cover image for the interview of TeqAtlas with Fusion Foundation

Q&A with John Liu of Fusion Foundation: potential of asset digitalization & other insights

  • TeqAtlas
  • Jul 11, 2019

TeqAtlas talked to John Liu, Chief Product Officer at Fusion Foundation to discuss how to validate the truthfulness of smart-contract triggering events, what disturbances smart contract may bring to stock markets and what potential asset digitalization has in terms of giving access to new pools of liquidity.

About Fusion Foundation

Fusion Foundation, is a nonprofit developer of an open-sourced blockchain protocol to drive the Internet of Value (IoV). The Fusion protocol empowers next generation interoperable architecture with the ability to express and monetize digital asset ownership over time. A growing number of companies are already integrating Fusion’s groundbreaking protocol into their blockchains systems. For instance, in just the last few weeks, Fantom Foundation selected Fusion’s cutting-edge interoperability solution to power their Distributed ledger; Grayblock Power partnered with Fusion to create investment opportunities within clean energy; and Fusion’s interoperable blockchain protocol was chosen for the first ever NFT (non-fungible) tokenization of archaeological assets, democratizing access to artifact-based investments.

1. Firstly, we would like to ask what kind of assets currently you think are the hardest to digitize and why?

I like to think of the question in the context of what makes sense to tokenize given the constraints of the existing legal restrictions, logistics, and the understanding of what value we're trying to digitize. The act of digitizing is relatively easy no matter the asset. You have a digital representation of whatever the underlying asset may be. Where problems arise is in trying to navigate the three aspects I just listed: 1) the legal considerations, 2) the logistics and 3) understanding how to value those assets.

On legal considerations, assets that both saddled with legal rules or assets that have not legal rules are difficult to digitize. Assets with heavy legal rules are hard to deal with because you have to make sure logic is coded such that whenever a new investor buys an asset, you're checking that they're in the right zone. On the other side, assets that are not regulated can also be very challenging to digitize. Consider a new area of asset digitization like the tokenization of art. There's nothing protecting the consumer buying this asset from being fed false information.

The logistics of managing an investment can also become a limiting factor, for example, from a regulatory standpoint. Let’s say I have to send documents such as tax withholding forms to each of my investors. Because of fractionalization, I may have a thousand investors who made a small investment of $10,000. That means the asset manager has to send out 1,000 documents. It can be done, but at high a cost. Fortunately, this is a transient problem, as we are moving from the old paper-based world full of inefficient forms of communication to the new fully digital world.

Finally, challenges related to valuing digitized assets due to lack of visibility or conceptual understanding make assets hard to digitize. What do I mean by lack of visibility? When an asset tokenizer can manipulate valuations because of information asymmetry. While blockchain is transparent, it is only as transparent as the data that's actually being put onto the blockchain. If an asset owner/tokenizer is able to falsify the data and there is no way for others on the chain to cross-check, then the network does not have a true understanding of the asset’s value.

What do I mean by conceptual understanding? Digitization has opened the door to value and trade assets that previously were not traded. However, that also means we don’t have a full understanding of what that value would be nor what the impact would be. For example, the McDonald's brand name is globally recognizable. It has a value that shows up on a corporation balance sheet. But if you're trying to tokenize just the brand, what defines its true value? Is it the person who's buying this tokenized value? Does the market determine the value? Is it McDonald’s who says, "Hey, my value is X"? We're not sure now, but the market will work these kinds of questions out over time.

2. Considering the amount of logic needed to be put into smart contracts for particular assets, are there some assets that are not worth digitizing? Would it be too costly to put lots of logic into smart contracts to digitize particular assets? Have you encountered some of them?

It comes down to why do asset owners want to digitize? One of the primary drivers is to access new pools of liquidity and sources of capital. If those are already being met in a cost-efficient manner, it doesn't make sense to digitize that asset. For example, in Silicon Valley, it's very popular these days to say you can digitize some startup's equity. But if regulations won’t allow investors to exit this particular asset for maybe 12 months, 24 months, then there's really no sense yet in offering a digitized version of the asset. Just go with a traditional raise. You may have benefited from a transparency standpoint, but is that benefit worth the cost? The jury is still out on that one.

