The top competitor we identified in our application — Aardvark — was acquired by Google for $50 million in February 2010, then shut down a year later. Quora launched two months after our rejection and grew to 300 million monthly users. Most of the other competitors we listed — Yahoo Answers, Google Knol, Mosio — are dead. And the problem we were trying to solve — “ask a question, get a satisfying answer from someone who actually knows” — is now handled by large language models. We wanted humans to be the neurons of a collective mind. Turns out, the neurons would be artificial.
Back in 2007, Antonio Orlando came to me with an idea for a semantic question-and-answer platform. We called it Bioniqa — bionic Q&A. Not a search engine — search engines crawl existing pages and rank them. This would be a system where people generate the content themselves. You ask a question, the system routes it to the right person, and over time it learns who knows what, where, and in which language.
We called users “Neurons” and the collective knowledge “the Mind.” Each user is specialized — a Neuron in a particular domain, language, and geography. When connected, information flows between them and the system adapts. The software would borrow concepts from nature: diversity, adaptation, neuroplasticity. Geography and language would be first-class attributes, not afterthoughts.
The core insight was that search engines are terrible at contextual, hyperlocal questions. “What’s the mineral water with the lowest residue currently on sale in Bari?” No amount of PageRank helps with that. You need an actual person who lives in Bari and buys mineral water. Our system would find that person for you.
The idea floated for a couple of years. Then in early 2009 I pulled in Michele Frettoli from Milan, and the three of us decided to take a shot: we applied to Y Combinator for the Summer 2009 batch.
We put together a specification document, filled out the application, and identified our competitors — Aardvark at the top (closest to our vision, but taking “bad” directions in our view), plus Yedda, Mosio, Yahoo Answers, MetaFilter, TxtEagle, Google’s Knol, and Wolfram|Alpha as a potential complement rather than competitor.
Part of the application was a video — YC wants to see the founders. Here’s mine, all 24 seconds of it:
Today, pg announced that all applicants have been emailed. Ours was a rejection:
We’re sorry to say we couldn’t accept your proposal for funding. Please don’t take it personally. Once again we got a record number of applications this cycle, and since there’s a limit on the number of interviews we can do, we had to turn away a lot of genuinely promising groups.
Another reason you shouldn’t take this personally is that we know we make lots of mistakes. It’s alarming how often the last group to make it over the threshold for interviews ends up being one that we fund. That means there are surely other good groups that fall just below the threshold and that we miss even interviewing.
We’re trying to get better at this, but it’s practically certain that groups we rejected will go on to create successful startups. If you do, we’d appreciate it if you’d send us an email telling us about it; we want to learn from our mistakes.
Y Combinator Staff
Fair enough. Record number of applications, limited interview slots, tough decisions. No hard feelings.
The idea is still good, though. Someone will build it eventually.