Developed by San Francisco’s Radar Networks, Twine is being billed as “the first mainstream semantic application.” Those are big words, and a lot is resting on them. Semantic capabilities have long been seen as the underpinnings of Web 3.0, but, to date, the technologies have been their developmental phase, and the semantic web has remained largely theoretical. Twine is a social bookmarking site that uses statistical analysis, natural language search, and semantic search; the three complement each other well. Twine was running a bit slow when I used it, but once I had created my profile and entered my interests, the results were impressive.

What is different about Twine is the depth and the breadth of the articles it finds for you. Often times a trip through a news aggregator or social bookmarking site reveals a lot of chaff; sure, there was chaff with Twine’s results, but it was stimulating chaff (that still happens less often than is possible with the internet, your mind being pulled in favorable directions it didn’t intend to go . . . aha! that’s the semantic part.)

Twine is less impressive in its groups, called twines. Articles obtained through twines were less diverse and interesting than those obtained through the tags generated by my profile. If Twine takes off, though, this will change. Tags beget better tags, so if Twine can build a strong base this will take care of itself.

Using your own intuition is only as good as your intuition; Twine’s AI shows how the semantic web could fundamentally change, at least enhance, our relation to information. The Wiki model has greatly increased our epistemological range; the advantage of semantic search could be speeding up knowledge diffusion until it approaches real-time, following the temporal acceleration we’re seeing as the Twitter generation.

By Mark Alvarez