The ancient trading hubs of modern science: Bridging the divide between microscopists and data scientists

Biomedical imaging is increasingly defined by a paradox: advances in microscopy now enable the routine generation of extraordinarily rich, high-dimensional datasets, yet the extraction of reproducible, biologically meaningful insight from those data often remains a major bottleneck. This mismatch has widened a growing divide between microscopists and experimental biologists, who generate complex image data, and data scientists, who develop the computational tools needed to interpret them. In this Perspective, we argue that this divide is not merely technical, but also cultural, shaped by different assumptions about validation, dissemination, workflow design, and what constitutes meaningful success. We highlight how artificial intelligence and computational microscopy are simultaneously overrated when treated as turnkey solutions and underrated when deployed thoughtfully to reshape experimental design, data acquisition, and biological discovery. We further argue that current validation practices often fail to capture biological fidelity, creating a dangerous gap between benchmark performance and real-world utility. Finally, we propose that imaging core facilities should evolve beyond service models into collaborative hubs and training grounds where acquisition, analysis, and interpretation are integrated from the outset. By fostering shared standards, iterative workflows, and cross-disciplinary training, core facilities can help bridge the divide between photons and pixels and enable a more rigorous, scalable, and biologically grounded future for imaging science.