After showing off his own Mendeley Profile, Luke spoke to the Mendeley team and guests about how, as a journalist writing for publications such as The Guardian, Fast Company and Wired he was keenly aware of the pervasiveness of technology:
“If you look at any period in history, the imagery and metaphors are drawn from popular science, and today there is no science more popular than computer science. My interest in technology comes from popular culture. If you want to understand popular culture you really need to engage with technology and the questions it poses, which are really key to understanding how the world works and our relationship it, as well as our relationships with each other, and issues with our own identity.”
As a filmmaker himself, he explained how he came across the famous quote from screenwriter William Goldman (who produced screenplays for All the President’s Men, The Princess Bride, and Butch Cassidy and the Sundance Kid, amongst many others) which stated that when it came to the entertainment industry, “Nobody knows a damn thing”. Luke then set out to discover whether this was indeed true, or whether technology could actually help us to, for example, accurately predict what films would succeed at the box office.
In his book, The Formula, Dormehl talks about how a company called Epagogix claims to be able to do just that by analysing scripts using over 30 million unique scoring combinations. Interestingly, it not only churns out a number, but is also able to make creative suggestions based on the data, adjusting scripts to make them more successful and profitable.
“This represents a vision of a future where machine logic can be embedded in the creative process”
But these processes are certainly not straightforward, even in fields such as academic publishing or law, which would seem, to an outsider, to be less subjective and therefore more suitable for automation.
He then outlined an interesting recent experiment, which illustrated how even turning a fairly binary traffic law into an algorithm that issued speeding tickets to infringers accordingly, could be a lot more challenging than you would think. Given the same datasets, two groups of scientists produced algorithms that issued vastly different numbers of tickets, which highlights the many potential difficulties facing the Google Driverless Car project, for example.
Luke concluded his talk by showing how nothing is sacred as far as algorithms go, not even love. He explained how it was even possible to create a virtual girlfriend though a relationship simulator called Kari.
Like with any research project worth its salt, writing The Formula left him with more questions than answers, and as you can imagine, the crowd listening to the talk followed up with quite a few insightful points of their own in the Q&A session that followed.
Do watch both videos on the Mendeley YouTube Channel and let us know what you think! We’re also busy arranging another talk on the 18th July, so be sure to watch this space and follow Talks@Mendeley on Twitter for more details!