I must admit it took me back a bit when Herb Kim announced that this was the fifteenth Thinking Digital. As I sat in the audience I tried to think of how many I had been too. It isn’t fifteen but it can’t be a long way off. I did notice, however, that I recognised fewer of the audience than in previous years. I still got to say hello to a lot of friends and colleagues but this is a good thing and shows that there continues to be a new intake to our fascinating sector.
Of course, as always, the event was amazing and gave me so many things to think about, so much so that I wish I could have my time over again. Even then, I would be only able to choose a small number of ideas to pursue.
I’m going to start, as the conference did, with the talk by Amanda Prorok, an Autonomous Transport Technologist, Professor of Collective Intelligence and Robotics in the Department of Computer Science and Technology at the University of Cambridge and a Fellow of Pembroke College.
I’m not sure if I got the terminology correct but she spoke of cooperative intelligence, how animals come together to create intelligent systems through cooperation and collaboration. I noticed such a system in action during the breaks where, despite the large numbers of people milling around, there were very few collisions. It was as if everyone knew the rules.
She described how driving is a 100% computable problem and is working on robotic systems to improve traffic flows and reduce vehicle costs, primarily by moving the sensors required away from the vehicle and into shared infrastructure. This would provide the intelligence needed to navigate the vehicle at a fraction of the cost and would make it easier to distribute this information across a much wider audience.
She also told us that early stage robotics systems had a lot of crashes yet that these were valuable learning experiences. The best way to have fewer crashes is to learn from having lots of crashes. This isn’t a very practical option in real life and so this is a really useful application for virtual reality. Let autonomous vehicles learn how to avoid crashes in an alternative reality.
The thing I am going to take away from her talk though is that developing cooperative systems to manage traffic flow, in which each vehicle takes others into consideration, improved flow by 35%. If ever there was an argument for collaboration and cooperation over raw competition then this is it. I don’t know how such an approach could be applied to commerce but a 35% improvement in efficiency is well worth the consideration.
Thoughts from a Guerrilla Worker 