DARPA is one of the most innovative government funding agencies around. Back in the late 19060’s – when they were still called ARPA – they backed ARPANET, which ultimately became the internet that we know.

More recently, DARPA sponsored the Red Balloon Challenge. This involved discovering the location of ten large (weather-scale) red balloons, scattered across the country. Maybe they were expecting satellite imagery, or powerful telescopes, or image recognition coupled to location information on Facebook photos. But what actually won the prize was a variant of good old-fashioned chain letters. A group at MIT led by my friend Manuel Cebrian devised a brilliant scheme where participants were incentivized to recruit team members by a recursive prize structure. They quickly built the largest team in the competition – and largest set of human eyes – and in only 8 hours won the prize.

DARPA is now at it again with the Shredder-Challenge, which is to reconstruct documents that have been shredded into small pieces. This is like a puzzle where all the pieces have random shapes. And Manuel is at it again, this time from his new home at UCSD. This time they are coupling the recursive prize with a computer vision algorithm that assists players in putting pieces together. Check it out!

Steve Jobs testing photo booth in 2005 – image by Mike Matas, available at Facebook

The world just lost a true genius. Rest in Peace Steve.

Health 2.0 conference fills up the ballroom at the SF Hilton

I just spent three days at Health 2.0 San Francisco earlier this week. Here’s some reflections:

• This year’s conference attracted 1,500 attendants, 50% more than in 2010. That’s a lot of growth. This is a reflection of two things; (1) great work by Matthew Holt, Indu Subaiya and the rest of the Health 2.0 team in growing their organization and (2) a big increase in the number of people interested in entering the field. Hard to say which one was more important, but it’s clear both played a big role. What was not clear is whether new business models will match all the enthusiasm. My take is that there will be some big winners (diabetes management?), but still very unclear who, when or how. The emphasis on shiny new things also makes it difficult to learn what’s already working.

What do 1,500 Health 2.0 conference badges look like? Like this

• Health 2.0 means different things to different people. There seemed to be (at least) three distinct subgroups at the event: patient empowerment advocates, early-stage consumer-facing (B2C) entrepreneurs, and more established enterprise (B2B) companies. They all have using technology for health in common, but sometimes not much else, given differences in priorities and capabilities. This makes for great breadth at the price of not much depth.

• A lot of companies are trying to build platforms, and they are going about this by announcing open APIs. Presumably this is to attract indie developers. As a hobbyist developer, I like the potential, but I would find it hard to tie myself to developing for a single platform with uncertain user-base and business models. It seems like a very fragmented market. The iOS and Android app stores work because they provide access to hundreds of millions of potential users, and there are easy-to-implement payment and advertising revenue sources. Lacking both of these, these platforms might struggle to attract indie developers that make up the bulk of the world’s apps. Maybe all the hackathons and app contests will help…maybe. It’s clear that managing developer relations is something that an ever increasing number of companies will have to become competent in, or outsource.

• I played around with a Basis watch, which is launching soon and will measure heart-rate, temperature, motion, and your aura’s minute-by-minute interaction with the universal soul-stream (j/k). It wasn’t functional, but it did look sexy. Will I pay 200 bucks for it? Probably. Will it prevent me from becoming obese? I’d like to think I have a good handle on that already. I think it will be more of a nerd-cred device for me than a health device, but I could be wrong.

• Related; an increasing number of apps and devices are generating personal fitness data…the success of apps like RunKeeper and devices like Fitbit is spurring a lot of new entrants. Is fitness data part of a medical record, and if so, does it need to be integrated and/or secured accordingly? Does the fact that I exercise regularly make me a better candidate for an employer who will be responsible for my health insurance? Or to a hot new brand seeking well-toned customers to best exemplify their target lifestyle? Or perhaps a dating site for the proven-fit? Interesting business and ethical questions here. I’m working on a related project at UCSD, stay tuned for more.

• Buzz-words of the year: gamification, API, health data

Fantastic video on media remixing by Kirby Ferguson focused on one of my favorite bands, Led Zeppelin.

Part of being a great innovator is knowing what is already working, and combining elements in new ways instead of always trying to reinvent the wheel. Google didn’t invent search, Apple didn’t invent the smartphone, Facebook didn’t invent social networking. They just made them better.

Everything is a Remix Part 1 from Kirby Ferguson on Vimeo.

Last week I spent a couple days with BJ Fogg, taking part in his Persuasion Boot Camp.

It was an amazing experience. BJ has a knack for communicating simple, concise, and useful information for anyone shaping behavior with technology.

The world of innovation is changing. With developer tools like Ruby on Rails and distribution channels like Facebook and the App Store, the biggest challenge in building a consumer-facing product is often not how to build, but rather what to build. Thus effective design can be more important than engineering resources. The proof is the tiny teams that built amazing products like Gmail, Reddit, and Instagram: all less than 5 engineers. Wow!

BJ teaches product design grounded in his experience as a behavioral psychologist, understanding what makes humans tick. If you are trained as an engineer (like me), it can be really difficult to put yourself in the shoes of a normal user. Here’s how:

B = mat: Behavior equals motivation, ability and a trigger (at the same time)
Don’t over-focus on motivation…people already want to lose weight, save money, and get work done. Instead make things easy to do, decreasing the ability required. Realize that your users will need a hot trigger – something in their path, that clearly directs them – or else the desired behavior won’t happen.

