Jimmy Wales: People are smarter than Google
Google may rule the roost when it comes to search but Wikipedia founder, Jimmy Wales, believes there’s a better way to do things.
With the backing of investors including Amazon, his company Wikia has developed Wikia-Search, a collaborative search engine that allows users to influence the rankings.
As Wikia Search hits its first-month anniversary, we catch up with Wales to chat about his vision for internet search.
Why does the world need Wikia Search?
Wikia Search is as much a political statement as anything else. We want to show that it’s possible to have good, decent quality search with open-source software and community participation. If we can do that it means we don’t have all the secrecy around search that we’ve had in the past. We’ll see if that’s true or not. Obviously, I have to build it first.
And how will it make that statement?
The main thing is putting the editorial control in the hands of the community, that’s one of the core principles. Doing it all with open-source software, so that we publish all the algorithms – I think that’s important. If you look at the results of Google, Yahoo, Ask, they’re very similar these days and we need to hit that level of quality otherwise we’re not very interesting.
If we’re successful at figuring at how to get lots of people involved, lots of human judgement involved in a productive way, we might actually surpass that quality and solve some of the problems of search. But that’s obviously a dream for the future – the first stage is just to replicate what the major players are doing.
How long do you think that will take?
I think it’ll take a good three years, but that’s just a figure I made up for my feel of what the product is going to entail. But we’ll be looking at search quality in six months and again in a year.
But if the relevance isn’t there now, how are you going to keep people coming back?
We need to get there pretty quickly in terms of getting some decent relevance and then we need to have tools that people find useful. For example, we’ll be coming out with a toolbar soon that people can use for social bookmarking, so basically tools that people can use to give us the information we need that aren’t necessarily tied up with doing searches on the site. That’s part of our early approach.
But if you publish your algorithm, what’s to prevent people gaming the results?
It’s the concept of the social network. If you look at the amount of true spam you get on Facebook, it’s basically zero, because you have a community of trust and people get bounced out pretty quickly if they’re misbehaving. And that’s the approach we’re looking at, broadly speaking.
The real concern about people gaming the algorithm is that they’re doing something through a variety of links and a variety of sites and through various keywords; they’re doing something to trick the search engine and it’s pretty hard to trick humans. If I do a search and the result I get is not what I wanted, then it’s not what I wanted – it’s pretty hard to game that.
But in order for that to work, you need a large userbase. How’s take-up been so far?
Cranking along. Obviously we had a huge spike of traffic in the first days because of the unreal amount of press coverage. It’s come down to be a little more sane now and we’ve lots of stuff under way. We’re pleased with the progress, but we’re not really giving out metrics right now.