Bionic Software and Programable People

Last week there was a quote that surfaced in a few posts related to the funding round announced by Codecadmy (in which OATV participated). The quote was from Douglas Rushkoff:

As we move into an increasingly digital reality, we must learn not just how to use programs but how to make them. In the emerging, highly programmed landscape ahead, you will either create the software or you will be the software. It’s really that simple: Program, or be programmed.

The primary thrust of the prior pieces focused on the need to learn to program and the massive competitive advantages that skill set bestows. And I agree with them.

But, what about the latter part of that paradox? 

Back in 2006 Tim and I were introduced to the concept of Bionic Software by You Mon Tsang, founder of Boxxet. Though his company didn’t last, the notion of Bionic Software has been bouncing around the OATV office ever since.

Shortly after our meeting, Tim took a stab at giving the term a little more clarity:

A bionic system is one that combines the biological and mechanical systems to create an enhanced system that is more powerful than either alone…one of the things that distinguishes web applications from PC-era applications is the fact that web applications actually have people inside them, working daily as part of the application. Without the programmers running the crawl at Google, filtering out the spam, and tuning the algorithms, the application stops working. Without the users feeding the spiders by continuously linking to new sites, the crawl turns up nothing new. In a profound way, the users are part of the application. This turns out to be true in one form or another for almost every breakthrough web application.

This interplay of humans and computers augmenting each others actions and amplifying one another’s understanding is at the heart of the Bionic Software meme. It also sets the foundation for what the notion of “programable people” implies.

A program, after all, is simply a set of instructions to accomplish a given task. If we’re fetching and organizing data for computer programs to crunch, essentially for free (or the privilege of being advertised to), how might they program us to perform tasks they simply can’t.

From my vantage point the waters are just beginning to swell for a new wave of applications that will rely on “programable people”. 

The category defining service, Amazon’s Mechanical Turk, is 6 years in beta and still going strong. The tasks on MTurk today look fairly similar to the tasks that were on there at launch- podcast transcription, photo tagging, writing and rewriting text, completing surveys, and the like rewarded by different levels of cash payout for each individual task Turkers are programmed to perform.

But since it’s launch in 2005, the landscape for programable people has changed on several fronts; namely, incentives, inputs and inventories.

  • Incentives- it was recently pointed out to me that the Zynga’s Offer Wall had actually eclipsed Mechanical Turk in terms of total tasks available and completed (I haven’t been able to find anything on the web to verify this but it seems plausible/provocative enough to highlight). If true, this would signal a marked shift in the acceptable, and even sought after, incentives for programable people. That players of Zynga games are willing to complete tasks for more, um, farm animals? seems an important shift. A recent entrant exploring these new incentive structures is Philip Rosedale’s Coffee and Power which creates a “need” and “offer” dynamic in which goods and services can be exchanged for points rather than cash payouts. Philip used his own service network to build his online service by breaking tasks down to their most atomic parts. The result, as highlighted in the NYT, was a set of 1,600 tasks from database setup to bug fixes completed by an interchangeable cast of programable people distributed across the globe. All were paid in C&P’s online currency “C$”. The impact on how we define “work” and “pay” is fodder for a future post.
  • Inputs- after wrapping up my last meeting in NYC yesterday I debated taking the train or grabbing a cab out to JFK. The thought of stuffing my 6’8” frame into a cab or taking a rush hour train wasn’t appealing after two days of timezone changes and late nights. So I fired up Uber, and requested a pickup via their iPhone app. 13 min later Frantzy arrived. The introduction of mobile phones will open up a whole wave of bionic software applications. Not just in mobile requests services like Uber or Cherry, but in other tasks that can only be completed on the phone with all of their native sensors. In the case of Uber, the GPS sensor on my iPhone pinpointed my pickup location. In the case of Google’s autonomous car, these human captured GPS traces provide millions of road miles which are being crunched to get a passenger from point A to point B. Gigwalk is programming people to take iPhone photos of venues. SeeClickFix (an OATV co.) is using mobile photos from residents to program city workers repair priorities. Virtual assistants, like Siri, open the door to a whole new set of tasks for real assistant services, like Fancy Hands, to fulfill. 
  • Inventories- In the last decade retailers like Wal Mart used local area networks to programmatically automated their supply chains and inventory management systems. We’re starting to see the same thing happen with programable people. Except today the inventory is comprised of time, skills and available tasks and distributed broadly over the open web. People with time and skill can search available tasks on services like Skillslate, Zaarly and Task Rabbit. For something more personalized, services like Etsy can connect you directly with artisans that have the time and skill available to complete a more tailored task. But hands are only one class of programable inventory. Kaggle is a new service that is looking to “stock” the brightest minds in math and science by offering a service that allows them to complete and receive cash rewards for cracking some of the worlds most ambitious algorithmic challenges.

Though Mechanical Turk has somewhat languished as a service for Amazon, I can’t help but think these new dynamics around incentives, inputs and inventories will enable the larger promise of bionic software and programable people revealed.

As I mentioned earlier, I believe we’re in the formative stages of this wave. Despite the fatalistic tone of Rushkoff’s quote, I can foresee a future where the impact of programable people isn’t dark at all; rather, it has the potential to unlock a massive amount of unrealized human potential which has been unconnected and untapped until now.

Ultimately, how these services will shape our relationship with the technology and technologist who are doing the programing is an ongoing storyline that I’m equal parts cautiously and curiously following.