ARCHETYPE AI

0

Unveiling the Future of Physical AI at TEDAI San Francisco

Archetype AI Team
  • October 30, 2024

Last week, we made waves at TEDAI San Francisco — on Day 1, we premiered our feature video, "A Phenomenological AI Foundation Model for Physical Signals," and on Day 2, our co-founder, CEO, and CTO, Ivan Poupyrev, took the stage for a panel discussion on the future of embodied AI. Read on to get an insight into how Archetype AI is addressing some of the biggest real world challenges with Newton.

AI Learning Physics from Sensor Data

Our feature video debuted with a bold vision: addressing some of the world’s toughest challenges – across healthcare, energy, transportation, manufacturing, and infrastructure – with AI solutions. It introduced Newton, demonstrating how this foundation model can learn principles of physics directly from real world sensor data, making it capable of tackling a vast range of complex, physical world challenges – such as optimizing energy use, improving infrastructure, and enhancing manufacturing processes.

The video showcased a milestone in our journey toward developing a physical AI foundation model. Traditional methods involve “teaching” AI about the physical world through structured data and preset laws, like physics equations. But our approach takes a new direction: enabling AI to learn through direct observations gathered from diverse sensors. Without relying on predefined data or pre-existing knowledge, the model decodes sensor input, understands it, and even predicts future physical behavior, allowing it to solve problems across an unprecedented range of applications. This adaptability brings AI closer to an augmented intelligence– one that complements, rather than replicates, human understanding and opens up countless possibilities for real world problem-solving.

Transformative Power of Embodied AI’s

In the panel discussion, alongside Field AI’s Sebastien De Halleux, NVIDIA’s Amit Goel, Principal Venture Partners’ Songyee Yoon and Bloomberg’s Shirin Ghaffary, Ivan shared his perspective on how embodied AI is poised to transform the physical world and the work underway at Archetype AI to bridge the gap between AI and the real world. He dove deep into the philosophy behind Archetype AI and the potential of embodied AI to operate seamlessly in real world settings. He described the development of a “foundation model for the physical world” – an AI capable of making sense of the diverse and complex sensor data that define our environments. From spectrograms to radar, our model is designed to interact with this data intuitively, creating insights that drive physical-world solutions.

During the panel, Ivan shared that one of the critical challenges for embodied AI is ensuring that robots or other other types of equipment and machinery can make decisions independently in unstructured environments, without requiring step-by-step human guidance. This challenge led partially to the development of our foundational model, Newton, which autonomously discovers and analyzes patterns within its surroundings. He likened this development to creating a “ChatGPT for the physical world,” with Newton offering autonomous insights that humans may not have even considered.

Bridging Physical and Digital Worlds

Together, our video premiere and panel discussion highlighted the potential for physical AI to adapt and innovate in real time, providing a glimpse into a future where technology doesn’t just support human capabilities but actively expands our understanding of the physical world.

Stay tuned as we continue to share advancements that expand AI’s impact across industries, making once-unsolvable challenges attainable.

Recommended posts

On October 17, 2024, Infineon and Archetype AI introduced the first-ever foundation model capable of understanding real time sensor data. They presented a demo at OktoberTech™ Silicon Valley, Infineon's annual technology forum. Read on to learn more about the event and our approach to building a large behavior model that can reason about events in the physical world.

October 24, 2024

The renowned tech journalist Steven Levy featured Archetype AI in an exclusive article on WIRED. In his piece, Levy explores how our advanced AI models serve as a crucial translation layer between humans and complex sensors, enabling seamless interactions with houses, cars, factories, and more.

April 8, 2024

Imagine a world where technology could help us make sense of the world's hidden patterns, understand the root causes of problems, and identify solutions.

At Archetype AI, we believe such a world is possible. We’re unveiling a new form of artificial intelligence that takes us a step closer to this reality.

April 5, 2024

Infineon will be the first company to utilize the Newton AI developer platform, offering device makers a combined package of sensor hardware and sensor AI.

October 24, 2023