The Future of Intelligence, Part 1

About six months after 9/11, I came across a book by Gregory Treverton, who served as the Vice Chair of the National Intelligence Council under Bill Clinton. The book, Reshaping National Intelligence for the Age of Information, was published shortly before 9/11, and its insights into the state of national intelligence were both revealing and prescient. It’s a remarkable book, and it got me thinking more deeply about the incredible cultural and organizational challenges of national intelligence. Not coincidentally, around the same time, I was starting to synthesize my ideas on collaboration, a process that resulted in my founding of Blue Oxen Associates.    (L7T)

Things are starting to come full circle. Over the past year, I’ve found myself engaged in conversation with a number of people in the intelligence community, and it culminated in a two day workshop with the CIA this past week. It’s been somewhat of a surreal experience, given that I’ve spent much of the past four years working side-by-side with progressive and Open Source activists, many of whom consider the government an antagonist at best, an enemy at worst. Moreover, my one previous brush with government work — a project with the FAA — left me with a less than favorable view of how our federal agencies work. The culture there is stifling, especially in comparison to the Bay Area. There is a stated desire to learn and to improve, but there is very little real commitment. Those who actually want to do something are trapped under a blanket of repressive indifference, and those who manage to do something anyway are usually completely marginalized by their superiors and even their peers.    (L7U)

All that said, the fundamental challenges regarding intelligence are dear to my heart, and I find myself paying a bit more attention when these conversations and opportunities present themselves. I firmly believe that deep knowledge about collaboration is spread across a number of domains, and the only way to acquire this knowledge is to engage with each of those different communities. This especially holds true with intelligence, where the knowledge product itself is sensemaking and actionable knowledge.    (L7V)

I am also a patriot. That word has attained somewhat of a negative connotation over the past five years, which is not necessarily a bad thing, because it has forced us to deeply reexamine our values. I’ve gone through this process myself, and I’ve walked away even more sure of my feelings. I’d like to make both this country and the world a better place. Those two goals are not orthogonal.    (L7W)

Last year, I met Darniet Jennings, an intelligence researcher, at WikiSym. We’ve had a number of interesting conversations since, and he participated in our first “Tools for Catalyzing Collaboration” workshop. About six months ago, he referred me to a paper written by Calvin Andrus entitled, “The Wiki and the Blog: Toward a Complex Adaptive Intelligence Community.” Andrus, who is with the CIA, wrote this paper last year, and it has since become the de facto reference on the role of Social Software in intelligence.    (L7X)

I’ve found the majority of these kinds of whitepapers shallow and uninteresting. Andrus’s paper is anything but. Rather than offer some simplistic portrayal of these tools while repeating the same tiresome anecdotes and misconceptions over why they’re useful, Andrus frames the conversation in terms of systems theory. He cites Ludwig von Bertalanffy and Jane Jacobs, and he praises decentralization, localization, and emergence. The depth of his paper comes from this framing, which establishes the correct higher-level goals and philosophy behind these tools and which surfaces the intelligence community’s real challenges. The paper has flaws, but they are minor.    (L7Y)

The fact that such a paper exists and that it has been embraced by the intelligence community makes me hopeful, but that hope is tempered by Treverton. Andrus writes about the importance of empowering local, bottoms-up action, and he cites Tip O’Neill‘s famous maxim, “All politics is local politics.” For those who might perceive of the intelligence community as being overly centralized, this seems to be a refreshing viewpoint.    (L7Z)

However, Treverton suggests that this view is not foreign to the intelligence community at all. In fact, he writes that “intelligence analysts tend toward the long view and to take the world as a given.” Treverton then cites the very same O’Neill quote, writing, “Because they [intelligence analysts] are so immersed in the local, they are by profession believers in the adage attributed to former U.S. Congressman Tip O’Neill that ‘all politics is local politics'” (181). In contrast, policy makers tend to care less about the long view. Transforming national intelligence is not enough. We need to transform the relationship between intelligence and policy.    (L80)

Online Community Summit: Friday’s Sessions

Notable talks and comments at last Friday’s sessions at the Online Community Summit:    (2F4)

  • Soren Kaplan described iCohere’s work with World Vision, a billion dollar nonprofit with 20,000 employees worldwide. There was an online collaborative process leading up to a conference, using iCohere‘s software.    (2F5)
  • Dave DeForest discussed the online communities at the Motley Fool. There’s a 40:1 read-write ratio on their bulletin boards. The Fool’s strategy for monetizing the communities was to get the readers to pay, and to comp the writers. The comping process is transparent in that participants know that some people are being comped, but the actual process for comping people is not concrete. When the Fool went to a pay model, it had a 90 percent attrition rate. 73 percent of its community participants are also likely to perform another transaction on the Fool.    (2F6)
  • Mark Williams discussed Apple’s support forums. Right now, he’s handling everything — management, development, etc. His managers tell him that the objective is to reduce phone volume, but he sees the two audiences as separate. Robert Labatt noted that Apple does a great job of converging threads on its support forums.    (2F7)
  • Anne McKay posed the following theory, citing last year’s The Atlantic Monthly article on introversion: You need extroverts for a successful online community. I would argue the opposite, although I have no numbers to back me up. We had an interesting discussion about this very topic in the Collaboration Collaboratory about a year ago. It would be interesting to do Myers-Briggs analysis on online communities in a future case study.    (2F8)
  • Steve De Mello told stories of “bad behavior” on some of ezboard’s online communities, and noted that hosts should only deal with black-and-white issues. The best pressure is peer pressure. Gail Williams agreed with Steve’s assessments, and noted that while paid communities helped filter out trolls, they didn’t eliminate them entirely.    (2F9)
  • During a breakout session on metrics, Gail suggested that our biggest problem is understanding and serving lurkers. (See my previous entries on lurkers.)    (2FA)
  • Tom Coates informed folks on the IRC Backchannel about Wiki proxy, a cool little proxy that automatically links terms to Wikipedia.    (2FB)

