Subscribe to

Archive for April, 2007

[berkman] John Clippinger: A Crowd of One

John Clippinger is giving a presentation about his just-published book, A Crowd of One: The Future of Identity. [As always, I’m typing quickly, missing some stuff, getting things wrong, and making a seamless talk sound all choppy. But in this case, the remedy is easy: If you want to know more about what John is saying, buy his book.]

John approaches human nature through evolutionary biology and neuroscience. Identity, he says, is social and multiple. Trusted identity is essential for community, he says. And he’s interested in how virtual worlds “allow us to build new kinds of institutions, economies and identities.”

The brain is not a blank slate, he says, citing Steven Pinker. The brain is “highly specialized, opportunistic, and jerry-rigged.” Some of our most important decisions originate at a prec-conscious level. This is very different from thinking we make rational decisions. “It’s more a reflex.” He points to our “mirror neurons,” that enable us to have empathy. Descartes, Hobbes and Rousseau, and the Enlightenment are wrong. Research shows that our natural inclination is to reciprocate, trust and coordinate. Virtual worlds are the new state of nature. You may think you can create any identity you want, but “our identities are socially embedded.” And we all have multiple selves.

How do you have a trusted community on the Net? You need a persistent, trusted identity, says John. “But the Web was born without an identity layer.” We need one. Just look at all the fraud, flaming and phishing. “How do you make people accountable for their actions without having overly draconian measures? You have to have some way of creating a cost for breaking the rules, being deceptive, etc.” John refers to biological signalling theory — there’s a cost for deception. [I may be getting this wrong.] You want to make the cost greater than the payoff. That’s essential to any kind of trust network, says John.

In re-imagining identity as the virtual and real worlds become more intertwingled, people will want control over their identities. They’ll want to have a persistent identity. They’ll want multiple identities, the ability to take their identity info in and out of different virtual worlds. They’ll want a range of degrees of identification, from anonymity to authenticated anonymity to complete disclosure. And they’ll want to develop peer networks of trust and authentication.

Over the past two years, John’s been working on a project called “Higgins,” an open source interoperable identity system. (It’s called “Higgins” because higgins is a long-tail mouse.)

We are getting “new narratives about cultural and political futures, not laden with moralistic doctrine.” This is a kind of “social physics”: there are some predictable behaviors and phenomena. It looks for “evolutionary stable strategies.”

There’s an opportunity, John says, to invent new digital institutions: governance mechanisms, more reliance about measured risk and reputation, transparency and accountability for all forms of authority, and acceserated social innovation through digital experimentation. He says the Chinese are very interested in social physics because they want to know if there are rules are principles they can use. [China’s interest in social physics as a way of predicting and managing social behavior is not necessarily a good thing.]

Q: [me] Having an identity layer would solve of bunch of problems, but is there demand for identity itself, as opposed to a demand for solving those problems?
A: At SecondLife I was surprised that people do want to be able to authenticate themselves to others. But that doesn’t mean they know your real world identity. There are degrees and types of authentication and identity. The user gets to control it. You may give up small attributes or fragments of your identity for particular purposes in particular circumstances. Community norms will arise to govern that.

Q: Is it to authenticate you as a consistent person or to get to a level of trust?
A: There is a need for persistence, frequently, although that can just be a number. And there’s another issue about whether you can authenticate the claims you make about yourself. Another party may have to authenticate those, and they may change over time.

Q: How will reputation factor in the changing nature of public opinion? E.g., Don Imus.
A: You have to be careful what you mean by reputation. It may be people rating each other for particular attributes, e.g., trustworthiness at eBay. Those are often easily gamed. I’m interested in work being done on understanding how the immune system [the real one] identifiers cheaters.

Q: Do you see a role for government?
A: Government is going to play an important role. When you have a Linden Dollars exchange, [where Second Life money can be brokered for real money], the government will get involved. And when you set up ecommerce sites, identity matters.

