Scuttlebutt

Scuttlebutt

Archetype AI CEO: 80% of the world economy runs on physical industries, short of workers

8 insights: why his AI model learns physics from sensor data instead of equations, why it has to run inside the machine, and why he says the American Dream comes down to acceptance

Scuttlebutt's avatar
Scuttlebutt
Nov 22, 2026
∙ Paid

Forbes · Nov 21, 2026 · 24m · 16m saved

With Ivan Poupyrev, Founder and CEO of Archetype AI

Archetype AI CEO: 80% of the world economy runs on physical industries, short of workers

Ivan Poupyrev, founder and CEO of Archetype AI, told Forbes in an interview that 80% of the world economy still runs on physical industries, and that his company is building AI for them rather than for the internet.

Large language models learn from text. Ivan's model, Newton, learns from the readings of sensors inside machines, and it is meant to work out the laws of physics without being taught them.

"The physical world around us... is not operating on language. Machines don't speak in language to us. It's mostly speaks in the language of sensors." — Ivan Poupyrev

Ivan left the Soviet Union shortly after it dissolved and reached the US by way of Japan. He worked on early virtual-reality headsets at the University of Washington, then held jobs at Sony, Disney's Imagineering research arm and Google, where he worked on Pixel phones and Nest devices, before founding Archetype AI.

The full interview is covered here so you can skip it. 24 minutes of audio, 9 minutes of reading.

Here are the 8 insights that matter.


This Scuttlebutt Summary is available in full to premium subscribers. If you're not one already, please consider subscribing to keep reading.


Key Takeaways

  • Ivan went from working at founder-built companies to founding his own, and calls the US the only country that supports and encourages it

  • Archetype AI's Newton is a foundation model for sensor data, not for language

    • It reads vibration, electrical current, voltage and chemical signals straight from machines

  • Newton is trained to work out the laws of physics for itself, the way Kepler worked out planetary motion from years of observations

  • Physics is more compact than language, so Ivan argues a physics model can be more efficient than a language model

  • Newton keeps learning after it is installed in a machine, adjusting as parts wear and conditions change

  • Ivan tells young workers to take AI into physical industries, where he says the labor shortage is the biggest crisis

  • The core of the American Dream, for Ivan, is acceptance: new arrivals are treated as if they already live here

1. A Route to America via Japan

The interview was filmed near the Statue of Liberty for America's 250th anniversary, and it opened on how Ivan got to the US. He grew up in the Soviet Union and left shortly after it dissolved.

Ivan reached the US by way of Japan, not directly from his home country

"My path to America was not direct from my home country straight to America. I went through Japan." — Ivan Poupyrev

From Japan he went to the University of Washington's Human Interface Technology Lab in Seattle as a visiting scientist. There he worked on early virtual-reality headsets, which were bulky and built on CRT screens. He went back to Japan to work at Sony.

Ivan concluded that a career in new technology had to be built in the US

"If I want to do this, I have to be in America, because in all the new technologies, America leads." — Ivan Poupyrev

He moved to the US for good in 2009, to join the research and development division of Disney Imagineering. He later moved to Google, where he worked on hardware including Pixel phones and Nest devices.

2. Founders Pushed Him to Found

User's avatar

Continue reading this post for free, courtesy of Scuttlebutt.

Or purchase a paid subscription.
© 2026 The Golden Age LLC · Privacy ∙ Terms ∙ Collection notice
Start your SubstackGet the app
Substack is the home for writing, podcasting, and media founders.