A large group of international tech rock stars are at the Digital Life Design conference in Munich today and friend of RWW Martin Källström of pre-launch search startup Twingly sent us a rough transcript of a particularly interesting panel this morning.
In this discussion, titled Humans Disrupting Algorithms, WIkipedia founder Jimmy Wales talks about his new search engine Wikia Search, Jason Calacanis talks about his human-powered search service Mahalo and there’s cameos by Google bigwig Marissa Meyer and international man of mystery Michael Arrington. Wikia Search and Mahalo are taking very different approaches to search. It was an interesting enough conversation that I read it from start to finish and thought readers here might want to as well.
JASON
We are an editorial company at Mahalo and we’ve been around for six months now. Here is an algorithmic search result for apple pie, filled with spam and weird stuff. Martha Stewart is not even here.
One person can pollute the internet with hundreds of thousands of pages in a matter of minutes. And the content…We are not trying to apply humans to any search imaginable. If you look at the long tail of search we are looking at filling the top spots with journalistic search results. Martha Stewart has to compete
with the slime buckets that are filling the web every day.
My belief is that for top searches human results will always come out on top. If you look at googles result for “paris hotels” it is filled with crap. Our search looks like this. When was the last time a clean directory was available on the web. Probably 10 years.
We have a social network with a purpose. Every link that is submitted are investigated by our editors to see if it is spam or not. We have 500 links coming in per day and 12000 people registered as editors. We
are not afraid to ban links or ban people. I have had seven of my own link submissions declined.
So that is basically what we are doing, six months in.
JIMMY
The thing we need more of is transparency and quality in search. For me a big part of is open source software and openness which traditionally has benefited us very much. The same things are happening in search.
JASON
A study shows that 60% of people are not happy with search results, up from 50% last year. But in order to take users from Google you need to be a magnitude better, not 15% better. But it is a fallacy to believe
that you need to be better than Google to compete in this space. We come in with nothing in market share, just as Google did some years ago. 1% of search is worth 1 or 2 billions of dollars.
JIMMY
My view is different on what it takes to beat Google. We have a situation where the search result quality is very similar between Google, Yahoo and Ask. We are no longer in the era when Google was a
magnitude ahead. They are basically competing on brand rather than technology. If search is competition on brand we’re ready to take them on any day.
Our product is not as mature as Jason’s. We have a miniarticle about the search and then the algorithmic search result. The clicks then go into refining the algorithm by being feedback. Users can also work with us to submit new algorithms and participate in the community.
The answer in search is not the next Larry and Sergey geniuses coming up with a new..
JASON
We’re a google partner, we have their ads and default to their search result when we don’t have one of our own. I’m not as brave as Jimmy in jumping in front of the Google bandwagon to compete with them
algorithmically. I’m pretty sure Google have people cleaning up the search result, even though they have
Voting is to easy, people just run down the page. It will cost us 50 cents to a dollar per link, but we will have the most beautiful, clean index available. We have 60 full time workers. I’ve come up with this wonderful model to actually pay people for work they are doing. Jimmy here invented work for free in open source and I’ve built on that and starting paying people for work. I like the idea that people can send their kids to daycare or maybe pay mortgages.
JIMMY
Nobody works for free. This is one of the major errors people make when they talk about things like crowdsourcing. It’s like basketball, when you see people playing in the weekends nobody says “these people are suckers, playing basketball without getting payed”
Everybody that sees a business model and moves in that direction is somehow exploiting it.
MODERATOR
You are getting terrible reviews and really no usage.
JIMMY
Thanks to all the bad reviews we are seeing usage increase a lot. We have now around 20000 people registered. It is hard to launch something that involves a lot of people. It is absolutely correct to say that
there is a chicken and egg problem.
JASON
This is not an easy space to compete in. If you are moving into this space you are probably looking to investing 50 millions of dollars during the first 4-5 years. It is probably insane to get into this space.
JIMMY
People love to talk about human powered search vs the algorithm. But I want to say that Google have hundreds of people looking to increase search quality. They are making editorial decisions even though they are engineers.
MARISSA MEYER Google
I have obviously a ton of thoughts. It is like before when you could organize your own documents into folders on your computer. Nowadays you absolutely need to have desktop search. The same thing applies to the web. To take an algorithm and enhance it with editorial without introducing bias is the solution, if you can find it.
JASON
I agree with everything she said. Can I work for you?
MARISSA MEYER
The problem isn’t with the searching process, it’s with the result. You can’t do the fat If you read Chris Anderson, the real value is in the long tail. So if I did a startup I would do something else. I would do a Facebook.
JASON
Ok. I will go kill myself now. My dream just got slashed by one of my idols. No, but I think fat tail equals humans. Long tail equals advertising.
MICHAEL ARRINGTON techcrunch
Jimmy, I was very hard on you in my review. Putting that aside I think you have a great service. But there are very few companies that have managed to make a for profit business engage people to work for them.
So in your heart, are you sure that people will do the work, or are you still not sure?
JIMMY
Again, people are not working for free, they do it because it is rewarding for them. Of course not, we might fail.
JASON
We might fail to. We could possibly fail to. In my heart I know that.
MODERATOR
Do you feel any resistance in getting people to work for you for free?
JIMMY
No, we don’t. And we do lots of other stuff. Of course we are for profit, we need to run our servers and so on. No one asks Youtube if they are meeting resistance in getting people to upload videos for free.
A final thought, We have absolutely no idea or any plans for introducing advertising.
Thanks a lot for the transcript, Martin!
