Entries in collective intelligence (5)


Collective intelligence at the Motley Fool to evaluate stocks

The Motley Fool, a financial services company and online investment community has a new service, CAPS, to find and evaluate the next big stocks. The site automatically screens for promising small-cap stocks and then publishes them on the site:

We'll tap the collective intelligence of our CAPS members to see whether these companies present real opportunities -- or whether the numbers fail to tell the true story.

CAPS members can then give the stock a 1 to 5 star rating based on their personal evaluations, presumably with the goal of sinking obvious-yet-less-promising companies and raising hidden gems. Here is the interface:

This service seems a bit superfluous. Since these are all publicly traded companies, I would think that the stock market is already pricing them to reflect their future value [and since real money is on the line, I would expect the real market to be more accurate and reliable than a 5 tier starring system]. But I will take the Hansonian* point of view that traders in this marketplace enjoy affiliating with the Motley Fool and trust that segment of the population to be wiser than the broader investing population.

*A philosophy, espoused by our friend Robin Hanson, that explains much of human behavior in terms of signaling and affiliation. See here.


Tap the Collective Video is Up

Our professional video from Tap the Collective on September 2 is now available: http://tappingthecollective.com/


Tap the Collective in DC Wrap Up

This week we co-hosted Tap the Collective, an event to bring together thought leaders in collective intelligence and forward-looking individuals from across government and technology firms. Through briefings and discussion, we shared lessons learned in implementing programs and platforms to leverage the diverse and distributed knowledge in government organizations.

Left to Right: Adam Siegel from Inkling Markets, Melody Hildebrandt from Transcapitalist, Ryan Hahn from World Bank, Robin Hanson from Consensus Point, Shyam Sankar from Palantir Technologies, Emma Antunes from NASA, Don Burke from Central Intelligence Agency (CIA) and David Resseguie from SensorpediaAdam Siegel of Inkling Markets and I kicked off the event before we turned it over to our six panelists who each spoke for 7 minutes about their initiatives. Here are the highlights:

Don Burke, Intellipedia doyen, opened by sharing his insights implementing a CI platform -- a Wikipedia for the intelligence community -- in a large bureaucracy. He noted that while individuals are finding value from CI, the organizational structure does not yet align incentives sufficiently to encourage wide spread adoption.  Among his ideas for how to advance CI in large organizations : "measure reputation not output" and eliminate the use of documents: "if you can't send it as a URL, don't send it".

Ryan Hahn of the World Bank, then discussed the use of prediction markets at the Bank. He cited the challenge of seeking to make predictions on future occurrences when the organization making those evaluations actually has the power to make the prediction a reality if the results were leaked. He learned this lesson the hard way, asking this question of the market : "Will X country default in the next year?" which nearly got the entire project shut down by the corporate communications team. While the prediction market experiment has seen notable accuracy, the challenge remains to convince senior leadership to act upon its results.

Emma Antunes, Spacebook manager at NASA, shared her lessons learned bringing social media to the enterprise with NASA's internal Facebook-like social network site, Spacebook. Spacebook was created to advance innovation and discovery; Emma described how innovation happens at the intersection of disciplines and the site helps people make those connections. Emma also shared some gems about NASA's culture, such as "no one ever retires at NASA, they just die" and also its "employee emeritus" program where those who do retire come back as volunteer employees.

David Resseguie of Sensorpedia at Oak Ridge National Laboratory, discussed the lab's "Wikipedia for sensors", a mashup of map and sensor data to be used by local law enforcement and the Department of Homeland Security. He spoke of the "citizen as a sensor" citing the first shot to come out of the Hudson river airplane landing, taken by a bystander on a passing boat on his iPhone, and how much rich data could come out of that simple, grainy photo, such as the number of people escaping -- especially the contrast between the first and second classes. Sensorpedia seeks to take advantage of the fact that everyone of us is now a sensor -- we all have cell phones in our pockets -- while figuring out how to aggregate this data in a way that is meaningful and actionable.

