Vivodyne, a biotech startup, has introduced HIVE, modular robotic labs designed to address the data gap in AI-driven drug discovery. These labs can cultivate 20 kinds of human tissue and autonomously dose and monitor them, producing causal biological data that current AI models are missing. This data, which comes from animal testing or single-cell studies, is insufficient for training AI to cure diseases like cancer.
The company claims its tissues closely mimic real human organs, with liver cells showing 94% predictive accuracy for toxicity testing and airway tissue matching real human tissue 96% of the time. Bone marrow tissue has achieved 100% concordance in tests of 20 chemotherapy drugs. Last week, Vivodyne opened what it calls the world’s largest 'human data center' near San Francisco, achieving twice the throughput of all U.S. animal trials combined.
Georgescu, Vivodyne’s CEO, argues that existing AI models lack the data to understand human biology’s complexity. He points to studies showing no clear data scaling laws when training generative AI models on static cellular data, which fail to capture causal relationships between cell states. HIVE’s approach, tracking ongoing experiments where diseased tissue is exposed to stimuli, aims to provide reinforcement learning for AI models to better understand human biology.
Source: techcrunch
