My first reaction was actually mixed and slightly negative because I see Google attempting to become an aggregator which oversteps the bounds of simple search.
With the example that Google shows for Yelp, which would apply to any directory or review site, rich snippets give searchers way too much information at the level of the SERP.
The searcher may spend more time on Google skipping through listings now that ratings are available and less time on respective websites. Imagine Google allows review results to be sorted by ratings and the process now becomes really scary. Now add to this scenario that many sites allow Google to display their cache pages and the review site itself can easily become completely cut out of the process with Google being the complete content provider.
With the way search works today, someone looking for a restaurant or hotel may delve deeper into the review site and find lots of options per site, which also gives their advertisers or products lots of views.
However, when Google has the ability to bring more information to the surface, and you can't control which item (high or low rated) from your page is displayed as a result, your ratings site may easily see a sudden drop in traffic because Google has now become an aggregator taking another step down the slippery slope from search to content provider.
The whole concept here is more functionality at Google means more face time at Google which means more face time in front of Google ads, not ads on the advertiser sites, with such "improvements" putting the ad-based model of survival of many sites further in jeopardy.
Many will argue rich snippets are good for the searcher but not when it ultimately results in less content to search as more sites are squished out of business by such predatory aggregation.
Furthermore, the pros and cons of microformats are certainly not weighing in favor of the webmaster because microformats identify data for easy scraping making the jobs of the scrapers easier than ever with Google ushering it in. The most amusing part of this evolution is that that content within the microformats doesn't contain anything that's uniquely identifiable, such as a paragraph of unique text, so scrapers can take the microformats data freely not even concerned with a DMCA complaint.
Anyone using these microformats will be putting a bullseye on their site from would be competitors and additional aggregators trying to inject yet another layer between the searchers and the content they seek in a power grab for eyeballs and clicks.
If you must use microformats, cloak them to the SEs only and let your visitors see the data without all the identification tags.
IMO, we've just entered a new era where content control is about to become more important than ever before.
[edited by: Robert_Charlton at 1:25 am (utc) on May 13, 2009]
[edit reason] fixed typo per request [/edit]