Honestly there’s a lot of valid points to be made here but the actual report reads as if their intention was to prove blockchains are insecure and centralized.

That is true for a lot of them, but true Nakamoto consensus is not quite as fragile as they suggest it is.

They don’t provide an analysis of the true cost of launching a 51% attack.

Their assertions about the security risk of “altering the software that nodes run” fail to mention how this is a voluntary process which all node operators choose to undergo. If a consensus emerges on the network or a subset of the network that the changes are problematic, these dissenting node operators can choose to hard fork. There will be few supporters of an obviously malicious attack in the network, so it would be unable to gain traction.

Their point about the number of entities in control of Bitcoin is technically correct, because of the way that pooling works in Bitcoin: many nodes send any propfs they find to one node, and that one node writes to the blockchain. So, there is a definite concentration of power. There are some in depth game theoretical analyses of why this is unlikely to become a problem but in general it is easy to imagine that, for instance, the US treasury would not want to destroy trust in the USD.

Interestingly, Chia, a new proof of work blockchain which launched a year ago, developed by Bram Cohen, has a unique and innovative solution to pooling which does not result in concentration of power: individual node operators submit proofs to the network, not to the pool, and the pool receives a fraction of the reward for minting a new block. Chia also has more full nodes than any oher blockchain, including Bitcoin. At this point it’s relatively unknown however.