This AI visionary is betting on physics instead of building another ChatGPT; All you need to know about Anima Anandkumar

The entire technology industry is racing to build powerful AI chatbots and models to counter the AI leaders like OpenAI and Anthropic. But this AI visionary just placed a very different bet on the laws of physics, and not the rules of language.

This week, Anima Anandkumar, a Caltech professor and an AI infrastructure engineer, and Benedikt Jenik announced Accelerated Understanding Inc, a startup that believes that the next leap in AI won't come from model training, but from teaching them to understand the physical world.

Must read: OpenAI reveals how Jalapeño Chip performs, wider deployment coming in 2027

Anandkumar and Jenik have also walked away from a lucrative offer to join Jeff Bezos-backed Project Prometheus to build something of their own. Here’s what you need to know about Anima Anandkumar and her vision for an AI breakthrough. 

Who is Anima Anandkumar?

Anandkumar is known to be a professor of computing at the California Institute of Technology. She was born in Mysore, Karnataka, to parents having background in engineering and mathematics. She also has an interesting educational background with a bachelor's degree in electrical engineering from IIT Madras in 2004 and a PhD at Cornell University in 2009, followed by postdoctoral research at MIT.

Anandkumar started her professional journey as a professor at UC Irvine before moving into the technology industry. She worked as a principal scientist at Amazon Web Services (AWS), and later became senior director of AI research at NVIDIA. At NVIDIA, she led research into how its powerful GPUs could be used for advanced AI applications.

Must read: Anthropic's Claude has some tricks ChatGPT doesn’t: Key features explained

Currently, she is the Bren Professor of Computing and Mathematical Sciences at Caltech and was previously recognised as the university's youngest named professor.

What’s the idea behind Accelerated Understanding Inc?

Anandkumar’s Accelerated Understanding believes that language-based AI has limitations, as AI models essentially work with language because they learn patterns in sequences of words or tokens. But Anandkumar argues that the physical world doesn't behave like a sentence, as they include continuous variables, mathematical relationships, and physical laws.

Accelerated Understanding is betting on the technology called neural operators. This will help the company build an AI system that can learn from physical systems and potentially apply that understanding across different problems. Now, the company claims that its system was able to process 5 trillion data points in a single prompt.