Hi George,
First of all, welcome! I hope you will get some clues from this community for the benefit of your venture. I read your CoinDesk article, and will try to continue your line of thought. To restate it, the issue is enabling cross-border micropayments. In that regard, cryptocurrencies have proven to bring forth one big advantage and an equally important drawback. The advantage is obviously that the transfer of value-bearing assets is made easy. The drawback is that one is limited to a specific asset group in such operations (i.e., cryptocurrencies), which obligates the existence of local cashflow nodes (i.e., fiat-to-crypto exchanges).
Let us restrict ourselves with US dollar remittances, since even that remains unsolved. If Alice, who is in the US, wants to send $200 to Bob, who is in EU, she would exchange her dollars for, say, bitcoins, and send the bitcoins to Bob, who will then exchange the bitcoins for dollars. In this case, the imbalance between US-to-EU transfers and EU-to-US transfers (cf. "P2P netting of payments" in your CoinDesk article) is resolved by way of arbitragers.
As for fiat-pegged crypto-assets, they seem to be an abstraction layer on top of the abovementioned dollar->bitcoin->dollar method. As such, they bring a new set of solutions and problems. The solution is, again, obvious; easy transfer of dollars. However, if Alice is to send a fiat-pegged crypto-asset to Bob, she will still need to convert her dollars to crypto-dollars to begin with. In this case, again, centralized local institutions may help. But this time those institutions need to agree on an interoperability framework which must provide not only the reliable transfer capabilities but also a crypto-dollar the validity of whose pegging is beyond local circumstances. This requirement holds regardless of how the pegging is undertaken.
Now, fundamentals: There is no taking without giving. In order to get a fiat-pegged asset, one needs to put some value somewhere. Let's say Alice buys 200 crypto-USD from the local vendor by depositing $200. As Bob gets the pegged assets, he will need to cash them out for dollars. But the $200 is still in Alice's jurisdiction. Thus, there needs to be a backend where the local vendor converts Alice's deposited money into the underlying cryptocurrency and sends it to Bob's jurisdiction at the same moment when Alice transfers the $200-worth crypto-assets. Not only this, but also the local vendor in Bob's jurisdiction should convert the underlying cryptocurrency into dollars in order to transfer them to Bob's bank account.
This model calls for the entire set of problems associated with liquidity and credit risk management. One might ask at this point why Alice buys the crypto-dollars by depositing dollars. She may as well use a cryptocurrency such as bitcoin to acquire the said asset. However, as has been said, the fiat-pegged cypto-asset already serves as an abstraction layer to this process. So it would be in this case meaningless to send the fiat-pegged crypto-assets instead of the underlying cryptocurrency. To approach the problem from another perspective, Alice needs to deposit dollars at some point in order to get the cryptocurrency in anyways. Hence, if Alice directly uses dollars to buy crypto-dollars, it means this: Alice asks her local vendor to make the conversion on behalf of herself. Well, it sounds like this is a service. As a service, it should cost something. So the question is whether or not it would cost more than SWIFT.
To restate a cardinal point, those considerations do not depend on how the pegging is undertaken. The vendor may faciliate Nubits or Bitshares, and bitcoins or other cryptocurrencies may be utilized as collateral. It does not matter. In the end, Alice will need to get that pegged asset. As far as my research goes, there is no fulfilling solution to this problem which is publicly discussed. Notwithstanding, I believe I have found a solution which does not at all depend on the idea of pegging, but rather on playing with the fundamental logic of "no taking without giving." If somehow Alice could deposit dollars without actually depositing cash, and if this process would yield a kind of blockchain entity, then there might be a novel path upon us. But I will not go into the details of that.
Returning to an earlier issue, if there are local exchanges enabling dollar-to-cryptodollar conversion back and forth, then different exchanges located in different parts of the world should work with one and the same crypto-dollar asset whose value they all agree is pegged to dollar. Begging help from central banks won't work, since using blockchain cannot magically help them solve the issues which prevent them to facilitate real-time cross-border settlements. Utilizing a centrally issued crypto-asset (such as TetherUSD) won't work either, since it involves a considerable default risk. One may think that a licensed multinational money transfer agent (such as PayPal) could issue the needed crypto-asset. However, again, utilizing blockchain could hardly solve the problems that prevent PayPal to allow seamless and cost-efficient cross-border operations.
As for bitUSD or Nubits, as far as the issue of cross-border settlement is considered, they serve as nothing but an abstraction layer on top of sending the underlying cryptocurrency, leaving all the liquidity management problems unsolved. Therefore, a shift of focus is needed. We need to understand that the problem is not in designing a fiat-pegged crypto-asset, but in managing liquidity and other risks. Take BitPay as an example. They do not have a fiat-pegged crypto-asset. They take 1% fee, and they make it possible to pay in fiat with bitcoins. We thus need a similar complementary structure with which one may pay in fiat with fiat - but with the intermediation of a reserve cryptocurrency.
That is all I have to say for now. There are much to discuss about how to structure a solution for the last point I made, but I feel that I have already written too much.
Best,
Jian