Science is a Strong-Link Problem

Science is a strong-link problem, so its progress depends on how good its best work is, and bad ideas mostly fade over time. The author argues that scientists instead treat it as a weak-link problem, relying on heavy gatekeeping such as peer review and grant applications, which can filter out novel, risky research. The author cites roughly 15,000 collective years of annual peer-review effort and notes that US science spending is about 10 times its 1956 level, yet scientific productivity has fallen for decades.