Imbue, an AI research lab, just secured a whopping $200 million in funding, pushing its value past $1 billion. This puts them in the ranks of well-funded AI startups.
Their mission? To develop AI systems that can think and code effectively, essentially helping us achieve bigger goals in the real world.
Initially focused on training AI in 3D environments, Imbue has pivoted towards building models that are “internally useful.”
These models can code, much like GitHub Copilot and Amazon CodeWhisperer, but what sets them apart is their robust reasoning capability. Imbue believes that robust reasoning is key to creating effective AI agents.
Imbue’s approach involves training very large models with over 100 billion parameters, optimized for reasoning tasks.
They’re utilizing a compute cluster co-designed by Nvidia, equipped with 10,000 GPUs. In addition to this, Imbue is also investing in building its own AI and machine learning tools for debugging and visual interfaces.
Their end goal? To develop systems that allow anyone to build their own robust custom AI agents, making the power of AI accessible to all.
This recent funding injection will be crucial in propelling their mission forward.