TEDxSoMa

Why Machines Need People

Damon Horowitz, Aardvark

From the program: Damon Horowitz is a leading thinker at the intersection of technology and the humanities. Horowitz is co-founder and CTO of Aardvark, the popular social search engine, where he oversees product development and research strategy. Prior to Aardvark, Horowitz built several companies around applications of intelligent language processing. He co-founded Perspecta (acquired by Excite), was lead architect for Novation Biosciences (acquired by Agilent), and co-founded NewsDB (now Daylife). Horowitz teaches courses in computer science, cognitive science, and philosophy at many institutions, including Stanford, NYU, University of Pennsylvania, and San Quentin. He has spoken at conferences ranging from Web2.0 to AAAI, and his work has been featured in media ranging from the New York Times to Discovery Channel to TechCrunch. Horowitz earned a B.A. in Computer Science from Columbia, a M.S. from the MIT Media Lab, and a Ph.D. in Philosophy and Linguistics from Stanford.

When Damon went to college, he became very religious, and technology was his God. The technology he believed in was artificial intelligence. He was an AI fanatic. Here’s how you can become an AI fanatic: learn some AI tools, and then get challenged to do something that seems really difficult (handwriting definition, summarizing a news story, etc). Then go away and do it. It’s an incredibly gratifying experience, and you start to think, “Wow, if I can solve all this, I can solve anything!”

So he spent the first decade of his career working on that at MIT with people like Noam Chomsky, and he developed some technology that got sold to Excite, and he should have been really happy. But he wasn’t satisfied, because no matter how good they got, the technology never knew what it was doing. It only processes data. He can’t brainstorm with his computer. While computers can process news stories, they can’t understand the meaning of the words. They’re very clever but they lack common sense.

So around the turn of the millenium, he went back to school to study philosophy, which struck many people as stupid. The reason he did it is because he wanted to understand the limitations of AI, and to do that you have to get out of the race. To be a philosopher is to think hard about thought itself, about what makes our language and life meaningful, and what is meaning itself. He spent years studying this.

While technologies can take us further than ever before, it is always humanity that creates meaning. Words have meaning because we speak them, not because they exist. Even the most advanced AI algorithms today come down to pattern matching. They don’t know why we react differently to a story about earthquakes in Haiti than we do to a story about vacations in Haiti.

Consider web search: you have some question in your head, and what you have to do is turn that question into keywords, find a browser, search through results, and extract information. It’s as though we’ve forgotten that a question is an invitation to a human experience. The amount of information in peoples’ heads dwarfs the amount of information in computers. So they started Aardvark — you send it a question, and Aardvark’s job is to find a person who can answer that question. They use AI not to replace people but to help connect people.

The amazing thing is that it works. They launched only a few months ago, but already over 90% of the questions sent get answered, and usually within a few minutes. Here’s an example of the type of questions they get: “I accidentally dropped my MacBook pro 15″ in a large bin of chili. It was in there for about 2 minutes or so. How can I salvage it?” http://vark.com/t/dd59e3 Luckily, someone replied within moments with instructions on how to save the chili.

He also told us about someone who asked how she could arrange a special screening of Julie and Julia for her friend who had been fighting breast cancer for 9 years, and had just moved into a hospice. How can you search for that question with keywords?

How valuable it can be to step outside of your technological world view and consider different disciplines. Philosophy isn’t a very popular subject with students today, but it should be.

Technology cannot solve all of our problems for us; the task of thinking is still ours.