Mickael Arrington interviewed David Kirkpatrick about his latest book: a very interesting conversation, less simplistic and as funny as his subsequent blog post where he compares Kirkpatrick to a teenager in love. The complicity between tech journalist is interesting—and characteristic of the intersection of two worlds where cooperating with your competitors makes your output, and doesn’t prevents ruthless competition.
The introduction lets Kirkpatrick repeat his view, after decades of covering the tech industry, on what makes revolutionary company: it is those who empower people. Facebook, according to him, allows people to make a statement and be heard—a essential point, and the main point of disagreement I have with his book. I wholehearted agree with the details he then gives: the Newsfeed reversed the model of messaging; now people do things and talk about it, while people who friend you actually subscribe to you feed of actions — the way readers are determined is what Kirkpatrick calls a “complex”: settings let people twist thing, but it it mostly an algorithm that pools together people who have more common interest than others.
My point of dissent is that this process is neither clear, nor neutral and no one seems to make fully sense of it.
“With much power comes much responsibility” the quote from Lord of the Rings, and a geek motto; it remembered many visionaries to prepare for the challenges their technologies would bring, but I don’t see too much of that concern around how the newsfeed changes politics and whether balances such as extremism vs. freedom of expression, local vs. global focus or more generally the role of personal kirks is to be addressed; according to Kirkpatrick’s book, Facebook calls it the Squirrel principle (people care more about a squirrel dying in their yard than a genocide in Africa) — but the treat it more like a human fact to be indulged, rather than the concern that journalist have to use local ads to pay for a war correspondent.
The rest of the interview is certainly worth the hour, if only for the many caveats from Arrington, who pretends to have to explain to a colleague why he has to be mean. This should be compared to Kirkpatrick refusal to be mean, answering one of the last question, about the board of the company.
Kirkpatrick repeats that Facebook was not the first or the only company to re-organise information that way, but is by far the largest, hence its importance. Arington’s joke on the telephone connecting everyone was spot on — but Kirkpatrick missed to opportunity of the connection to show how Facebook treated public discourse. I’m not sure Kirkpatrick insists enough on how difficult this really is: being large not just a question of numbers, but no service that size can ignore new types of problem—generational, cultural, etc.— and Facebook succeeded where say Digg remained a tool for a single, albeit diverse, community.
The first question corresponds to Arrington’s main criticism: isn’t the book too one-sided: World Peace, really? I regret this question focused on the Winkelwoss twins (I like Kirkpatrick justification of his talking to them, though: we’ve heard enough in court) while Open Social Graph advocates have far more points — I did like how Kirkpatrick introduce the likelihood of seriously regulations, but regret he wasn’t more specific on what could happen; “corporate identities are not new” no, neither are laws on what corporation should do with those, nor internationally coherent. The scope of issues is too vast not to be addressed.
The discussion on how Facebook handles censorship, using local legal basis would be a riveting story on cultural adaptation, and the threats of physical violence against Zuckerberg in Pakistan show how important.
The rest is rehashing questions well covered in the book (but probably a faster way to understand them, including the tone at stake here): the attitude towards growth over the business model, that Zuckerberg considers this more a project about society than a for-profit company, Sheryl Sandberg’s and Sean Parker’s role, users’ paradoxical attitude towards changes and privacy and the way Facebook uses that to force innovation while limiting the offenses through opt-out features.
Kirkpatrick confirms that the internal attitude towards innovation is indeed “Push, push”. I like his description of his and many pundits’ opinion on the matter: the attitude is cavalier and shouldn’t be encouraged, but it’s hard to argue against it because so few users behave like they care. Zuckerberg’s enlighten dictatorship does indeed look like Steve Jobs’ “I know better than the users do want they want.”
Kirkpatrick sees a pattern among the five crisis against users (no surprise), and also confirms that the first one, around the Newsfeed was the most contrite Zuckerberg ever got, down to a more recent “Things will blow over”: I’m not sure why he doesn’t see it either as a diminishing trend (because of an habituation of critics, a dilution of users who care less, or an improvement of Facebook staff to present the change), or corresponding to the magnitude of the change: Facebook Open Graph API would then be legitimately second in volum. He also confirms the technical possibility that (some) Facebook employees can access private account information (chat and message) and the dependance of Facebook’s internal respect on its own policies: employees were fired upon this, but the scandal prone public/media might mostly care about Facebook’s own CEO, being “sole in charge”, to enforce the reserve on himself. Controlling such accesses by making them more visible might be a better, more pediagogical solution. Eg.:
One of your friends (or friends of friends) has signaled your profile photo as being inappropriate; we are investigating it—you can anticipated in three ways:
- My bad: I see why; let me remove that picture myself;
- I see why, and I would like to make a case against that person;
- How dare you write to me in that way?
Someone not related to you/
One of our algorithms has signaled your profile as problaby not from a real person, or not using your real name; we are investigating it—you can anticipated in three ways:
- No, my name is really [James Bond] and you can’t imagine the grief I get for it;
- I’m modified my name to protect my privacy; please show me how to set my profile so that no one I don’t approve doesn’t see it anyways;
- I am not a “real” person; please show me how to set up a page for an institution.
This would lead to more reflexion on how to give access to Facebook data, at the statistical level or otherwise, and later include social scientists to study them, and the risk of dis-anonymisation through such process, etc.
Back to the interview, one interesting point that was Facebook makes money from several, distinct sources (four, and apparently evenly split): that is very reassuring news for the business model. However, they still make two-orders of magnitude less money than Google. Zuckerberg cares deeply not to have irrelevant information on the site, and I’m not sure that revenues from relevant ads along the Newsfeed can soar two orders of magnitude… except (and I blame both tech journalists for not drawing that parallel) that the relevancy criterion was, and still is, Serguei Brin and Larry Page’s original critic of ads — yet, search advertising seduced them right-away.
“You friend Lucy will attend Twilight Première, will you come too?” is a great pitch, but there is only so much ad that you can sell by using such pretexts, while using social recommendation to non-newsfeed triggered queries (about travel, services or product to buy) should be huge. All this should be a very big signal that Facebook is considering Search, very much like Google does. I can only interpret Arrington and Kirkpatrick silent evasion as an embargo on this issue. Tom Krazit stirred up the flames on that, just last week.
I’m waiting for John Battelle to weight on the Search aspect of things—but things are clear to me on that front: it’s a race between giants to make the web semantic through meta-data. Facebook leverage of the Like button to stir users’ initiative is better adapted to their passivity, and far more encouraging for website owners than Google’s Intelligent Snipets. What company will make a similar step forward to encourage the upstream, language-processing kind of semantic against Google’s Voice? Apple-controlled Siri?
All in all, a great interview, but I still believe their are some misconception on the service that need to be addressed: the biggest is the actual, fair empowerment of users; the second is new models to explain how Facebook handles privacy; competition and data-portability would demand another book, very different from Kirkpatrick’s, while a row-full of books should be written on the second search battle looming.