I'm torn between 4 and 5 stars. Near has borrowed a lot of the best elements from some of the leading blockchain projects and incorporated it into their own to create a very solid product. They have also made clever improvements like Randomized Block Proposers, which makes it more decentralized than other POS networks and dynamic validator sets, which allows the consensus mechanism to be adjusted as the network grows and requirements change.
One of their main contributions is a more scalable blockchain, which is achieved through sharding. However, instead of sharding via sidechains they partition the network into smaller sub-networks that process transactions in parallel and communicate with each other. This allows for better scalability without compromising on decentralization or security (but it compromizes on flexibility. (See image attached)
These are all good things. In my opinion it's lacking a game changing innovation of its own. They mention that the rainbow bridge is the world's first trustless bridge. But their docs say it depends on "relayers, which run on traditional servers", which makes them centralized and therefore not trustless as far as I understand. I would love it if someone can convince me that the rainbow bridge really is trustless.
Anyways, I'll still give them 5 stars because I think their sharding model will offer a more decentralized and secure option in a future of scaled blockchains, and decentralization is the whole point of blockchain. Show Less