Pathway, as Cohere and Writer explore the ‘Live AI” arena, joins with a $10M funding round
As big companies struggle to integrate AI into their platforms, they are facing a challenge: Generative AI must have a memory and the training data it uses needs to be updated constantly for any real use. The area has been renamed ‘Live AI,’ and several startups including Cohere are involved. Pathway has raised $10 million in a Seed round for the development of live AI systems, which, according to the company, will think and learn as humans would. The round was led TQ Ventures with participation by Kadmos Innovo Market One Capital Id4 and Angel investors. Another investor in Pathway includes Lukasz Kaiser, the co-author of Transformers and a key researcher behind GPT o1 from OpenAI.
Pathway’s offering includes what it calls ‘infrastructure components’ that power live AI systems, feeding on structured and unstructured data, meaning that enterprise AI platforms can make decisions on up-to-date knowledge. NATO and La Poste are among the customers so far.
Zuzanna Stamairowska, CEO and Co-Founder of Pathway told TechCrunch on a phone call that deep learning assistants and LLM assistants work by taking the training data, and then training models. The question is how do you deal with knowledge and memory? A LLM is currently acting like an intern who was given a book on his first day at work. They can’t memorize it. It’s also not real-time, but static.”
To fix this, Stamirowska said Pathway allows developers to “build a pipeline that feeds live data into AI systems.” Right now we do it during the prompting stage of when you build LLM applications or Gen AI applications.”
Stamirowska — who is moving to Menlo Park, California — has assembled an impressive, highly technical team to achieve the startup’s goals. Her co-founders include CSO Adrian Kosowski, CTO Jan Chorowski who worked previously with Geoff Hinton, the recent Nobel Prize winner in physics and “Godfather” of AI. Stamirowska is the author of the Academy of Sciences of the US’s state-of the-art forecasting models for a maritime network. “The company began with an idea I had on a sunny morning in Chicago,” said Stamirowska. “I was accompanying a colleague to a conference on theoretical computer science. We had a disagreement and I decided to start my own company. I then took my laptop out and began writing to my network to find ways to make this happen. “I still remember the taste the coffee had at that time.” She said that in enterprise deals, we often encounter Palantir for AI transformation tenders. However, they are less product-oriented than us. “Whereas in enterprise deals, we often encounter Palantir for AI transformation tenders, although they are less product-oriented than we are.”
Commenting in a statement, Schuster Tanger, Co-Managing Partner and Co-founder at TQ Ventures, said: “Zuzanna and the team at Pathway possess bleeding-edge insights and expertise in one of the most exciting fields in modern business… Last and hardly least, the response from the developer community has been powerful.”
