Recent advancements in deep learning have been remarkable, with innovative systems and models such as GPT-3 demonstrating improved capabilities in understanding human language. This progress is enabling developers to integrate these technologies into a wider range of applications, as evidenced by the significant improvements we’re seeing in speech-to-text applications and translation software.

But what further capabilities can be unlocked by this evolving AI infrastructure? Hebbia is dedicated to discovering the answer.

Hebbia is currently operating as a startup and a product studio – essentially a testing ground for AI concepts – established by George Sivulka (a Stanford PhD candidate currently taking a break) along with a team of three other AI researchers and engineers from Stanford. Utilizing the latest deep learning techniques and models, the team is striving to expand the potential of knowledge graphs, semantic analysis, and AI to enhance human productivity.

Sivulka’s motivation for concentrating on this area stems from observing the challenges faced by his peers in the professional world. “Many of my colleagues… everyone enters these office-based jobs where they spend their days reading large amounts of information,” Sivulka explained. “Individuals become financial analysts and examine SEC filings for specific details, or pursue legal careers and perform similar tasks… [They are] overwhelmed by extensive texts, by a constant stream of information that is difficult to interpret.”

(I can certainly relate).

His team’s objective is to significantly improve human productivity by developing search, analysis, and summarization tools designed to help individuals effectively manage their own personal knowledge base. “The core concept is that Hebbia is creating tools for thought that improve how people work. These tools directly manage the information you encounter daily,” Sivulka stated.

Recognizing the ambitious nature of this goal, they began with a specific project. Their initial product, which initially sparked my interest in their vision, is a Chrome plugin that has been in private testing and is now being released to a broader audience (please note: it is currently not listed in the Chrome Store). This plugin enhances Chrome’s search functionality, moving beyond simple text matching to actually understand the intent behind your query and identify relevant answers within a webpage. A demonstration of the plugin can be found on TechCrunch:

For example, when using Ctrl-F on a Wikipedia page, you could ask “Where did this person live?” and the plugin will recognize that you are seeking location information and highlight relevant text on the page. As it is still in early development, the AI’s performance may vary. However, as Hebbia refines its models and improves its understanding of text, the goal is to fundamentally transform browser search and provide a substantial boost to productivity.

Sivulka has demonstrated exceptional talent from a young age. He worked at NASA during his teenage years, completed his bachelor’s degree at Stanford in two and a half years, finished his master’s degree shortly after, and began a PhD program before dedicating his focus to Hebbia.

Hebbia’s vision has already garnered attention from investors in its early stages. Floodgate’s Ann Miura-Ko led a $1.1 million pre-seed funding round, with participation from Naval Ravikant, Peter Thiel, Kevin Hartz, Michael Fertik, and Cory Levy.

Sivulka emphasizes that their Ctrl-F product is the company’s primary focus at this time, serving as an entry point to the broader possibilities offered by knowledge graphs and personal productivity tools. “This represents one of the final challenges for computers,” Sivulka said, pointing out that computation has already revolutionized numerous fields by digitizing data and simplifying processing. With Ctrl-F, “this is a fundamental technology, [we are] only beginning to explore its potential.”