They should be allocated towards intellectual freedom biased towards randomness.

Existing knowledge is preserved implicitly and be the public, and well-trodden ideas are furthered by industry. Academia is the best place for experiments, which are necessary to avoid stagnation, because there’s only so much obvious (low-hanging) research which isn’t experimental.

Related: “Can random experimental choice lead to better theories?”, https://journals.sagepub.com/doi/10.1177/26339137261421577