After reading this year's Google I/O summaries, my first thought was, 'Right, even more AI.' AI matters, but there was so much packed into the announcements. So I went through the list again, ignored the spectacular demos for a moment, and picked out the features that could remove JavaScript from everyday UI work.
One direction was especially clear in 2025: interface patterns that once required JavaScript components are gradually moving into the browser platform.
CSS carousels can cover part of a slider's interaction, while Anchor Positioning lets a tooltip or menu follow another element. Baseline provides a simpler label for whether a feature has broad support across the major browsers.
New platform features still have version and implementation differences. It does mean that, before installing another dependency, we can check whether HTML and CSS already solve the problem.
Three changes worth watching#
-
CSS carousels bring scroll buttons and position markers to CSS, so a simple carousel may not need a full slider package.
-
Anchor Positioning addresses the old problem of keeping a floating layer attached to its trigger, and it works well with popovers.
-
Baseline is appearing in more development tools, turning browser compatibility into something more readable than a long table of version numbers.
Built-in AI APIs attracted plenty of attention, but they still face model downloads, device requirements, and cross-browser support questions. The HTML and CSS improvements above are more likely to reach everyday projects first. When a release is this large, choosing what is useful soon matters more than translating every announcement bullet.
I now sort conference announcements into three columns: usable today, suitable for progressive enhancement, and worth remembering by name. Baseline helps with the first. Anchor Positioning often fits the second because an existing positioning fallback can remain. Device-side AI can comfortably stay in the third when the product has no concrete use for it.