Lately I've been looking into IBC / cross-chain messaging, and the more I look, the more I feel that a single "cross" is actually signing a bunch of invisible trust contracts: the chain itself must not have issues, the light client / verification logic must not be written crooked, relayers must not go offline or act maliciously, the other application must not parse things arbitrarily, and ultimately, the order of the transaction landing / MEV still depends on luck... Honestly, it's not about whether there's a bridge or not, but about which components you're willing to trust.

I've been tortured by impermanent loss in market making for a long time, and now I'm especially sensitive to "single points of failure." If I can add an extra layer of redundancy, I want to add one—just as a backup for myself: it's best if the same message has an alternative path / can be rolled back in design, otherwise, getting stuck once really blows your mind.

By the way, the NFT royalty disputes happening recently also seem quite similar: everyone wants to "trust less, friction less," but when liquidity tightens, who will cover the losses, who will pay the bills, starts to become a tug-of-war... Anyway, with cross-chain stuff, don’t just look at the smooth experience; first, consider whether you can handle the worst-case scenario.