Humans are considered to be highly prosocial, especially in comparison to other species. However, most tests of prosociality are conducted in highly artificial settings among anonymous participants. To gain a better understanding of how human hyper-cooperation may have evolved, we tested humans' willingness to share in one of the most competitive fields of our current society: academia. Researchers were generally prosocial with 80% sharing a PDF of one of their latest papers, and almost 60% willing to send us their data. Intriguingly, prosociality was most prominent from male to male, and less likely among all other sex-combinations. This pattern suggests the presence of male-exclusive networks in science, and may be based on an evolutionary history promoting strong male bonds.