A biotech startup called Vivodyne believes the AI drug-discovery industry has a fundamental flaw: a lack of human tissue data. Their HIVE machines can grow 20 kinds of human tissue and autonomously dose them, generating the kind of causal biological data that today’s AI models are missing.

The company’s CEO, Andrei Georgescu, argues current AI models rely on static snapshots of cells rather than understanding how those cells change over time. This limits their ability to accurately model complex human biology and diseases like cancer.

While a handful of AI-designed drugs have entered clinical trials, the vast majority of drugs fail in these trials despite promising results in preclinical studies. Vivodyne’s HIVE machines aim to address this by providing more accurate data before clinical trials are conducted, thus reducing costs and improving drug development efficiency.

Georgescu envisions a future where AI can better understand complex human biology through the kind of reinforcement learning generated by his HIVE machines. This could be crucial for developing combination therapies that target multiple pathways simultaneously – key in tackling complex diseases like cancer.