To your second question, there may be complex logic to capture conditional payouts, making the smart contracts built on certain blockchains too expensive to execute due to the gas consumed. I recently did speak with a company focused on cash payments on chain, where the cost of paying out was 2-3% of fee, making the economics not much cheaper than non-blockchain payments.

3. Regarding the obstacles that currently Fusion Protocol faces when identifying events that are supposed to trigger certain transactions. Could you elaborate a bit on this issue?

With any type of event you're monitoring, it's hard to validate what is the truth of that event. What is the source of truth from the external world that should trigger contracts? In blockchain, we recently gave a presentation at Digital Asset Summit in New York on decentralized oracles and why they are so important. In short, an oracle tells the blockchain what is happening with the world so it can trigger events in the smart contract. A good example of a trigger would be when the stock price of Apple actually crosses the $800 mark. That will trigger some type of event. Today, that's easy to do because banks have an agreed upon source of data. They say, "We're going to look at the Nasdaq close as of 4:30 PM." When you move to the decentralized world of blockchain, if you tokenize Apple stock, for example, there really is no market close for this stock anymore. So, what is the close price that should trigger this event?

You say, "Let's all agree that the close price is still 4:30 PM," as quoted on your blockchain. One person sees a close at $800, another person using another data source sees a close at $799. What now? Being able to justify contradictory sources of data is probably the most difficult thing we face when we're trying to identify the events that are triggering transactions. Decentralized oracles are a system wherein a multitude of oracles get rewarded based on how accurate their data is. The system can then aggregate answers from accurate oracles to provide a single source of truth. For example, everybody on the network agrees to look across 50 different oracles and select only the top 25 oracles giving accurate answers. The average of the data provided by these 25 oracles will be the source of truth that triggers all smart contracts in this network.

4. Right now, you're working on obtaining the most truthful information. In terms of validation, if a particular oracle is true or not, do you use empirical evidence? Do you use humans to check up on the oracle?

Yes. We're still working on the exact rating system. It's like a marketplace. The same way as when nodes are competing against each other to try to close the block and post transactions, certain nodes don't seem to be posting the right transactions. They get dinged or penalized because compared to the rest of the nodes, they're not giving the right answer.

Let’s say in my previous example, I had 50 different oracles and 30 of them said Apple stock was $800 and one of them said Apple stock was $750. When the network is operating efficiently, we can say the majority said it was above $800. So, this one that has been substantially below the others is probably not giving us accurate numbers and will be penalized. We're still discussing with partners what the best way is, but initially, you would probably configure an algorithm with some human involvement—similar to semi-supervised machine learning—to fine tune and train the algorithm. Then you let the algorithm take care of itself to rank which of the outliers is correct and which is not.

5. What kind of tokens or coins are the hardest to fuse together in order to create a smart contract across different tokens and across different blockchains to solve this interoperability issue and why?

For us, it hasn't really made a difference which tokens and coins there are because with both our centralized and decentralized solutions, we're able to create mirrors of those tokens within our system. It doesn’t matter what the technology is for the external chains as long as I can bring them into my chain. There's no difficulty with that. I wrap the token in the Fusion wrapper and now can trade it with ease along with all the other Fusion wrapped tokens. We then rely on the external chains’ signature and consensus for their respective coins to confirm any action that is enacted on them. That is the beauty of implementing interoperability at the signature-level.

6. What kind of off chain values have been the hardest or impossible so far to map on chain and why?

The difficulty to access off-chain data is the same problem that any off-chain solution would face when accessing data. If data is behind a paywall or secure login, that makes it very hard to access. If the server's down, sometimes I can't get to it. That problem has nothing to do with whether it is a blockchain solution or not. That has to do with whether the source of data itself can be exposed to anyone who's trying to access it.

7. So far, we know you have main trigger functions: time, event, and transaction sales. Do you think that in the future, you'll need more kinds of trigger functions to satisfy more sophisticated attributes of new digitalized assets?

I think these three trigger events can capture every single more complicated business logic that exists. Tell me any type of transaction you have, and I bet I can break it down into time, event and transaction based triggers. The short answer is, it seems like these triggers are sufficient for now, but in the long term, as we develop more things, none of us know what's actually going to happen. If we require additional characteristics that we want to put into the protocol level, well then, we'll add it.

