Freebase's $42M! Can the top-down semantic web scale?

January 15, 2008 Filed under: Web Startup

I just read this: MetaWeb, the folks who power Freebase, a semantic database, got another $42 million in funding.

This shows a very serious investment in the semantic web... but instead of a distributed semantic web, which is a strictly top-down or bottom-up semantic web, it seems to be an investment in the private, centralized semantic web (and a strange mix of top-down and bottom-up, adding in the proprietary nature of it).

Now, $42 million dollars is a lot of effing money -- like Ning size money (well, they got $44 million) -- so the folks over at Benchmark and Goldman Sach must this MetaWeb's technology can really scale from a business development side of thing (substituting for the other way to scale, which is get actual people to use it and survive on much more modest amounts of cash).

But I'm not sure this is how things work on the the Internet. Wikipedia, the most direct competitor to Freebase, scaled without such money; they scaled with people.

Freebase may want to take a different approach, of course. Remember that open Wikipedia API I called for a few days back, and apparently Wikipedia is working on? Well, Freebase has an API, and it's open to web developers to build their applications on it (we plan to have some integration with BricaBox).

But a quick glance at their Featured Applications list show very little activity on the developer front (I've also subscribed to their developer mailing list for the past 9 months, and seen very little activity). One problem is that there aren't many categories which are fully seeded with information -- a problem Tim O'Reilly found in his gushing article about them last year.

And here-within lies the problem, perhaps:

With an API out there, it may be that Freebase is expecting to GET more information from third-party developers than they GIVE. And with a $42 million bounty over their heads now, you know Freebase is going to be trying to GET a lot.

comments powered by Disqus