I think this a fairly important inflection point. It's when semantic technologies begin creeping in to the mainstream as semantic technologies rather than obscure under the covers functionality as with firefox.

It's also interesting timing, coming as it does on the heels of their "Unreasonable Effectiveness of Data" paper.

I'd have to re-read it to comment, but at least some felt that it was being rather dismissive of semantic-web structures: http://www.betaversion.org/~stefano/linotype/news/275/

I've always thought they were a little too dismissive of semantic web structures, Peter Norvig in particular. But I also thought it was a ploy.

I think the problem is that the semantic web is fundamentally very similar to the sort of 'logical ai' (Prolog, Cyc, etc) that is often blamed for the AI winter.

Eliezer made a fairly convincing argument that a lot of this is the result of clever marketing campaign a few decades ago (http://www.overcomingbias.com/2008/11/logical-or-conn.html)

Well, so far nobody's really addressed the known issues with self-supplied metadata (see http://www.well.com/~doctorow/metacrap.htm for one good and old roundup)...

It's also much more comprehensible than previous attempts. For your average web developer it's just adding a few attributes here and there as tidbits, and as a benefit, you get bells and whistles from google. The big turning point will be if a browser maker with market share builds support for recognizing them - the minute that happens they will pop up on web pages everywhere.

Seems like Google is playing catch-up to yahoo for once. This has searchmonkey written all over it. Yahoo though, seems to be still without its mojo so cant see it capitalizing on this thing - even assuming it goes big.

Wolfram Alpha already breathing on their necks?