Liquid AI Builds Personal AI Around Device-Level Context, COO Says

Li made the remarks in an interview with theCUBE Research's Dave Vellante and John Furrier, broadcast on theCUBE, SiliconANGLE Media's livestreaming studio. He said the company's aim is to move AI closer to the user and to run it on the hardware surrounding them.

"The vision we have is that we should bring AI … closer to the user," Li said. "So we focus on building these AIs to run on the devices all around us."

The form factors best suited to capturing that signal, he said, are phones, wearables, watches, PCs and cars. Model builders have traditionally assumed elastic cloud capacity, but the edge offers fixed hardware alongside a far richer view of the user, which Li described as the natural home for personal AI.

Sitting between the models, the agents and the hardware is Liquid Context, a component optimized for Snapdragon processors. It uses device signals to build an understanding of who the user is and what they are trying to accomplish, according to Li.

Managing that context falls partly to agent harnesses, the software that turns a model into a functioning agent. Li said keeping user context in an ever-growing text file creates problems on devices with limited resources, and that Liquid AI uses its own models to decide which information to retain and how to compress it.

"The problem with devices is that you have fixed compute," he said. "You have to fit within the zero-sum compute. That means a lot of the assumptions around how harnesses today are built no longer hold at the edge."

Liquid AI's work includes a collaboration with Mercedes-Benz Group AG to bring on-device AI to its cars. Li said the next step is keeping deployed agents aligned with user expectations over time, which the company intends to address through systems that improve and personalize on their own.

"From here, we want to build these systems and these agents to be able to self-heal and improve and personalize on their own over time," he said. "We're building observability loops and continuous improvement loops that will improve both the model and the harness over time through natural usage."

TheCUBE is a paid media partner for the Fully Connected event. Neither CoreWeave, the sponsor of theCUBE's event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.