The points on immutability are incorrect. When Ethereum was rolled back to reverse the 2016 DAO hack the data was not lost, it continues on the Eth Classic chain - it is still immutable.
That chain is no where near as valuable today - that's true of most forks, one wins and the other loses. For most Ethereum users, and for the Polkadot users for which a similar fork happened recently, the early reversion to remedy loss caused by an exploit is considered a feature, not a bug and it was the community that decided which fork would win.
There has been much more money lost to similar exploits in various contracts since the infamous DAO hack, and the chain was not rolled back - if it ever was it would likely be a losing fork unless the community of users agreed on its value.
For Ethereum specifically this is more problematic as its full nodes are expensive to run and don't have much say compared to miners or centralized node providers like Infura, but this type of fork in principle isn't an example of immutability, and chains which do subsidize node costs to avoid the centralization seen in Eth's user facing nodes remain in control of the users.
In total, picking out the worst parts of the Ethereum network and generalizing it to 'blockchain' is lazy, and doesn't even fulfill the modest goal of the essay to "[discuss] how these technologies work in practice today," as it ignores the nascent systems whose developers were the first on scene to many of these problems and decided that they were solvable rather than bloggers with slightly deeper than surface level understanding seeing the same problems years later and deciding they were unsolvable.
I don't think it's fair to say their point is incorrect. If you're talking in absolute technical terms about what "a blockchain" is in the abstract then, sure, a blockchain is immutable... but cryptocurrencies don't use "a blockchain" they have "the blockchain" and "the blockchain" is just whichever blockchain the network has decided is the blockchain. For all intents and purposes, "the Ethereum blockchain" is mutable because it can be swapped out if everyone agrees.
That chain is no where near as valuable today - that's true of most forks, one wins and the other loses. For most Ethereum users, and for the Polkadot users for which a similar fork happened recently, the early reversion to remedy loss caused by an exploit is considered a feature, not a bug and it was the community that decided which fork would win.
There has been much more money lost to similar exploits in various contracts since the infamous DAO hack, and the chain was not rolled back - if it ever was it would likely be a losing fork unless the community of users agreed on its value.
For Ethereum specifically this is more problematic as its full nodes are expensive to run and don't have much say compared to miners or centralized node providers like Infura, but this type of fork in principle isn't an example of immutability, and chains which do subsidize node costs to avoid the centralization seen in Eth's user facing nodes remain in control of the users.
In total, picking out the worst parts of the Ethereum network and generalizing it to 'blockchain' is lazy, and doesn't even fulfill the modest goal of the essay to "[discuss] how these technologies work in practice today," as it ignores the nascent systems whose developers were the first on scene to many of these problems and decided that they were solvable rather than bloggers with slightly deeper than surface level understanding seeing the same problems years later and deciding they were unsolvable.