These past few days, I've been looking into IBC/messaging/bridges, and the more I look, the more I feel that cross-chain is basically about "who do I trust."

A cross-chain transfer from A to B might require trusting: the source chain won't rollback, the target chain can verify correctly, the light client/validation logic isn't flawed, the relay/forwarding isn't malicious, and most importantly, when dealing with "external bridges," there's an additional layer of multi-signature/oracles/operational trust assumptions...
The more components there are, the less surprising it is when something goes wrong.

Then I see the shared security and staking models being criticized as "yield stacking," and honestly, I kind of envy how well they can tell the story, making the yields look very attractive.
But after doing risk control backtests for a long time, my first reaction is: which layer of trust risk is that extra yield actually compensating for?
If I can't figure it out, I prefer to avoid it for now—I'm not good at chasing high returns anyway.