Recently, I’ve been analyzing my “Mileage” system again: a cross-chain transfer from A to B. Basically, you're not just trusting one bridge; you're trusting a series of components. IBC’s message passing sounds great in theory, but in practice, I first think: are the light client/verification methods reliable, will the relayer slack off, is the security of the other chain sufficient, and are there pitfalls in routing/bridge contracts? Missing any link could lead to a situation where “the money arrives but the message doesn’t,” which is awkward.

These days, modularization and DA layers are being hyped up, making developers very excited, while users are confused… I’m confused too, but at least I know: where the data is stored, who can prove it, who can do evil, all ultimately come back to “who do I trust.” Anyway, now I split cross-chain transfers into two transactions, with smaller amounts. If native IBC can be used, I avoid complicated bridges; the process is like going to work—automated. If it doesn’t work, then forget it. Let’s not discuss it for now.