In addition to the comment already brought to light by Marko, the article could mislead readers on other blockchain frameworks.

Case 1: For example, with regards to the claims on the transaction throughput, the reader may think that all other blockchain networks are too greedy in block creation and they execute transactions later once the ordering is done. The fact that they need state comparison contradicts that they cannot merely blind-add transactions to order without knowing what the expected state after execution is.

Case 2: The continuation of it on the throughput of a blockchain framework is debatable. These numbers could significantly go down based on the contents of the transactions, and their correlation, policies set for endorsement rules etc. Other blockchain frameworks generally lack flexibility to configure these aspects. But the recent trend has shown that they have realized how useful the consortium operations are.

On the other hand those frameworks too have execute, order, verify mechanism. It's just that the way they order transactions is different from the Hyperledger Fabric model. You wouldn't require a client/external agent to collect endorsements and send to another cluster for ordering. Rather, those blockchain frameworks do these by themselves without the help of an external client.

Case 3: It is claimed that transaction determinism is required only on other blockchain frameworks, reducing their performance. This isn't true. Transaction determinism is required in every stateful blockchain framework. What that would mean is a transaction executed any point in time, on any machine, consuming State1 produces State2. This is the proof that nobody tampered with the history.