Q: [me] Right now, sites solve their identity problems differently, and generally satisfactorily, pretty much. Given that there are risks to having an identity layer, at what point do we say the ad hoc system is broken enough that we want to have such a layer?
A: The layer won’t be uniform. There are risks of abuse, of course, but the identity layer will be an interoperable set of tools for disclosing what users want to disclose.

Q: [chris meyer] Massachusetts no longer uses the SSN for drivers licenses, presumably because it’s insecure to have a single number encode so much…
A: There may be one number that makes multiple sign-ins far more convenient. That will enable innovation. But you can’t get that without a pretty sophisticated layer underneath. Ad hoc-ery will give way, but not necessarily to uniformity.

Q: People worry about uniform identity not in Second Life but in larger systems. E.g., people have proposed used SpeedPass to use to issue tickets for speeding in the tunnel.
A: They’d be persistent, not consistent. It’d be hard to link them. And people will not do business with businesses that betray them.

Q: [chris meyer] Transparency is two sided. When you suggest it, people get worried that they’ll connect up too much information. When does transparency engender trust and when does it not?
A: Transparency may be transparency on not your full identity but on a chosen set of attributes.

Q: Integrated health care records are important for healthcare. If you try to set up a false identity, you could hurt yourself badly from a healthcare perspective.
A: [irving wladawsky-berger] When it comes to health care and children, I believe there will be legislation.
A: [someone else] Yet at Virginia Tech, people didn’t know the killer had been hospitalized because of privacy laws.
A: [clippinger] Right now it’s ham-fisted. It’s either/or. We need it to be more flexible so people can see what they need to see. That’s the new generation of social technology we now need.

[Fascinating, although I remain skeptical about the need for an “identity layer.” And the reception afterward was a great time to talk with some amazing folks, including the Clipmeister himself.]

[Tags: ]

Berkman Center launch party for EIM

The <a href=””>Berkman</a> is holding a launch party for <a href=””>Everything Is Miscellaneous</a> on April 30. I’ll give a talk at 6pm in Pound Hall Room 335, and then there will be a reception at 7pm at the Berkman Center at 23 Everett Street. (Pound Hall is a block away.)

You are invited.

Last night the Center threw a similar affair for John Clippinger’s new book, A Crowd of One. These are really nice events. <a href=””>John’s talk</a> was terrific and engendered a lively discussion, and the wine-and-cheese party at the Center embodies much of what’s best about the Center. So, I hope you’ll come.

Gender Genie confirms I’m a man, pretty much

The Gender Genie at the BookBlog uses an algorithm “developed by Moshe Koppel, Bar-Ilan University in Israel, and Shlomo Argamon, Illinois Institute of Technology, to predict the gender of an author.” You paste in some text (preferably more than 500 words) and it guesses the author’s sex.

The first 1,200 words of the first chapter of Everything Is Miscellaneous gets 2,634 male points and 2,215 female points. The last chapter (5,000 words) scores 8,560 male points and 6,506 female points. My 1,200 words post live-blogging Wendy Seltzer’s talk about copyright scores a similar 1,684 male points and 1,452 female.

The consistency suggests that the Gender Genie is measuring something real, some implicit metadata invisible to me but characteristic of my writing. And if it in fact gets authors’ sex right, it is even stronger evidence — as if we needed it — that so much of our meaning lives in what we don’t see. [Tags: metadata gender ]

Colleges marketing through blogs

The Boston Globe has a good article by Marcella Bombardieri about colleges using students to blog to give prospective students a sense of what life is like there. About 25% of colleges do this. Some pay, some don’t. Some see the blogs before they’re posted, some don’t. All say they have a high tolerance for negative or embarrassing posts.

Wouldn’t a prospective student do better to find students who are just blogging, rather than ones who are sponsored by the school admissions department? On the other hand, have you tried to find, say, MIT blogs at Technorati? Let me give you a hint: The “related tags” listed for “mit” are “technology, und, der, zu, den, das, von, ein and auch.” Who tags anything “zu” or “von,” the equivalent of tagging an English-language post as “to” or “of.”

