Sudarshan Babu

The paper argues that scaling model capacity alone is insufficient for Virtual Cells, emphasizing the need for contextual diversity and causal representation learning over data volume.