Note that one of the assumptions isn't necessarily true.

> The culprit has to fit the crime. Most importantly, it has to fit the timing of the crime. What we’re looking for is something that happened around 2020 (uh, seems obvious) and then didn’t recover (ah, that’s the hard part). This timing rules out several otherwise plausible suspects.

You can pile straws on a camel as part of a continuous process and then observe the breaking of the camel's back as a discontinuous result.

Any explanation that doesn't fit the timing (like the "decline of religion" example he uses) may still be relevant. It can't be automatically ruled out, but the timing is a strong piece of evidence against it. The theory needs to include a solid explanation for why the timing doesn't seem to match. I don't think decline of religion has such a solid explanation, but other theories might.