This is a pretty good and brief introduction to recent history of replication bias, and why it's there. It goes hard into the Registered Reports solution, which I don't think is that workable (too strict), but that's alright. Definitely recommended for curious lay people. The math is not difficult.
He discusses bias towards positive and novelty results a lot, which is great, but there is no mention of the rather extreme political bias. This is arguably more important because the positive and novelty biases do not distort the field in any particular direction, but political bias does. Very important for psychology to serve as basis of policy, as it does.