8. In terms of smart contracts posing a threat of damage or any kind of market disturbances, could you please elaborate on that? In case there is a specific event that triggers smart contracts of many brokers or many security settlers, can that have a domino effect and lead to a stock crash?

That's a good question. I always like to ask, what is the purpose of the smart contract? It is to capture existing logic more efficiently. In the example of a stock market crash, there are already a lot of circuit breakers in place. There are stop algorithms that get triggered at a certain price to prevent these flash crashes. When a trigger is exceeded, a bunch of sell orders can trigger cascading orders. All of a sudden, the stock market is down 20% or 30%, and then it goes right back up. We've seen a few of those.

Now, is that the problem of the smart contract or the fact that these protections exist? I think you'll find the answer is it's because that logic, the protection, exists. The smart contract itself is as much at risk of creating a stock market crash as the existing circuit breakers. The benefit though with the smart contracts is at least people can start seeing the logic that's in these smart contracts. If someone codes an incorrect smart contract, of course, that's a risk. The technology itself is no more dangerous than the legacy processes.

9. As per your personal sentiment, considering human nature, can it [smart contract] aggravate the next crisis? We're humans after all, and with all of this technology more and more penetrating our life and financial markets, do you think the potential of human error increases or decreases?

The potential of humans coding incorrectly is the same as if smart contracts did not exist. Stated another way, if we never had this concept of smart contracts, and stocks are trading more efficiently and a lot more black boxes are being introduced, all the inherent biases of people are already being captured in the logic of these trading algorithms. To answer the question, I don’t think smart contracts would aggravate the next crisis any more than existing trading algorithms would. Do I think the proliferation of automation could aggravate the next crisis? That is certainly a well-known risk.

10. Following the idea of creating more ways to unlock previously locked capital and assets, creating digitized versions of assets that weren't previously available for trading, do you think unlocking these new large amounts of capital pose any threat to financial markets so far?

I don't think the new amounts of capital will cause additional stresses if we include appropriate protections. If you just went and created a fully decentralized world where people randomly issued assets with no protections, that world would be fraught with risk. Think about ICOs as a great example. The market itself worked just fine. There was nothing wrong with the technology. Supply and demand were there. But people lost their money because there were a lot of "questionable” projects in a new and largely unmonitored marketplace. As we're digitizing new assets, I think it's great to embrace the liquidity and all the new opportunities they bring, but we still have to protect the asset originators and investors through proper custody, transparency, KYC/AML, and other legal enforcement.

11. Let's say we've completely migrated to smart contracts. They basically substitute a middleman like escrow brokers and other security settlers. Do you think that the blockchain service providers that develop different kinds of protocols will be those middlemen instead of brokers and settlers? Basically, they would receive the same fees as brokers do now, but they will provide infrastructure rather than handing all of the settlements and other brokerage functions manually.

One of the promises of blockchain has always been that we cut out the middleman so that the asset owner and investor can exchange more efficiently. In actuality, what we're doing is cutting out the inefficient middlemen. If the blockchain innovators and integrators are as inefficient as the existing middlemen, then I think we've all failed. It clearly added no benefit. I don't think an interoperable protocol like ours imposes anywhere near as much cost as a having a bunch of manual processes or costly middlemen trying to settle these instruments.

Furthermore, because blockchain is transparent, you can see what fees are being charged. If any protocol is charging an exorbitant amount of gas fees when a transaction crosses the interoperable solution, that will be seen as a point of disruption by the more innovative protocols. They'll come in and knock that fee down pretty quickly as they take market share away from this expensive provider.

12. In the long term, you do not forecast the cost of this infrastructure to rise?

That’s right. The cost of the infrastructure, if we have learned anything, only gets more efficient. The reason certain regions have been inefficient is because there have been gatekeepers, people who prevent transparency from being introduced. There is no transparent network that is disrupting that market right now. Middlemen that charge exorbitant fees can do so because there's no information symmetry between buyer and seller. If we introduce new tools so a network can exist where buyers and sellers can interact, then the need for that expensive middleman drops dramatically.