Finally, Reid Hoffman and Ross Mayfield gave a quick walkthrough of Social Software. I was amazed at the blank stares in the room during this talk. These folks seemed to have some awareness of Social Software, but certainly not a deep understanding. The previous day, Marc Smith said that threaded forums aren’t going away. I agree, but the common wisdom in the group seemed to be that threaded forums tend to be the end-all and be-all of online communities. I strongly disagree with this assessment. Dave asked what Wikis offered that threaded forums do not. That question missed the point: It doesn’t have to be one over the other. (Ross enjoys needling folks by claiming that email and threaded forums are dead, but I don’t think he actually believes it.) My response to Dave was that there’s great potential for integrating Wikis and threaded forums (as I and others do with blogs).    (2FC)

A Walk with Howard Rheingold: Collaboration as Craft

I had the great pleasure of walking and talking with Howard Rheingold last Thursday. Howard lives in Mill Valley, a few blocks away from some of the many trailheads leading up to Mount Tamalpais. We had exchanged emails a few times and had met briefly after his talk at Stanford in October. I had invited him to coffee, and he suggested a brief hike instead, which I gladly accepted.    (MG)

Winter is one of the best times to go hiking in the Bay Area. We started walking around 4pm as the sun was beginning to set. The sky was a deep blue with a solitary streak of clouds overhead, and the air was cool and crisp. We walked about a mile to the top of a hill, where a rock formation seemed to form a natural bench around the crest. Looking north, we could see the peak of Mt. Tam. To our west were neighboring hills and the Pacific Ocean. To our east was a beautiful view of Mill Valley, where the city lights were beginning to come on. All of this served as a vivid reminder that I had, as usual, forgotten to bring my camera.    (MH)

Nevertheless, I was there to talk, and talk we did. One topic that came up — and a key reason for wanting to talk to Howard in the first place — was my desire to see the emergence of collaboration and community-building as a discipline, a widely acknowledged craft.    (MI)

People sometimes ask me what I know about collaboration that other people don’t. My response: Nothing. The reality is the reverse. There are many, many people in the world today who know significantly more than I or anyone else associated with Blue Oxen Associates about collaboration.    (MJ)

The problem is that this knowledge is scattered around the globe in isolated pockets. These folks all speak different languages — not just English versus French versus Korean, but also geekspeak versus Wall Street versus academia. Even when they know about each other, they can’t always talk to each other.    (MK)

Even worse, there is no group memory. Narrow the field to online communities. A lot of folks in the field have heard of Howard. How many people know what he’s accomplished beyond his excellent books? How many people have heard of the WELL? How many people know who founded the WELL? Going further back, how many people have heard of PLATO? Most importantly, how many people can cite lessons learned from the WELL or PLATO? Online communities have been around for decades. How many people can trace the lessons learned from these different communities over time?    (ML)

Howard told me that one fellow — perhaps one of the most knowledgable people in the field of online communities, with the credentials to match, and someone whom I’ve admired from afar — is working in retail right now to make ends meet. There’s no shame in working in retail, especially when times are tough like they are right now. Nevertheless, this strikes me as the worst kind of cosmic joke. Venture capitalists are spending millions of dollars on fast-talking entrepreneurs selling Social Software, trying to figure out how to make this stuff work (and profitable). There’s someone out there with decades of experience to share, someone who can undoubtedly help make these efforts successful. And yet, he’s currently working in a strip mall, addressing the needs of last-minute Christmas shoppers.    (MM)

Who’s at fault? You can say that this person — for all his skills — is poor at marketing himself. You can say that companies are short-sighted, and that they don’t understand what they need or how this person can help. There’s probably some truth to both of these statements. But, it’s still a travesty. This guy should be a hero to everyone claiming to be in the business of collaboration.    (MN)

That’s the crux of the matter. This is a field that is in desperate need of self-awareness. If we in the business truly want to improve, we need to be aware of our history and our heroes.    (MO)

