yeah, there's a lot more to it than just "the blockchain says you have influence over it", sure it's a self-referential system, but so is just standard PoW so it's not really a good argument in my opinion.
The author is right about one thing, though: Conflicts of consensus will absolutely test blockchains of any type or flavor, and doubly so when Dear Leader is gone. Ethereum does have the most prominent leader of any chain I can really think of, and that changes how the community and the chain at large reacts to external stimuli.
PoW is not self-referential. If someone shows you three blockchains, say pow1, pow2, pow3 you can roughly calculate the amount of real-world resources that were required to create each chain based on the hash difficulty. That work/effort cannot be faked. If the difficulty is high enough, you can be certain that this chain was not just generated in someone's basement last night.
With PoS, if someone shows you pos1, pos2, pos3, you have no idea which to trust without external information/trust. They all could have been generated the night before in someone's basement.
If it was discovered tomorrow that there was a massive secret bitcoin mine in a cave in Siberia that produced a chain 2x as long as the main chain and we didn’t know about it, and this miner also gave themselves all the coins, the community would choose to ignore that and use the shorter one we have built consensus around.
Note that this weakness was discovered early on. Bitcoin actually uses a sum of difficulty, not longest chain. You can't just mine more 1 difficulty blocks. But your point stands, if the US Government is mining bitcoin it is plausible that they could invalidate the entire chain.
Whats a community? How can it ignore the longest technically valid chain if people come and go, centalized "reference" websites change hands? The Bitcoin is the longest chain, most work burned, it's the definition and trivially verifiable by anyone now or 1000 years from now (if it retains the majority of humanity hashrate)
This doesn't mean that bitcoin is self-referential. It means the consensus model encompasses more information than purely the length of the chain. But that should be obviously true; there have been countless hard forks like BCH and yet we still follow the chain we all call BTC.
The problem with hard forks in the PoS world is that you cannot judge the health of each chain by some external factor.
Imagine that for a chain to be acceptable, it must chain back to a known genesis block, and that that genesis block distributed stakes to 100+ people of whom you trust at least half.
That's more equivalent to the PoS system Ethereum is using, and has much stronger guarantees.
Getting agreement on a genesis block is a theoretically difficult problem... But practically it's as hard as agreeing if it was George Washington who was the first president of the USA or Bill Gates.
Well it's not and been solved for the first time by Satoshi Nakamoto. You pick the first block of the most worked chain. Any other thing is no better than WebMoney.
The author is right about one thing, though: Conflicts of consensus will absolutely test blockchains of any type or flavor, and doubly so when Dear Leader is gone. Ethereum does have the most prominent leader of any chain I can really think of, and that changes how the community and the chain at large reacts to external stimuli.