That requires trust between the two gateways, and the sender trusting both of them.
Using bitAssets for remittance the only trust needed is that the gateway not cheat you while you're physically standing there face to face.
Not necessarily. You have IOU USD on Gateway 1 but want to send funds to Gateway 2 for cashout in another country. For the other person to even get out of the Gateway 2 he needs to be signed up with Gateway 2 first. In this scenario bitUSD could be used as a better ripple-like XRP (more stable). This would provide
incredible liquidity to bitUSD and as a result greater stability and more reason to hold (probably a better yield too).
However, what the gateway providers really want isn't a counterparty-less token as a middlemen for reduction of their own risk profile, they rather want a way to create a risk contract directly between the two gateways. Think of it as a CFD between two gateway IOUs. Banks already to this though the interbank lending market and Credit Swaps. Such a crypto contract would be a lot cheaper overall, but not liquid enough because it would be a A-to-B contract rather than an A-to-N and N-to-B contract (e.g. Credit is time delayed split barter. Money is counterparty-less credit).
In this regard, what Bitshares really needs isn't just more gateways, but adding automatic pathfinding just like ripple, whereas the main high liquidity and non-counterparty token would be the bitUSD.
Interestingly enough, the bitUSD yield to be expected in such a setup would be capped by the market as a function of the total perceived counterparty risk of the gateways + the network cost of providing liquidity among sufficient pairs of the gateways.