Introducing Primer EnginesLearn more
Our world has become increasingly complex and more difficult for us to understand. Yet, as individuals and organizations, we still need to make decisions in this world, many of which will have significant consequences. We can’t do this with human cognition alone. We need to build machines that can help us understand this complexity.
At Primer we’ve been working on this problem for the last two and a half years, working to build the core technologies of a system that can automate the generation of intelligence and allow us to construct the base layer of global knowledge. I’m excited to finally share some of what we’ve been working on and the progress we’ve made.
We started Primer by focusing on a core problem: training machines to read and write at the level of a specialized human analyst. To solve this, the machines must look at different types of data across multiple languages – English, Russian, Chinese – from a diverse set of sources. They must be able to combine them together to understand what’s being said. Finally, they must communicate these insights by generating human-readable language and charts, without any human involvement.
We’ve raised a total of $13.5 million in two oversubscribed rounds from a world class group of investors with extensive expertise in AI, deep tech and intelligence. Our Series A was led by Data Collective (DCVC) with additional investments from Lux Capital, Amplify Partners, Bloomberg Beta, and Avalon Ventures. Our team has grown to 36 people and we’ve been heads down building out the core machine learning infrastructure. We are a talented team of engineers and other specialists pushing the boundaries across a range of important problems in machine learning. Our mission is to accelerate our understanding of the world.
Building machines that can read and write is a difficult task and one that will require continuous improvements to current machine learning models. We’ve pushed technology forward in key areas such as language generation, multi-document event detection, and unsupervised entity-to-entity relationship extraction, as well as advanced fundamental lower level technologies like novel information detection and Chinese character segmentation.
With this technology we have built a system that can identify events in the world, geolocate them and describe them in human-readable format. The system is then able to predict the significance of what’s being observed. This allows us to construct an interconnected graph of events that updates itself as the world changes, generating a base layer model of the world where insights can be communicated to the user in natural language. This model can be enhanced by connecting it to proprietary data sets and deploying it onto customer cloud platforms and hardware infrastructure.
Though still evolving, this technology is already opening up a range of exciting possibilities. For example, Primer can take 40 million unclassified documents in English and Russian, read them and identify all terrorism-related events. Then our platform can automatically generate an interactive map that detects and highlights key parts of the narrative as it unfolds across two different languages.
This core problem of understanding the world is shared by the intelligence community, finance sector, and global corporations. These sectors have invested heavily into the collection of massive amounts of data over the past five years, but do not have enough human analysts to look at it all. This creates a delta between the amount of information we should be looking at and the amount of information we are actually looking at. This gap won’t close by throwing more people at the problem. We need machines that can understand this data for us.
Our platform is used today by some of the most important companies and government organizations in the world to augment and amplify the skills of human analysts. Primer is helping them to understand their world. This commercial traction allows us to continue investing in our long-term technology roadmap and further advance our machine learning capabilities
As the world becomes more complex, it is increasingly important for us to quickly and cost-effectively jump outside our filter bubbles and connect the dots between seemingly disparate ideas. If, for example, an analyst wants to investigate a weak link between their analysis and that of the Saudi sanctions on Qatar, they no longer have to spend their weekend doing this. They can simply pick up the Primer-generated analysis and use it as a starting point. Primer’s technology enables analysts to do this by massively reducing the cost of curiosity.
With technology that can read and write, you have the flexibility to generate custom insights in any format or level of detail. If you’re a subject matter expert, Primer can tell you a detailed story that takes your knowledge into account. If you’re new to a subject, it can generate an introduction to get you up to speed quickly. If you have an interest in a particular angle on the story, or a geographic lens that you want to zoom in on, the insight can be customized for you.
Imagine the possibilities if you had one thousand analysts working for you, all day, every single day. What questions would you ask, what kinds of briefings would you have them prepare? What would you be curious about? What would you learn about today?
As a physicist I was trained to look for the underlying truth in a complex world, to identify from first principles the structures that give rise to everything we see. It’s this philosophy that influences how we think about artificial intelligence at Primer.
Primer was founded on the idea that computers see the world differently than us and that they should be able to teach us about the world that we live in. That artificial intelligence should be used to help us understand our world.
We believe the most intelligent system in the world will be created by combining the best of human intelligence with the best of machine intelligence. This is our start.
_Sean Gourley, CEO and Founder of Primer.ai