Why we're here
The problem we're solving
For billions of people across the world "the internet" is a very different place. Echo chambers and interest-based algorithms create closed-loop systems that lock people into corners of the biased applications of their choice. Within the confines of these artificial walls, discussions are fluid, narratives emerge, and slang evolves, but they're restricted to the corners they were born in. Sometimes trends have enough escape velocity to "break-through"; we call this going viral. But without virality, and without infiltrating every corner of the internet, how can we answer the following questions?
What are people talking about?
What do people think?
How can I find the conversations I'm interested in?
And as generative AI becomes more accessible,
What narratives are being pushed?
Who's pushing them?
We view the social web as a fractal-like tree of subcultures. It's heavily fragmented across platforms, protocols, lexicons, audiences, and content.
Our goal is to traverse this infinitely expanding tree, aggregate the information, use advanced AI/ML methods to understand it, and create an index that powers search, discovery, and analytics.
Our story
Siftree was created by Kyle DeSana and Ehren Marschall in Chicago, Illinois. Kyle was working on a way to rank comments by relevance, to help people find comments they're interested in faster (like Google PageRank for annotations) and quantify battling perspectives. Ehren was exploring better ways to use machine learning to cluster conversations across platforms and see what the world was talking about. We found similarities in what we were building, believed there was a way to combine the two, and decided to work together on it.