A new race in artificial intelligence: the startup aiming to build "general biological intelligence"

As the AI industry increasingly focuses on the concept of "artificial general intelligence" (AGI), a new American startup is proposing a medical-specific variant: "general biological intelligence." This is Radical Numerics, a company recently featured in an episode of the AI Prognosis podcast, hosted by journalist Brittany Trang for STAT News AI.

What Radical Numerics is trying to achieve

According to STAT News AI, the company's CEO, Eric Nguyen, explains that the startup's goal is not to build a simple diagnostic tool or a narrow predictive model, but rather to develop AI systems capable of "understanding" biological mechanisms at a level of generality comparable to how large language models process text. The underlying idea is that if an AI system can learn the fundamental patterns of biology — from molecular interactions to complex cellular processes — it could later be applied to a wide range of medical problems, rather than just one specific task.

This approach fits into a broader trend in AI-driven biomedical research, where companies and academic labs are attempting to build "foundation models" for biology, similar to those used in natural language processing. The major difference, the source notes, lies in the extreme complexity of biological data, which is far more heterogeneous and harder to standardize than text.

Challenges and uncertainties

In his conversation with Brittany Trang, Nguyen acknowledged that the path toward such "general biological intelligence" remains uncertain, and that concrete, clinically applicable results are still a long way off. According to STAT News AI, the startup is at the very beginning of a long-term research process, with no guarantee of immediate results.

The original article also highlights the skepticism within the scientific community toward ambitious AGI-related promises — skepticism that naturally extends to the narrower concept of "general biological intelligence." Experts cited by STAT News AI warn of the risk that investor enthusiasm could outpace the technology's actual maturity.

For now, Radical Numerics remains an early-stage project, and only time will tell whether its approach will lead to concrete progress in the computational modeling of biology — or whether it will remain just one of many ambitious initiatives in the field of AI applied to healthcare.