Z-Blog

Image via WikipediaAs Zemanta’s CTO I spent last three days at Semantic Tech Conference trying to figure out where all this semantic stuff is going. The question if this semantic technologies are going to benefit enterprise and vertical markets is largely resolved. The answer is clear yes.

Large scale consumer applications on the other hand have some problems, reception of two “white knights” which launched recently (Powerset and Twine) has been cold. The issue is not that semantic technology does not deliver. The issue is an overpromising they ere doing. But luckily both companies already know that and are starting to manage expectations and hopefully media will stop looking for next “google/facebook killer” and start focusing on direct benefits these applications are providing.

Powerset is great way to search and browse Wikipedia. If we look it only in this optics then it has delivered. Yes, they might index whole web at some point in time and we will judge them on that goal when they release the product. The same goes for Twine. Yes, tagging objects with objects can bring something big down the road, but for now we just have to take it as it is - good colaborative information filtering.

Zemanta also had similar problems with user expectations. User that wrote two paragraphs about how he is testing Zemanta, and then just two sentences about pop culture, was expecting us to know that he really wants images to illustrate that minor part of the story. And user that got 6 good , directly usable images, complained that other 3 were wrong.

Is there a way for startup to spark imagination of its users and media and still not overpromise and disappoint? Yeah, there was a star-up that has done it in 1999. But I don’t believe we can expect such extreme jumps in technology as Google has made at the time. Semantic tech conference was great, but no revolution on the horizon.