Shyam Sankar, Palantir Technologies, profiled the disruptive collaborative information analysis platform, Palantir, that came out of efforts by the team from Paypal and engineers from Stanford. Paypal stumbled upon CI while dealing with the existential risk from hackers, as Shyam states, "Paypal wasn't about sending money, it was about sending money without being defrauded by the Russian mob." When fraud was at its height of 150% of revenue, Paypal recognized that humans were much better than machines at catching the fraud, but they didn't have the power to scale sufficiently. Paypal eliminated the fraud through intelligence augmentation, leveraging the collective abilities of the organization through a robust technology platform. Palantir is now applying that same approach to counterterrorism, cyber threat analysis, and open government (open to the public through analyzethe.us). Here is a great article/demo about their capabilities.

Robin Hanson of George Mason University, presented last, exploring how organizations can make decisions based upon unverifiable or controversial data. He described prediction markets as a great way to evaluate controversial and political topics and to incentivize people to "shut up when they don't know the answer". Best Buy's success in crushing Circuit City, for example, may be attributed to their markets asking employees about the direction of certain technologies, leading them to be better stocked with Blu-Ray players when that was decided upon as the HD video player of choice.

The event wrapped up with a lively 40 minute Q&A session with questions ranging from how to get more raw data out, to creating the incentives to enable culture change, to how to fight forces who lobby against pursuing these more innovative paths. We were really thrilled to see the level of engagement by the audience and the diverse group of attendees, coming from government agencies, professional industry organizations, startup technology firms, and nongovernmental organizations.

The full Flickr stream is available here and I'll update the post later this weekend when the video is finalized.  Thank you again to Boalt Interactive and Palantir Technologies for sponsoring the event.


Intellipedia takes collective intelligence to the intelligence community

Part four of the Tap the Collective profiles...

The open and collaborative information analysis platform that debuted to a wall of skepticism in 2006 may now, in 2009, be demonstrating a legitimate alternative way to think about analysis in the intelligence community. Intellipedia now boats an average of more than 15,000 edits per day with 900,000 pages and 100,000 user accounts, a remarkable level of adoption in only a few years in an environment known for being resistance to big change.

Yet Don Burke, Intellipedia Doyen, finds that Intellipedia is still largely used by early adopters rather than institutionalized across the community. It seems, though, that its proponents are strategizing well to make this social software eventually a central part of analysts activities rather than an afterthought.

One important factor is that all posts are attributed through official accounts, meaning that quality control is high since people cannot hide behind anonymity, and also perhaps more importantly, analysts can actually build a reputation on the site. This additional incentive is important to adoption and active participation -- as we have seen with prediction markets and the leaderboard system. If young analysts can become known for publishing high quality content perhaps this effort could gain the attention of senior officials, much like Inkling Markets has rewarded successful traders with the opportunity to meet company executives. The wiki format seems to create a flatter mechanism for generating high quality content regardless of the level or organization of the contributor, unlike for example, National Intelligence Estimates which follow a strict formal process from originator through commenters. Chris Rasmussen, knowledge manager at NGA and one of the original proponents of Intellipedia believes that they can go one step further and make social software contributions part an employee's performance and compensation plan.

Another important decision in the roll out of the site is that the developers chose to use nearly the exact same interface and functionality of Wikipedia. A problem seem all too often in government is a customized technology solution introduced to allegedly make life easier for analysts that is so unintuitive as to make it not worth the hassle. I have technology-savvy friends in the Department of Defense who groan when a new application is announced simply because their expectations are so low. The result is that they go out of their way to avoid using it. Wikipedia's developers and proponents were very wise to stick to an interface that works and that the community's analysts are already using at home and would be comfortable interacting with immediately.

We look forward to hearing Don Burke, Intellipedia doyen, share his first hand account of implementing this collaborative information analysis platform on Wednesday at Tap the Collective. Mr. Burke is a finalist for the 2009 Service to America Medals, sponsored by the Partnership for Public Service, for his work in this area.


Tap the Collective coming September 2 to Washington, DC

We're excited to announce an event we're co-hosting with Inkling Markets in Washington, DC on September 2nd: "Tap the Collective". We've assembled a great lineup of technologists, thought leaders, and do'ers to discuss how collective intelligence initiatives can be used to overcome organizational barriers in government.

More information and registration information are here: http://tapthecollective.eventbrite.com

Our event is generously being sponsored by:

Palantir: http://www.palantirtech.com/
Boalt Interactive: http://www.boalt.com