I’ve long been a fan of Tim’s, even before I had the chance to write and publish a now-quite-out-of-date book for O’Reilly media. Tim did not disappoint, bringing a wealth of stats and facts to share from his vast experiences in the tech world. He started with an excellent quote from Edwin Schlossberg (“The skill of writing is to create a context in which other people can think”), expounding on how his publishing career has resulted in him addressing the top minds in healthcare. His talk quickly covered topics as diverse as global conciousness (Danny Hillis described it as: “that thing responsible for deciding that pots containing decaf coffee should be orange”), architecture for participation (small modular pieces joined by open protocols using simple standard data formats that are extensible by design), and the wonder that is Wikipedia (especially the ability to animate changes to a page over time).
One of his core topics was crowd sourcing and the ability for huge data sets to triumph over even the best artificial intelligence design. I loved his account of Garry Kasparov vs. the World, a 1999 online chess game in which more than 50,000 people from 75 countries collectively decided on the next moves for the black side of the board. Garry eventually won after 62 moves and four months, after which he said “It is the greatest game in the history of chess. The sheer number of ideas, the complexity, and the contribution it has made to chess make it the most important game ever played.” Tim focused on the tenth move of the game, an unexpected suggestion from Irina Krush that took the game in a completely new direction. Irina isn’t a player on par with Garry, but she did spend a lot of time figuring out that particular move, which is the advantage to a crowd of narrow experts over a single, broad expert.
Data and Open Data are one of Tim’s passions. He talked about the difference between the DARPA Grand Challenge winner Stanley and Google’s Autonomous Car project is the volume of data available to the AI. Stanley was one of the first true autonomous vehicles, but his knowledge base was limited to what the team was able to build into it. Google spent hundreds of thousands of hours studying real driver behavior in real situations, and their cars are working from models built on that huge amount of data. That’s the same approach Google has taken to advertising, using their AdWords program to solve the (in?)famous Wannamaker problem: “Half the money I spend on advertising is wasted; the trouble is I don’t know which half.” Data can have huge leverage in healthcare if we built proper systems to be able to share it. Right now all of that data lives in silos with incompatible data formats – Tim’s call to arms is to free the data and solve health problems through shared knowledge. He’s right about this: the growth in data science and analytics job postings on LinkedIn speaks for itself.
That chart comes from O’Reilly Media’s excellent (and free!) ebook Building Data Science Teams. You can download the book from that link. A number of the ideas sketched above are covered in more detail in that blog post too, so the whole thing is highly recommended.