Decision, decisions: What do I fancy for lunch today? Do I want a girlfriend? What should I do with my life? Too many questions and not enough time for me to resolve them. So what if there were a Website that knew me well enough to automatically answer all my questions? Wouldn’t that be the next big Internet thing, the definitive answer to all our questions?
Divining intent has always been the Holy Grail of academics, doctors, marketers, and snake-oil salesmen. At the beginning of the 20th century, Sigmund Freud founded the school of psychoanalysis, a science of intentionality based on interpretation of dreams. And today, 100 years later, a new generation of quacks has founded a school of crowd-sourced intentionality.
The father of this digital school of intentionality might be Eric Schmidt, the CEO of search engine Google. Back in 2006, when asked by a Financial Times reporter where he wanted Google to be in five years, he answered that he imagined the search engine would become so intelligent that it would know not only what its user wanted to do that afternoon, but also what career they wanted to pursue.
While Schmidt’s hubristic dream is nowhere close to being realized by his Google search engine, other Silicon Valley entrepreneurs and investors have continued to chase this happy dream by using all the latest algorithmically personalized crowd-sourced tools of the Web 2.0 revolution. The latest entrant into the race to automate human intentionality is Hunch, a year-old company that is building what it calls a crowd-sourced "recommendation engine," which knows us well enough to answer all our questions. Hunch claims to build a personalized algorithm that supposedly is able to learn our intentionality.
Hunch represents la crème de la creme of Silicon Valley. Founded by the mother of Flickr 's Caterina Fake, with illustrious board members like Wikipedia co-founder Jimmy Wales, Hunch has just raised a $12 million round of venture capital investment led by blue chip fund Khosla Ventures.
The truth is that Hunch, which is a bad mash-up of Google and Wikipedia, doesn’t work on any level. The crowd-sourced content on Hunch is mostly banal, sometimes absurdly so. For example, in response to the question, "How should I express my condolences?", one user-generated answer suggests "Attend the deceased funeral," while another advises us to "Send flowers."
Duh. Is this the best that Silicon Valley can offer us? Is this the sum intelligence of Fake, Khosla, and Wales? Do I really need a crowd-sourced Website to explain to me that the best way to express my condolences after a death is to attend the funeral of the deceased?
And the algorithm doesn’t work either. When I signed up for Hunch, it asked me a series of dumbly automated questions about my consumer preferences that barely skim the surface of the tangled web of desires that feed my intentionality. Human beings like you or me don't operate exclusively from silicon-based logic (well, at least I don't). We can’t be figured out by answering a few random algorithmic questions of an unthinking machine.
The problem with Hunch is that it treats me as a crass consumer rather than as a complex human being. Web 2.0 algorithms might succeed at building search engines that are, at best, 50 percent accurate. As such, Hunch represents the intellectual bankruptcy of the Silicon Valley revolution that has attempted to replace real human expertise and knowledge with bad algorithms and even worse user-generated content.
When it comes to determining what I want, crowd-sourced algorithms are about as accurate (and scientific) as Freudian dreams.
— Andrew Keen, Silicon Valley author, broadcaster, and entrepreneur, can be reached on Twitter at @ajkeen.