Thursday, August 30, 2007

Professor Scoble on the Future of Search

Professor Scoble is lecturing us on search engines. He tells us how they work. How facebook works. How Techmeme works. Thank you professor!


It hard to describe how I feel about Robert Scoble. He used to work at Microsoft, he was sort of cool then. But then he quit and started working for PodTech, a podcasting company. His videos were real podcasting material -- badly filmed, hardly edited. Even though he had a $4,000 camera, which he could not mention enough during his interviews in the beginning.

Scoble believes himself to be a visionary and profiles himself like that. He feels he is important. He is a true pet peever. Most of his talks are about what he cares about and his super power RSS-reading abilities. His current pet peeves: facebook, techmeme and twitter.

These three lectures are no different. It's about facebook, Techmeme and Mahalo. He thinks that these three together can beat Google in four years.

Here's why he's wrong. To start off, he's addressing a problem that does not exist. It's not like Google's search results are bad or not relevant or of poor quality at all. You couldn't say that of Altavista, they were bad and everybody knew it. Second of all (and I assume you watched the videos I linked to) his idea does not scale. He mentions this as a problem in his last video and does not convincingly explain why it does. Second of all, the system is based on trusting humans. Humans cannot be trusted, machines can. The system is based on human labor. Humans are expensive and slow. A search engine like Mahalo only returns results for a few terms, it's really unusable. Sure, in 4 years it will be much bigger (if it still exists, because projects by Jason Calacanis -- who founded Mahalo -- tend to die), but by the time they covered all search terms, the early ones will already be outdated.

As I said, stupid idea.