(Disclosure: I’m on Technorati’s board of advisors.) [Tags: ]

Why I like Twitter

Twitter limits you to 140 characters per posting. You see the postings of people in your social network. The limit encourages frequent postings of small significance.

Twitter thus sounds dumb.

In fact, Twitter is about the intimacy of details. Through it I see small events in the lives of friends about whom I otherwise might only learn the Big Events when we “catch up” after long intervals. [Tags: ]

Podcorps Nation

The Conversations Network (a non-profit from the same folks who bring you IT Conversations) has just launched Podcorps, an all-volunteer team of “stringers” who will record the audio and sometimes the video of public events that matter to people.

Once you register, you can search for events near you that you can sign up to record. Or, if you know of an event you’d like covered, go stick it into the calendar. (The FAQ says that some stringers may want some help covering expenses, but this is intended to be an entirely non-profit enterprise.) The stringers can then publish the media where they want, although Podcorps expects most will post them at and the Internet Archive where they are freely available to anyone.

I hope this takes off. More is better than less. (Disclosure: I’m on the board of the Conversations Network.) [Tags: ]

Networked truth, part 2

I’m still sleep-dprived, but I’ve had a day to think about what I posted yesterday about truth being a property of networks.

It would have been clearer for me to say understanding is a property of networks. Then I wouldn’t have left the impression that I think facts are a matter of majority opinion. Facts are facts. That’s pretty much their essence. Understanding, however, is plural, at least in many domains — less so in the sciences, more so in the humanities.

On the other hand, our age should be embarrassed that we’ve reduced truth to mere facts.

[Tags: ]

Networked truth

It’s three in the morning in the US. I am in the Zurich airport, waiting for my flight to Helsinki. I am high on Dramamine. All of which will help explain why at the moment it seems plausible to me to say: Truth is a property of networks.

I can only guess at what I mean, starting with the obvious: Rather than thinking that truth is a relationship between the propositions we believe and the way the world is, such that the propositions represent the world, in the networked world the truth is argued for and connected via links. For all but the most mundane of truths, the network of conversations gives us more shades, nuances, and reasons to believe. Which leads me to think that if truth isn’t an emergent property of networks, then understanding is.

It is, of course, an unowned, self-contradictory, unsettled truth that is too big to be contained by any individual. It is outside of us and among us. It is gained not by trying to contain it but by traveling through it.

Of course, the fact that I’m traveling at the moment has no effect on my choice of metaphors.

And the fact that I’m dog tired has no effect on my decision to post this instead of letting it melt in the light. [Tags: ]

HIV workers’ social knowledge

Partners in Health is going live with an HIV website that “allows visitors to share insights about the manual and experiences in the field, to ask questions of each other, to answer others’ concerns, and to foster a community of care.” The editor-in-chief, Joia Mukerhjee, says: “This on-line manual is distinctly a work in progress. We intend to keep it that way. ”


(Needless disclosure: My cousin-in-law is an engineer on this project.) [Tags: ]

Code? Nah. Codes? Maybe.

We’ve all got a real problem. On some sites comments are so nasty that they are driving people off the Web. Even if the comments on your own site are always respectful and sweet-natured, the verbal violence on other sites is your problem. Our problem. It’s not as bad as some in the media portray it, but when Kathy Sierra gets over a thousand messages, mainly from women, saying they’ve been stalked or bullied, it’s an issue we can’t ignore.

Unfortunately, there is no easy solution. A Blogger Code of Conduct goes down the wrong path. Codes only can play a role if they’re plural. Very plural.

Lisa Stone puts it all well when she explains why a “one-stop-shopping” code can’t work for all:

Images that are appropriate for a blog devoted to the war in Iraq would never work on a parenting site, for example. They shouldn’t have to play by the same rules. And we all know how I feel about the First Amendment. :)

So, here’s a longer way around to the same point. (More of Lisa here.)

