While foundational model companies like OpenAI and Anthropic are now hiring philosophers to argue about the consciousness of their LLMs, the real minds and, increasingly, monies, are pouring into physical AI - the application of machine learning to discover new materials and conduct experiments in the real world.

The last thing any of us want is a world where every company across every sector is ceding value to a few models that eat everything they see. If all the value is accrued by only a few models, the political economy will simply not tolerate it. There is no societal permission for an AI future that hollows out entire industries. [...]

In my view, our priority has to be building a frontier ecosystem, not just a frontier model, so value flows broadly across every company, every industry, and every country. One where every organization can own the learning loop that encodes its institutional knowledge, compounding its human and token capital.

-Satya Nadella, a most singular propper-upper of said models

Unlike Sam Altman or Dario Amodei, neither Kirsten Persson nor Ekin Dogus Cubuk is much of a household name.

Neither has so far peacocked about for many onstage interviews or been prone to dark, sweeping statements about how their creations might spell the end of mankind or any future human employment.

Yet the body of their work, both individually and collectively, represents the foundation of AI materials science - a field that has attracted more than $1.3 billion in investment from top VCs. And the two companies with which each is involved are leading the crop of new startups driving this space forward.

With OpenAIs IPO now delayed until 2027, investors, nervous about the gargantuan spending and lackluster adoption in the LLM space, are looking for the next iteration of AI soonicorns, and this batch offers the promise of eking more efficiency out of chips and batteries.

Doge, as Dogus Cubuks friends call him, is the co-founder and co-CEO of Periodic Labs, which has raised $300M to date from a16z, Felicis, DST, NVentures (Nvidias venture capital arm), and Accel and individuals including Jeff Bezos, Elad Gil, Eric Schmidt, and Jeff Dean. Another half-billion, led by AMP, is reportedly in the works at a $7.5B valuation.

Persson, along with Geoffrey Hinton and Yann LeCunn, serves as an advisor to Cambridge-based CuspAI, which, as of publication time, has just confirmed (but not yet closed) a $400M round led by Bezos Discoveries and Kleiner Perkins at a valuation of $2.6B. 

Still, the gulf between the computational promise of AI materials work and its physical success in creating those materials yawns large. 

Over the course of their careers, Persson and Dogus Cubuk have been both collaborators and competitors, and their journeys illustrate both the promise and the perils of applying AI to materials science.