Even more unexpectedly, Newton RF also performed strongly on non-RF tasks like vibration, audio, EEG, and seismic data. That means we should be able to expand Newton into the one model that understands any physical phenomenon, reads any sensor, and works for any use case. One world model that encodes the entire physical world. Full research: archetypeai.io/blog/newton-r…

Newton now understands radio frequency (RF) data! We just released results comparing a new Newton RF model against specialized RF models. Newton RF was trained on a large volume of RF data that looks nothing like other physical measurements, and we improved the model architecture to capture the spatial signal characteristics inherent to radio frequency. The results: Newton RF beat all five specialized RF models in 18 of 20 radio tasks, including localization, channel modeling, and blind demodulation. This is remarkable. In RF, models are highly specialized, each carefully designed for a single task: one model to decode, one to localize, one to identify the transmitter, and so on. Jonah Casebeer and @_jaimelien_ have proved that our Newton world model can rule them all!

Physical AI systems are already pretty good at describing the world. The next logical step is to act on it: perceive, decide, act, verify. Next week, I’ll be on a panel with @ethanjb, and @IanRountree at TechCrunch Disrupt. Our panel is called “Physical AI isn’t a Model Problem, it’s an Execution Problem.” Join us to understand what separates intelligence that can generalize to the variability of the physical world from intelligence that doesn’t.

Proud of our team! @FastCompany named Archetype’s world model, Newton, one of its Next Big Things in Tech 2026 for foundational AI. Our ambition: encode the entire physical world in one AI model, extending our understanding beyond the limits of human perception. Honored to see @PhysicalAI alongside @cerebras @liquidai @GoodfireAI, pushing the boundaries of AI together. Onward!

Physical AI pays off when it starts from the right use case. Tomorrow I'll talk about the use cases at @nttdata_inc's Physical AI Summit. I'm joining leaders from @HYGCareers, Litmus, and @ericsson for a fireside chat to discuss three key decisions behind every real deployment of Physical AI: which use case carries the value, which model and architecture fit it, and where AI can intersect with operations on the factory floor. I will talk about our deployment of Newton, our physical world model, at the Hyster-Yale Kentucky Factory, where we power the forklift manufacturing quality assurance and support workers by verifying critical production tasks step by step, in real time. Looking forward to the conversation.

Today I’m attending the inaugural #NovaFutureSummit26 in Napa with @PhysicalAI's COO Brandon Barbello. The summit presents a unique opportunity to gather with other leaders in tech, telecoms, and business for substantive closed-door discussions on the future of AI and global digital infrastructure. Reach out if you're here.

In "The Evolution of Physics", Einstein says that what matters is to pick and frame the question; the solution is mostly technical. That’s even more true now: AI is commoditizing technical. Humans' role will be to decide what truly matters and frame it so machines can solve it.

Excited to be a part of this amazing peer group of incredible entrepreneurs and visionaries and talk about our vision of @PhysicalAI.

Early experiments with Newton running on a Raspberry Pi, discovering the physical world through sensor data, not video or images.

The Manufacturing Leadership Council recently published a survey that validates some of my observations. 53% of organizations report that their data comes from incompatible systems and formats. Another 28% lack the skills to analyze their data effectively. Counterintuitively, more data often means less clarity, and more sensors make it harder to make decisions. Real-world industries need tools for turning data and observations from sensors into actionable understanding. That's the core vision behind our physical world model, Newton. It is pre-trained on billions of cross-modal sensor measurements (vibration, pressure, temperature, current, acoustics, radar, video, time-series telemetry, and more). It generalizes across machines, sites, and use cases without a custom model built for each one and delivers insights, predictions, and automation to our customers. Newton eliminates those pain points and turns data into insights and decisions. Full report: manufacturingleadershipcounc…

I just returned from a vacation that included a stop in Istanbul, a city layered with millennia of history. The artifacts I saw—some more than 2,000 years old—struck me because they reminded me that no two physical objects are ever exactly the same. Time leaves its own imprint on each one. That is the essential difference between the physical and digital worlds. Many algorithms assume that a model trained on one motor, press, or robotic arm should work on the same machine in another factory. Replicable. Deterministic. Software-like. Reality is different. No two motors spin identically. Every press behaves differently. Materials wear and shed atoms. Sensors drift and their signals distort. Each asset is shaped by its environment, maintenance, usage, and history. The physical world is always changing. Trying to anticipate and account for every possible variation by collecting data in advance—even for a simple system—is prohibitively expensive and ultimately impossible. With Newton, we are exploring a different approach: models that understand the fundamental physical behavior of systems, observe each machine, and then dynamically adapt to its particular behavior and conditions. We call these adaptive local world models. This is how we believe Physical AI can scale.

I'll be speaking at @encord_team's Long Horizon Physical AI Summit this October. I'll be joining Eric Chan (Rhoda AI), Sankaet Pathak (Foundation), and Ashish Kapoor (General Robotics) on our panel on whether world models will eat Physical AI. Let me know if you'll be there. 📍 Fort Mason, San Francisco 📅 October 20, 2026 🌐encord.com/long-horizon/

If you know what you're doing, you're not doing anything new.

Thanks to @Politico for including @PhysicalAI in this week's piece on the world models and the key use cases for Newton. Read on: politico.com/newsletters/dig…