Traditional 2D cell cultures can be misleading. Advanced 3D models, by reconstituting the tumor microenvironment with stromal cells, can uncover mechanisms of drug resistance (e.g., to ADCs) that are completely invisible in simpler systems, providing more clinically relevant data.

The next major AI breakthrough will come from applying generative models to complex systems beyond human language, such as biology. By treating biological processes as a unique "language," AI could discover novel therapeutics or research paths, leading to a "Move 37" moment in science.

The next leap in biotech moves beyond applying AI to existing data. CZI pioneers a model where 'frontier biology' and 'frontier AI' are developed in tandem. Experiments are now designed specifically to generate novel data that will ground and improve future AI models, creating a virtuous feedback loop.

A major challenge in phenotypic drug screening is determining a compound's mechanism of action. AI models can analyze the complex visual data of cellular condensates after drug treatment, extracting maximal information to understand how the drug is actually working inside the cell.

The progress of AI in predicting cancer treatment is stalled not by algorithms, but by the data used to train them. Relying solely on static genetic data is insufficient. The critical missing piece is functional, contextual data showing how patient cells actually respond to drugs.

Traditional methods like crystallography are slow and analyze purified proteins outside their native environment. A-muto's platform uses proteomics and AI to analyze thousands of protein conformations in living disease models, capturing a more accurate picture of disease biology and identifying novel targets.

The low-hanging fruit of finding a single predictive biomarker is gone. The next frontier for bioinformatics is developing complex, 'multimodal models' that integrate several data points to predict outcomes. The key challenge is creating sophisticated models that still yield practical, broadly applicable clinical insights.

Following the success of AlphaFold in predicting protein structures, Demis Hassabis says DeepMind's next grand challenge is creating a full AI simulation of a working cell. This 'virtual cell' would allow researchers to test hypotheses about drugs and diseases millions of times faster than in a physical lab.

A 3D model is considered "advanced" when it's a bioactive system recreating a tissue's microenvironment. It's not just about three-dimensional growth; cells must both influence and be influenced by their surroundings, including architecture, diffusion gradients, and mechanical cues, to be truly representative.