Derya Unutmaz presents a compelling vision of the future of scientific research, catalyzed by the integration of AI models directly into the laboratory workflow. The central pillar of his work is the use of AI to bridge the gap between biological complexity and computational modeling. By using LLMs to generate functional code, he creates tailored applications that allow for the visualization of single-cell data, such as flow cytometry, which was previously confined to highly rigid, expensive third-party software. This transition represents a fundamental shift in scientific agency, enabling domain experts to build tools that would have previously required months of dedicated software engineering.

Central to his thesis is the concept of the 'digital twin,' a comprehensive model of an individual's biological system. Unutmaz argues that current medicine is hindered by a 'one-size-fits-all' approach, where effective drugs for some patients are administered to millions regardless of their unique genetic or immunological profiles. The move toward digital twins will allow researchers to test hypotheses and pharmaceutical treatments on simulated versions of individual patients, reducing the risk of clinical failure. This methodology, he predicts, will eventually make clinical trials significantly faster and more accurate.

However, this progress is contingent on a substantial expansion of global compute power. Unutmaz acknowledges that simulating the trillions of reactions occurring in the human body exceeds current capacity, but he remains optimistic that the exponential trajectory of AI advancement will solve these limitations within the next decade. He asserts that this evolution is not just an incremental improvement but a total reinvention of the scientific method, which he terms 'Science 2.0.'

Unutmaz addresses the skepticism surrounding AI in science by pointing to his own experience, noting that the reasoning capabilities of recent models have surpassed human intuition in specific, complex tasks. He challenges the notion that scientists are being replaced, suggesting instead that the role of the researcher is shifting toward that of a master orchestrator who defines the hypothesis and lets autonomous agents execute and iterate on the experimental design.