> the three things the author argues cannot occur in a proof of stake system.

The author appears to be saying that "any decentralized consensus via proof of stake system is vulnerable to timing attacks"

The counter-argument that "This here proof of stake system has not been successfully attacked ... that we know of ... yet" does not seem to be watertight.