The first and least debatable Blogger Code of Conduct is the body of law that sets limits on what we can say in public. Death threats, libel, and giving away state secrets are all out. But when we try to get more specific than “No death threats! No nuclear secrets!” what do we really all agree about? A Code of Swimming Pool Conduct that says “Swim safely!” is of little use. The only code worth posting poolside says things like “No diving. No swimming without a buddy. ” But what’s the equivalent for blogging, that is, for talking together in public? A single code of conduct would need to drive down into specifics about which bloggers disagree.

Further, no single code could cover all the different ways we want to talk. Conversation shapes itself to its topic, venue, goal and personal relationships. For example, if I’m arguing with a like-minded friend about politics, my social group’s norm allows me to be more interruptive and use more curse words than if I’m talking with an acquaintance from the other side of the fence. Our norms tell us exactly how much bad language we can use with our family, at work, at the sports stadium, and when meeting our future in-laws for the first time. We know how loud we can talk whether it’s sermon time at the synagogue or South of the Border Night at the bar. There is no possibility of coming up with a single code of conduct because there are too many circumstances in which we conduct ourselves. We are left, ultimately, with our judgment.

Behind the drive for a single code of conduct is often the idea that there is one particular type of conversation at the pinnacle of all conversations: The rational discourse in which two people who disagree work toward the truth. Civility is important there. I’m thrilled to be at an institution — the Berkman Center — where those sorts of conversations happen every day. But those are not the only sorts of conversations we should, could, would, will or do have. Some conversations should be raucous. Some should get people red in the face. Some should have us leaving muttering under our breath. Polite, respectful civil conversations are not the only ones worth having because conversation is about much more than the mutual discovery of truth. Conversation is how we’re social, and thus is as rich, ambiguous, implicit, and multipurpose as we ourselves are. Yes, as Tim O’Reilly says, “Free speech is enhanced by civility.” Definitely. We need more civility. But free speech is also enhanced by healthy doses of incivility. In our drive to limit harmful speech, we need to be careful to preserve risky speech.

Of course, that’s assuming a particular model of civility. If, instead, by “civil” one means only that the conversation should be respectful, then I agree that many more conversations need to be civil. But: (a) Respect is not always the highest value of a conversation. (b) What constitutes disrespectful or injurious speech depends upon the target, the speaker and the context (again, ruling out posts that cross the boundaries of the law and our shared sense of decency). (c) A code of conduct that says that, for example, we should be “respectful” will founder on the details of implementation since there are so many norms about what constitutes respectful discourse — sitting in a quiet room with our hands on the table and our heads cocked attentively being only one scenario. Without the implementation details, the code is as useful as the “Swim safely” poster at the pool.

But then we come back to the problem: People violated – threatened, bullied and stalked – by thugs wielding keyboards. When those comments cross the legal boundaries, there may be legal recourse, although usually that’s not practical. It is a problem with no easy or short-term solution. When the comments are posted on the victim’s own site, there are tools for dealing with them, although none works perfectly. A blogger can moderate the comments, perhaps add a reputation system, or even forbid anonymity. A code of conduct is one more tool in the box. Such a code makes explicit the rules already implicitly governing a comment space. As we come across blogs more and more randomly, it often doesn’t hurt to be told that a site won’t tolerate bad language or wants commenters to stay on topic, if those are the local norms. Bloggers can of course state that already — there’s an infinite supply of sentences — and many do, but coming up with standard ways of expressing the rules would encourage their expression.(That’s what I was suggesting 1.5 wks ago, and it’s what I like in Tim’s idea.) Transparency generally is good.Posting rules of the pool that make explicit the existing implicit norms can be a worthwhile tool…although pasting a long list of precise rules can indeed inhibit free swim.

As for encouraging civility: Absolutely. I like civility. Truly. I encourage it on this blog’s comment pages, and I even try to model it on occasion. But I also like a good fart and a high five now and then.

[Tags: ]

« Prev - Next »