Ross Mayfield on Social Software, Social Networks

Last Friday, I finally got a chance to meet Ross Mayfield, who was speaking at the November meeting of the Bay Area Futurist Salon, held at SAP Labs in Palo Alto. Ross is CEO and founder of Socialtext, a company that is selling social software to companies. Some people have compared Blue Oxen Associates to Socialtext, which is apt in some ways, but is not quite right, as I’ll discuss below. Christine Peterson (one of our advisors) had said great things about Ross and Peter Kaminski (Socialtext‘s CTO), and Ross and I had exchanged some pleasant e-mails. The meeting confirmed Chris’s judgement. Not only is Ross a good guy, he had some interesting things to say about Social Software and Social Networks.    (D8)

Social Software    (D9)

I mostly agreed with what Ross had to say about social software, and I especially liked his overall philosophy about what makes a good tool. I did have a few nitpicks, though. Ross defined Social Software as follows:    (DA)

  • Software that supports group communication    (DB)
  • Software that adapts to its environment, rather than requiring its environment to adapt to it    (DC)

I asked him to clarify the latter statement, and he explained that most collaborative software tries to enforce too much structure. These tools force users to figure out how to fit the data into the tool, whereas the tools should fit the data. In this vein, Ross spoke highly of Wikis and blogs, and also of human filtering (such as Google’s technique of measuring backlinks) as a way of organizing information.    (DD)

I strongly agree with Ross’s philosophy, although I don’t like how he worded it in his slide. His statement is equivalent to the first part of Doug Engelbart‘s philosophy of coevolution of tools and processes; however, it leaves out the second part, which is equally important.    (DE)

Doug says that tools ought to augment human processes. However, as we learn more about the tool, we also must evolve the processes to adapt to the tool. An example that Doug often cites is the bicycle. Riding a bike is not intuitive, but it offers significant performance advantages over a tricycle. (To illustrate this point, Doug likes to show this picture.) “User-friendly” tools can be useful, but they should not be the end-goal.    (DF)

One “semistructured” tool that Ross does not like is e-mail. In predicting its demise, he cited several statistics regarding the amount of time required to deal with e-mail, especially but not exclusively due to spam. Ross introduced a cute term that I had not previously heard: “occupational spam” — e-mails from “legitimate” sources that have no bearing on your life or work whatsoever. This often manifests itself as being cc’d on threads of little or no relevance. Ross claimed that 30 percent of our inboxes consist of occupational spam. I should have asked for a citation on that stat, because it’s an interesting one.    (DG)

One of his main arguments against e-mail as a collaborative tool was that it encourages discursive discourse, and that it’s hard to make any sense of the sum product. I agree entirely with this argument, but would use it as an argument for how to use e-mail effectively rather than against e-mail entirely. Internally at Blue Oxen, our e-mail and Wiki usage complement each other quite nicely. We kick off the discussion using e-mail, and we synthesize the resulting thoughts into the Wiki, with links back to archived e-mail.    (DH)

Social Networks    (DI)

Ross’s presentation on Social Networks was outstanding. He’s obviously thought deeply about these issues, and he’s collected quite a bit of interesting research. He first described his and Valdis Kreb’s Blogmap Project, which mapped relationships between members of Ryze‘s Blog Tribe.    (DJ)

He then discussed the power law of blog space, which states that the distribution of links to a blog is inversely proportional to the blog’s rank. In other words, the number of links to the most popular blog is exponentially greater than links to the next-most popular blog.    (DK)

Ross then proposed three types of collaboration: the publishing model (one-to-many, where the power law applies), the individual’s social network, and small teams. He cited the primatologist Robin Dunbar’s work to suggest that social networks generally numbered about 150, and said that people’s interactions with their social networks generally followed a normal distribution. (Interestingly, he showed his RSS aggregator, which had 146 subscriptions.) Finally, he suggested that the optimal size of teams is about 10, and that the number of interactions between team members tends to be about equal.    (DL)

Socialtext Versus Blue Oxen    (DM)

Socialtext and Blue Oxen are similar in that we both are interested in collaborative tools, and that we both share similar philosophies about tools. Socialtext sells these tools as enterprise applications, however, whereas Blue Oxen is focused on understanding how best to use and improve these tools and on disseminating that understanding widely.    (DN)

In that vein, we both have embarked on similar projects, but with different goals. Ross talked about Socialtext‘s experiences with Social Software supporting face-to-face events, and those tools were available during his talk. We did a similar project with the PlaNetwork 2003 Conference, for which we set up a Wiki, RSS aggregators, and an IRC channel.    (DO)

We host collaboratories, and plan on charging for membership to those collaboratories once our infrastructure is up to snuff. However, we’re not an ASP, although we will make (and already have made) ASP-like services available to communities. Our infrastructure is meant to be cutting-edge and rough around the edges. It’s a way for people to experience first-hand what it’s like to use those tools, an opportunity to learn (and share) how best to use them, and a platform for coevolution. Despite our intentions, we’ve already discovered that even with promises of little/no “official” support, people enjoy using our tools for their real-world collaborative needs. As members, they’ll have free access to these tools for their groups. When they start needing better support and enterprise-level features, we will happily point them towards Socialtext and similar companies.    (DP)