As large enterprises work to integrate AI into their platforms, one major hurdle has emerged: Generative AI systems need memory and constant updates to their training data in order to be practically useful. This area, known as Live AI, is becoming a focal point for several startups, including Pathway, which has recently raised $10 million in a Seed round to develop AI systems that think and learn in real-time, much like humans.
The round was led by TQ Ventures, with participation from Kadmos, Innovo, Market One Capital, Id4, and several angel investors. Lukasz Kaiser, co-author of Transformers and a key researcher behind OpenAI’s GPT-01, is also among the investors in Pathway.
What Pathway Offers:
Pathway is building infrastructure components designed to power Live AI systems. By enabling the integration of structured and unstructured data into AI models, Pathway’s platform ensures that enterprise AI systems make decisions based on the most up-to-date knowledge. Their technology promises to transform how AI platforms operate, offering real-time learning and adaptation. Notable customers include NATO and La Poste, the French postal service.
Challenges in Current AI Systems:
Zuzanna Stamirowska, Co-Founder and CEO of Pathway, explained that traditional deep learning models and Large Language Models (LLMs) rely on training data that is static, which limits their ability to adapt. She compared these AI models to “a very smart intern on their first day” who is given a book but is unable to memorize or use the information. Pathway’s goal is to fix this limitation by enabling real-time data feeds into AI systems, allowing them to continually update and evolve.
The Pathway Vision:
Stamirowska outlined how Pathway’s solution allows developers to build a pipeline where live data can be fed into the AI models during the prompting stage, ensuring that the AI is always working with current information. This capability is key for businesses that require up-to-the-minute insights.
The startup’s founding team includes Adrian Kosowski (CSO) and Jan Chorowski (CTO), both of whom previously worked with Geoff Hinton, the “Godfather of AI” and recent Nobel Prize winner in physics. Stamirowska herself is recognized for her forecasting model for complex networks in the maritime trade, which was published by the Academy of Sciences of the US.
The Journey to Launching Pathway:
The company’s inception was sparked by an idea that Stamirowska had during a trip to a theoretical computer science conference in Chicago. After a discussion with a friend, she realized the potential for a new AI solution and quickly began reaching out to her network to bring the idea to life. She fondly recalls the moment of inspiration, saying, “I still remember the taste of the coffee at that moment.”
With the $10 million funding, Pathway is poised to advance its mission of creating AI systems that continuously learn and adapt in real time, meeting the growing demand for dynamic, responsive AI solutions.