(reprinted with permission from BJ Fogg)

Baby Steps Don’t try to move mountains at the beginning. Instead, try out and validate simple behaviors, quickly. Don’t be afraid of seeming silly…for example, get your mom to understand how to use your app before you worry about what the perfect logo is. This is actually the normal engineering approach of breaking a big problem into manageable pieces, except the elements you are building with – human behaviors interacting with machines – are inherently fuzzier than material or mathematical properties. Because of that, more validation is required.

Iterate Build on successes. Understanding failures is really hard, because so many things can go wrong. Instead, focus on what’s working. This includes understanding the competition.

These take-aways are simple…like consumer facing products should be.

There are many ways to get answers online. I am a huge fan of Google and use it every day to find information that is already published and indexed. But today [EDITOR: actually June 30th, 2011, when this was published on my Posterous] I had a great “Aha” moment when I found an answer through Twitter instead, because it connected me to the best person in the world to answer a very specific question. Here’s how it went down:

My question:
How many physician bloggers are there in the United States?

Post #1

Response #1:

I already know that Doctor V, who follows me, is a physician blogger - one of the best, in fact. The point of this tweet was to alert our mutual followers that this question was worth reading, and that we would both be paying attention. This tweet did not go out to all of Doctor V’s 6,154 followers, because by starting the tweet with an @ mention, he limited it’s distribution to our mutual followers.

Response #2:

Barbara’s retweet is amazing, because it just alerted all 3,240 of her followers to my question. At least the ones that happen to see this tweet.

TwitterBot chimes in:

Whoa! Twitter is apparently scanning every tweet, detecting questions, and suggesting users who would be well-suited to answer them. DanielCass works at Kaiser Permanente, a very large healthcare provider organization here in California. Well done, twitterbot.

Responses #3-#6:

Some responses trickle in. Not exactly to the question I asked, but it’s a start. Everyone who is @mentioned in RayGoldberg’s and MatthewBrowning’s tweets is now alerted in a special way to my request - their “@mentions” inbox will include this tweet. That’s how Twitter lets conversations expand across follower streams. The @mention is just high-friction enough – there is a cost to publishing it, of having your followers know who you are trying to reach out to – that it is not abused by most users. Twitter’s interaction design is absolutely sensational.

Response #7

Grandrounds is an aggregator of medical blogs. Getting warmer.

Response #8

DrVes is a physician, professor, and expert in online presence of physicians. He is an adviser to NEJM, and his website has a lot of useful resources for physicians. And he gives a me a rough answer to my question.

Response #9

The interaction between DrVes and scanman leads to several pieces of useful information. Call it instant-peer review. Did I mention DrVes is in Chicago, scanman is in India, and I’m in San Francisco?

Post #2:

I clarify my question, and reach out to Danielle based on Twitter Bot’s suggestion.

Response #10

Another link in the chain

Response #11

And another link

Response #12

mR is Manhattan Research, which specializes in health care market research. Bingo.

The Lowdown:

20 hours of asynchronous interactions eleven people involved, 10 of which I have never met before, from San Francisco, Chicago, New York City, Washington DC, New Haven, and a city in India called Salem. one twitter bot
I get my answer Twitter shows off how powerful a communication tool it can be

Twitter versus Google:

What happens if I google my query instead? The first result is the exact same question on Quora, which I asked a few weeks ago, unfortunately still answered. The advantage of Quora would be that the answer would become permanent and well-organized, instead of vanishing in the void of tweets past. But right now the community I need is on Twitter.

Starting in the mid 70’s until the mid 80’s, a group of nerdy folks met regularly in Silicon Valley to talk about personal projects with hard-to-use, clunky, and unfriendly machines called computers. They were called the Homebrew Computer Club, and while their interests seemed crazy at the time, they were laying the groundwork for the personal computer revolution that has changed all of our lives. I’m always wondering, in 30 years time, what will have been the equivalent group of the present time?

There’s a few candidates, like the open source community hanging out at Github, or the DIY folks reading magazines like Make and posting projects on Kickstarter. But being obsessed with building tools to measure and improve human health, I’m most fascinated by the movement called Quantified Self.

QS is all about constantly measuring indicators of health, like body movement, weight, heart rate, and meals. People have done this informally for a long time (e.g. food diaries), but the willpower and discipline required turned off all but the most determined. This is now changing. In the past few years, three huge trends have made self-tracking easy and fun for endlessly curious people like me. First, the costs of the technology required to automate these measurements has plummeted, because popular consumer devices like the Nintendo Wii have commoditized sensors like accelerometers. Second, the social web and app economy have driven development of open APIs, and demonstrated how awesome sharing digital data can be for consumers and companies alike. Third, mobile devices like the iPhone combine multiple sensors with the computational power, connectivity, and interactivity to make data accessible and meaningful.

I’m just beginning to pay attention to this community, but I’m already sure it will spawn cool things in the future. Like personal computers, these products will first be used by techy early adopters and people with specialized needs, like elite athletes, but will eventually have a much broader impact. I recently bought a Fitbit, a pedometer that syncs wirelessly with my laptop, and I’m sharing my daily step tallies through a Google spreadsheet. Thanks to Ernesto Ramirez for posting instructions on using a script by John McLaughlin to automate exports of data from Fitbit to a Google Spreadsheet.

After several months of hacking Rails, and building an Android app for fun, I’ve recently found myself doing more paperwork and less coding. I miss using terminal and getting my hands at least a little dirty with code.

To fix this, I decided to use Octopress to power marciovm.com, where I will put more professional posts, and to relegate my Posterous to more personal things like quotes and my pictures. Octopress is a blogging platform for hackers, like Jekyll which it is based on. I’m hosting this site as a GitHub page, which is free and reliable.