When the Q&A site Quora launched in 2010, it seemed like there was little space for a major new source of information on the Internet. Social networks and news sites took care of current affairs, from personal lives up to worldwide events; Wikipedia rounded up reference information from online and offline sources.
Three years later, Quora has made a good case that the Web was missing out on a valuable knowledge source. By encouraging people to write down their personal experiences, expertise, and advice, the site has compiled some arresting content of the kind that marks the most memorable of face-to-face conversations. Examples range from an astronaut trainer explaining prelaunch rituals to an inside account of Apple’s culture of secrecy.
Quora’s cofounder and CEO, former Facebook chief technology officer Adam D’Angelo, told MIT Technology Review’s Tom Simonite about his mission to make the Web more interesting by having people tease out information from one another and using technology to curate and filter it.
What is Quora for?
Our mission is to share and grow the world’s knowledge. There have been a lot of algorithmic attempts to organize the world’s knowledge: there’s Google, there’s IBM’s Watson. But there’s all this information in people’s heads that’s not written down on the Internet. We want to get to the state where if there’s someone in the world that knows something, we can tell you that. After the Asiana crash, for example, we had all these pilots giving their expert opinions. Or sometimes you want to hear about someone’s direct experience, when you have a question like “How should you raise a dog?”
Doesn’t letting people ask questions and write answers lead to a lot of bad content, like what exists on Yahoo Answers?
Other Q&A sites all end up in this state where there’s some good information but there’s a ton of low-quality answers. That’s why we are more of a tech company than a lot of Internet companies. We can predict which content is good and only show you that. Over time the average quality contributed may go down, but as long as the amount of good content that we show goes up, we’re doing fine. When people say Quora is great now but how will it scale, that’s the answer. You will only see a slice of it.
What does that technology do?
We use machine learning to understand what you are interested in, based on signals like which topics you read or follow on the site and answers you voted on. People give us access to their accounts on other services and so we can see what they do there. We also have this map of who are the experts in different areas, so we can route questions to the people that will give the best answers.
Lots of websites already rely heavily on recommendation systems, like Amazon and Netflix.
Those systems are different. There are not that many movies that have ever been made and recommending products requires less accuracy. We have tens of millions of pieces of content, and most of our answers are not that interesting to most people.
If you suggest a random product or movie, it’s not going to be that bad, but if you show someone information about B-52 bombers and they don’t care about it, it’s very bad. We have to use our data to predict what people are interested in very finely. Maybe a particular user likes answers about physics but not quantum physics.
When someone writes a new answer on Quora, we also have to figure out very quickly if it’s good. We already know something about the person who wrote it, and we know their previous answers and who liked them and who didn’t. We also very quickly show new answers to people. We can learn within minutes whether it is likely to be a blockbuster answer and if it meets a certain quality threshold. If it does, then we surface it on the site and notify people following that topic or question.
Can that information be really valuable, rather than just nice to know?
Yes. If you wanted to know how to run a startup really well, you used to have to do a kind of apprenticeship and work in a lot of different places. Now you can read about it from experts on Quora. I think we’ve done really well at collecting knowledge that you would otherwise have to do time in sort of indentured servitude to access.
How do you make money doing this?
It’s likely that we will run advertising that is finely targeted to users. But we’re going to invest in other ideas as well. We think about it, but we’re not building anything. For now we’re focused on growth. That’s what the investors want.
This new data poisoning tool lets artists fight back against generative AI
The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.
Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist
An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.
Data analytics reveal real business value
Sophisticated analytics tools mine insights from data, optimizing operational processes across the enterprise.
The Biggest Questions: What is death?
New neuroscience is challenging our understanding of the dying process—bringing opportunities for the living.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.