These days, I've been seeing more talk about bridges being hacked, which makes me even more hesitant about cross-chain interactions… Just to review: what do we really trust when crossing over?

In the idea of IBC, where chains verify each other's light clients, it sounds the most reliable, but you still have to trust: that the other chain itself doesn't go down, that validators don't collude in malicious behavior; the relayer just moves messages, in theory it shouldn't be able to steal, but if it goes offline or gets stuck, you're just left waiting.
Compared to regular message passing/multisig bridges, it's more straightforward: you trust that the set of signers/relayers and the contract aren't maliciously written.
Recently, everyone’s been waiting for oracle errors to be confirmed before acting, and I’m pretty much the same—prefer to wait longer, wait for a few more finality rounds before taking action. Saving on transaction fees isn’t worth it if assets are lost.
That’s all for now.