WebSocket is a single TCP connection, which means one lost packet can freeze every message behind it, and it only knows how to deliver data one way — reliably and in order, whether you wanted that or not. WebTransport is the newer browser API that runs over HTTP/3 and fixes both limits: many independent streams, plus a lossy express lane for data that's worthless once it's stale.