DLD: Wales v. Calacanis
Jason Calacanis makes the case for human v. algorithmic (SEOed or spammed) search and then shows off his apple pie and his new social network. Jimmy Wales says what we need is more transparency and openness in search. “This is something that has benefited us across the internet,” he says. David Kirkpatrick, the moderator, says that given Google’s incredible growth, people are satisfied with search. Calacanis mentions a study that says 60 percent are dissatisfied (anybody have that citation?). Jimmy says that the leading search engines — Google, Yahoo, Ask — return essentially equivalent-quality results. So he argues that they compete on brand.
Marissa Meyer of Google is sitting in the front row right in front of them. I’m betting she’s not scared. She’s the fourth member of the panel just by her presence.
Jimmy says that Wikia will have not only the functions Jason’s Mahalo has but also user-submitted algorithms. This goes to yesterday’s theme of software being content — and open software requires open algorithms. Jason says he doesn’t think Jimmy will be unsuccessful trying to “jump in front of the Google train” in algorithms.
Jason says that he has a “pretty good suspicion” that Google is using people to clean up results. Marissa shakes her head no.
Jason says he has 60 fullitme people and 400 freelancers. “I believe in paying people for work,” he says; his zing on Jimmy. “I’m a writer by trade and I take offense when people try to devalue writing.” Do you think this way about Wikipedia, Kirkpatrick asks.
“There’s a very strange element to it, cult-like some people would say, that they want content to be free.” Jimmy responds: “Nobody works for free… What people do for free is have fun…. We don’t look at basketball games and people playing on the weekends and say these people are really suckers doing this for free.”
Jason has his teeth in the leg. He says it’s hypocritical that “everybody in the Web 2.0 industry should become millionaires and billionaires except for the writers.”
Jimmy: “To me this like the debate we got over a long time ago because it was so stupid that open-source software is communism.”
Yes. It’s about aligned interests.
Kirkpatrick pushes Jimmy and says he has a chicken-egg problem: Wikia is getting “terrible reviews” and he needs people to build the product but he needs the product to attract the people. Jimmy agrees about the chickena and the egg and says the bad reviews also hatched ore audience.
“If you’re going to come in the search space you need to invest at least $50 million to do this… This is not for the faint of heart. A major of search companies fail.”
Jimmy says it will take at least two years to get to industry-standard results.
Marissa from the front row says that it’s a mistake to return to the human-created directory. she agrees with them that there’s a false-dichotomy to look at this as all algorithmic or all human and, as Jimmy pointed, out, Pagerank is edited by humans — engineers, he says, but humans. She pushes back on Jason’s model of trust and says that may not be the right model; expertise matters.
“I agree with everything you said,” Jason says. “Can I go work for you?”
Esther Dyson says the problem with their model is the long tail. If she were starting a company she’d do somethign else. “Ok, I’m going to kill myself now,” Jason said. Punch line. He says that the fat tail will be human, the medium tale social, the long tail algorithmic. And he says that the advertising interest is in the fat tail.
Michael Arrington says he was very tough on Wikia when it started but has “promising aspects.” He returns to the playing basketball for fun question and says “that there are very few examples of for-profit companies getting people to do their work.” Not sure I buy that as a rule. He acknowledges Digg. I’m seeing networked journalism. I do believe in sharing revenue (a better model than staff work given the business realities today). Jimmy responds that “if we ask people to do work we will fail” but if he provides tools they want to use for their purposes they will succeed.
At the end, I ask about advertising and whether they will concentrate on the fat-tail brand, display ads and Google will have the scale to do automated advertising. Turns out that Jason will likely turn over all his inventory to Google to start and then, as with similar companies, will grow sold advertising. As the last word, Jimmy says, amazingly, that he hasn’t thought about advertising.
January 21st, 2008 at 10:01 am
Mahalo is just SEO.
January 21st, 2008 at 10:08 am
Information retrieval is complex. The big three do a fairly good job on the straightforward types of searches that are most common, that is: people, products and events. If they didn’t people wouldn’t be using their services and they wouldn’t be attracting advertisers.
Specialized searches (say for scholarly research) are not going to be moneymakers and this could be a useful area for an open source or non-profit entity to tackle. One size does not fit all.
January 21st, 2008 at 12:13 pm
I’m sure the same stats Jason is offering apply to lots of things (Windows OS comes to mind). LOTS of people are ok with “good enough” even if they say it ain’t so in a survey. So much so that huge brands are made on “good enough”.
What matters is how consumers actually behave and consume. Not so much how they answer a survey.
January 21st, 2008 at 12:38 pm
what’s a short sweet definition of the “fat tail” in the context of this discussion?
January 21st, 2008 at 4:10 pm
Nice coverage…Hybrids appear to be a safe passage in light of the dissatisfaction. Sorry - don’t have the actual stat but have heard it more than once. Local search in particular suffers from poor experience which is a key driver to directories rushing to get “human” generated content in the form of reviews.
January 21st, 2008 at 4:24 pm
@Matt - I think he means “head,” not fat tail, which doesn’t really make sense unless he’s referring to the larger part of the tail in a typical Pareto or Zipf distribution curve.
January 21st, 2008 at 9:51 pm
Awesome break-down of the panel. Thanks Jeff.
January 22nd, 2008 at 1:25 am
[...] de post van Anderson grijpt weer terug op een verslag van blogger Jeff Jarvis] Stem of voeg toe aan [...]
January 22nd, 2008 at 8:06 am
[...] Jason Calcanis squares off against Jimmy Wales in a discussion around human vs. algorithmic search. Jeff Jarvis’ post picks out a juicy Calcanis quote: [...]
January 23rd, 2008 at 11:08 am
[...] Jeff Jarvis, der die ganze Diskussion gewissenhaft mitgeschrieben hat, [...]
January 23rd, 2008 at 11:08 am
[...] Jeff Jarvis, der die ganze Diskussion gewissenhaft mitgeschrieben hat, [...]
January 23rd, 2008 at 11:41 am
[...] It is an interesting way to think about it. Of course, it is all ad-supported and it does make a good case for Mahalo. [...]
January 24th, 2008 at 5:51 am
> Jason says he doesn’t think Jimmy will be unsuccessful
Is this correct? Shouldn’t this be “doesn’t think Jimmy will be successful”? An important difference in this case. And what would success mean, slowing the train, or being run over by it (:-)?
February 20th, 2008 at 7:20 pm
[...] for video online Interesting discussion out of a panel at the Digital, Life, Design conference from Munich a month ago. Mahalo, Wikia, [...]
March 9th, 2008 at 3:35 am
[...] Jimmy Wales (Wikia) and Jason Calacanis (Mahalo), with Marissa Meyer (Google) discussing human-driven search (Mahalo and Wikia) vs. algorithmic search (Google). “Esther Dyson says the problem with their model is the long tail. If she were starting a company she’d do something else. “Ok, I’m going to kill myself now,†laughed [Calacanis] He says that the fat tail will be human, the medium tail social, the long tail algorithmic. And he says that the … [...]
March 18th, 2008 at 12:23 pm
[...] (Wikia and Mahalo) vs. Algorithmic search (Google). Jeff Jarvis of Buzz Machine was live blogging here and the Long Tail blog covered it [...]