When Andrew Cobb doesn't write, he uploads results to Google Analytics or AddThis. He searches online for articles about carmakers, university presidents and philanthropists, and he types in hashtags and email addresses to flag potentially problematic content.
He often finds it—for instance, just by Googling an association executive’s name and seeing that other people have often followed his lead. But that gets a little complicated. Cars, universities and nonprofits are each such a nuanced thing, and so, according to Cobb, it’s impossible to find something valuable every time. This leaves even people like Cobb scrambling to remain informed. The problem is that people can’t, right now, find information they’re looking for fast enough, because much of the information out there doesn’t currently appear in search results.
For a new startup called Wordflow, the answer is machine learning. These algorithms, combined with other data sets, can add new information—whether it’s geographical location, news articles published by journalists, or social media posts—to search results and news aggregators, thereby helping users quickly and easily find content they’re searching for.
Wordflow aims to improve the information users find in search results, news aggregators and search engines—all the while increasing the speed at which the information is supplied to users. By using artificial intelligence, data and machine learning algorithms, Wordflow gets around the problems of over-reaching and curation, as well as even providing users with news alerts for specific topics.
“When people come to search for something or use social media, that’s where the whole conversation is happening,” said cofounder Andrew Cobb, who cofounded Wordflow after being frustrated with the information he and other like-minded users wanted to find online. The 24-year-old Cobb was also interested in building something to increase people’s well-being and make them feel more connected. “So much of our lives are mediated by a finite selection of digital media,” he said.
Cobb studied data science and statistics at the University of Washington and received his MBA in software engineering from Stanford. He now works at his own startup. Wordflow is founded in the same building as GitHub, and Cobb and cofounder and CTO Adam Gray both previously worked at GitHub. Before GitHub, Gray was a design consultant at UXPin and built more than 300 websites for brands including Target, Expedia and Burger King. In fact, cofounder and CTO Adam Gray did his PhD on search and natural language processing, and Cobb previously taught a class on machine learning, in which he and Gray worked as a team.
Cobb hopes Wordflow will help people stay connected in a post-Apple announcement world where people tend to make decisions based on information they consume through search engines like Google. As of this writing, the site had about 200,000 users, but they have about 18 million content combinations they’re working on. These include location, news outlets, cities, universities, charitable organizations, major charities, and other large bodies of information that can be used in search results and news aggregators.
The wordflow site shows what kind of stories people are searching for. However, the site will continue to update—and strengthen—their search algorithm and social-network data as it gets better.
“We want a world where every user gets this fast and relevant information,” Cobb said.