I read the paper. It's well written, short, accessible, and answers most questions a reader would have about the claims. It introduces a complete system and eva

I read the paper. It's well written, short, accessible, and answers most questions a reader would have about the claims. It introduces a complete system and evaluated performance tradeoffs involved in using different cryptographic primitives. I this paper shows that simple digital currency systems do not need to be complicated to be fast. Crucially it includes benchmarks on real hardware, although the simulated configuration doesn't correspond to the real world in important ways, such as with latency and failures. I also don't understand how sharding interacts with the quorum thresholds. Is there a separate q of n quorum for each shard? Or is sharding completely separate, ie each validator must run one separate process per shard? As for the relationship to cryptocurrency systems, I think this paper is overly naive. The idea that the client should be responsible for providing the existence of a spendable input, and that commuting transactions don't need to be sequenced by the protocol, is not novel and exists in many cryptocurrency systems. The most well known is probably the planned ETH 2.0 system, which uses receipts to send information between shards.