A new intelligent behavioral system called "Newton" uses messaging to comprehend millions of data points.

With the help of an assistant, the startup Archetype has created an AI model named Newton that can analyze millions of data points from a vast array of sensors located all over the world, respond to queries, provide graphs, or even write code to report events. A clever texting experience similar to ChatGPT.

One of the company's founders and CEO, Ivan Popyrev, stated to Wired that "Architype" aims to enable direct, tangible contact between the digital and physical worlds.

 Arquitype is seeking funds through a $13 million seed round.

The Newton model, according to the business, is more of a huge behavioral model than a linguistic model since it demands the user be specific about what he wants to know from the data when he interacts with it to grasp what the many sensors connected to it are collecting.

An illustrative example of the model's operation was provided by "Architype" in a video clip that the company posted on its official website. Several sensors were placed inside an envelope, treating it like a package of goods. These sensors were connected to "Newton," which asked the user if the envelope's contents had been handled carefully.

On the side of the carrier or not, then the model will examine the data of the vibration sensors and the directional compass to determine the level of rotation of the content, as well as whether it is subjected to any vibration during transportation. if the user requests to know where the package is currently located, the model will focus on the data received from the GPS sensors.

"From the heart of Google"

The Archetype team emerged from the core of Google, where Bobyrev led a group working on the "Soli" project, which involved creating tiny radar devices and integrating them into wearable technology so that they could react to a user's gestures and movements. 

The primary obstacle to this team's progress was the intricacy of deciphering, interpreting, and analyzing radar sensor data. 

Therefore, Popyrev and his associates recognized that when large language models (LLMs) surfaced, they could enhance the power of sensor data by enabling people to examine data gathered from extensive regions of the planet through straightforward adjustments. Place and time are readily apparent. 

The team was eager to develop their idea outside of Google, so in March 2023 the team members resigned and decided to create Archetype.

Practical applications

Through numerous examples, the company explained some possible scenarios for using its new smart system. For example, for a quality officer in a factory to be able to verify the efficiency level of performance of various production lines, he would previously have needed to move between all of them to ensure their safety, but with Newton, he can verify this while he is in his office room, by asking the smart assistant a question about the status of the production lines, and in turn, he will answer by analyzing a huge amount of data collected from the sensors deployed on the production lines, which saves time and effort, and also allows for executive decisions to be made in real-time. Which increases productivity.

Franziska Bossart, head of Amazon's Industrial Innovation Support Fund, indicated that the company is one of the first investors in "Architype," explaining that the emerging company's technologies will benefit the e-commerce giant in improving the flow of goods through collection centers and warehouses and improving the speed of delivery to customers.

Stefano Bini, a professor in the Department of Orthopedic Surgery at the University of California, San Francisco, praised that the new smart chat platform provides access to a set of graphical statistics collected from sensors installed on a patient’s foot, to evaluate his level of recovery after knee replacement surgery.

Volkswagen, one of the investors in "Architype,"  has begun several early tests of the "Newton" model, including linking sensors in one of the company's cars with sensors in the smart garage door lock at home, so that the user can address the garage door lock to open automatically, as soon as it detects "Newton" returns user.

“This mission is just the beginning,” said Brian Lathrop, senior researcher at Volkswagen’s Innovation Center. “Imagine what would happen when you linked all those cameras from millions of vehicles on the road, in parking lots, and on driveways together? If you had an artificial intelligence that looks “to all those graphic feeds, it opens up a huge amount of possibilities and uses cases.”