Prediction markets are interesting but I don't think they've taken off at all. One of the larger issues with Ethereum is that Solidity, and the EVM as an exten

Prediction markets are interesting but I don't think they've taken off at all. One of the larger issues with Ethereum is that Solidity, and the EVM as an extension, make it really, really difficult to write correct, cheap contracts. Off the top of my head, choices like: * uint256 as a standard word size * the possibility of reentrancy attacks * overflowing integers wrap around * contract upgrades are non-trivial (particularly upgrading solidity versions using an upgradable contract paradigm). * Operator semantics are confusing * State mutability is not as explicit as it ought to be I know bad language features like hoisting were fixed a while back (and maybe some of the stuff I listed as well), but many of us are stuck on pinned older versions of solidity because upgrades. Questions like: "what transactions are present for a particular address" have non-trivial answers. Key management is difficult even for experienced users, as are addresses and identity management. There is some momentum toward using WASM instead of EVM in future Ethereum versions, but I honestly think the proper move is to develop and use a standard language for the purposes of building secure financial contracts. Running a full-node and keeping it healthy (let's say with state pruning, i.e. not a full archive node) is actually pretty difficult. Most people probably use infura because of this. I still don't have a great answer on how to properly determine that a node (geth, unsure about parity) is effectively 1. up to date and 2. healthy (i.e. can be load balanced), making it difficult to automate stuff like load balancer promotion / demotion. Last week the gas price on the network went up like 20-40x for 4-5 days because of an individual contract spamming transactions with a high gas price. It seems like few people noticed and few people cared.