We have to look not far from our own field to see a replication crisis brewing. The results in ML papers are obviously moving in the same direction, all while beating SOTA by miniscule amounts.

IMO psychology became a lot less interesting when it tried to become more scientific.