'People Will Work on Their Mobiles in Africa, We Just Don't Know How Yet'

As I mentioned in a recent post, one of the most striking aspects of the African Internet economy is that we don't know what the mobile Internet will look like when it is proliferated in Africa, much the same way we didn't know that banking would like M-Pesa.

My best glimpse yet of the future came yesterday at Nathan Eagle's talk at the Kennedy School. To be honest, I was somewhat embarrassed that I hadn't come across some of the brilliant work he is doing.

He started off the talk by saying "people are going to do work on their mobile phones in Africa, we just don't know what it is yet." However, three of his projects give some hint of what is to come.

Entrepreneurial Research and Programming on Mobiles
A program to support computer scientists and other engineers in their efforts to create mobile applications for the developing world context. They host an SMS Bootcamp, mobile phone programming for entrepreneurs, and mobile web apps. The program is now in 10 countries with 15 local computer science professors and lecturers.

These workshops have spawned a dizzying array of companies, in the fields of MDSS, mobile medicine, SIM-based application development, reality mining [w/ Christopher Waranga at Univ. of Nairobi], SMS bloodbank, Boonanet [commodity pricing], SMS gateways, air time regulator, stolen car alerts, business directories, pre-paid electricity, weather forecasts, MoKoSo [craigslist] and crush lists.

txteagle
A start up based on three premises: (i) there are 1.5 billion poor literate mobile users in the developing world with time on their hands; (ii) corporations would benefit from 'crowd-sourcing' millions of tiny tasks that will always be done better by people than by computers; (iii) mobile phone networks have tons of underutilized capacity.

Txteagle "enables these tasks to be completed via text message by ordinary people around the globe." Say, for example, that Mozilla needs a Luganda language version of Firefox. They would ask people who speak Luganda to translate some of the key words. The machine parcable nature of text responses means that algorithms can identify talent and weed out those gaming the system.

The project is in a very early stage of identifying which tasks are best answered by SMS, and at this point 'they are throwing ideas against the wall, and seeing which ones stick.'

Reality Mining
Nathan's more academic work is on behavioral inference of complex systems in the developing world. At the Santa Fe Institute this summer, he built a super computer that can map, for example, where all of Rwanda's international phone calls in go in a single day.

This could be useful for understanding human patterns and affinities in outlier events like earthquakes, financial networks, urban planning, housing management, and movement dynamics. The most fascinating measurement of the economic influence of mobiles is a study that compared graphs of fish prices in three different regions of India before and after mobile base stations where installed. Before the base stations, the prices were erratic and subject to the whims of the larger buyers. After base stations, the data flat lined, showing that the ability to communicate has a profound effect on the price of tradeable goods in the developing world.