Until now, including data feeds in web applications has been a painful process involving server-side proxies and a fair knowledge of XML data formats. The new API takes care of proxying automatically and offers a straightforward interface for developers to access syndicated data. No server-side code is needed at all. According to Google:
In contrast to the drag-n-drop Yahoo Pipes and the rumored Microsoft Springfield tool, the Google AJAX Feed API is oriented towards developers and not casual users. As such, it's a little harder to use but far more flexible.
Bret Taylor, group product manager for Google developer products, will feature the Google AJAX Feed API today in a talk at O'Reilly Web 2.0 Expo titled, "Who Needs Server-Side Code? AJAX APIs and Product Integration 2.0."
Update: Now that the documentation is available I've noticed several interesting things. For example I think the work being done for cross-browser XML access, the repurposing of the AJAX Search blog and group for all AJAX APIs, the details about FeedFetcher and the relationship to Google Reader, and the two-step process to load the Feeds library (needed because "Google is moving to a new model of loading AJAX APIs to make it easier to include multiple Google APIs on your pages") may be significant to developers.
For more information see: