Theoretical physicist and artificial intelligence/blockchain expert Stephen Wolfram believes in a computational universe.
From Mathematica, the programming language that Wolfram wrote and designed, to Wolfram|Alpha, the answer engine that underpins Microsoft's Bing and the Siri artificial intelligence in your iPhone, Wolfram has built his career on the premise of a computable world: that if you break knowledge into easily digestible bits and pieces, you can model anything—from financial markets to musical theory.
ETF.com recently spoke with Wolfram at Collision, one of the fastest-growing annual U.S. tech conferences, held in New Orleans, May 2-4, 2018.
Wolfram was there to give a talk on computable or "smart" contracts, or contracts written in computer code that could be distributed on a decentralized blockchain network, without the need for any third parties. We asked him what impact such contracts might have on financial markets, where billions of contracts are traded every day.
A lightly edited transcript of our conversation follows:
ETF.com: The idea of a "smart contract" is fascinating. What sort of applications could it have in the financial markets?
Wolfram: In the world of finance, it's already happened a bit. For the last, I don't know, 30 years, options [contracts] have been expressed in essentially an algorithmic way.
But mortgages haven't. So, if you take this multipage mortgage document and express it in computable form, then you can take 10,000 and do systematic analysis on them.
That's a rather sterile example. But there are plenty of other examples; say, whether you’ll deliver 10,000 pounds of grade-A bananas to this site at this time.
ETF.com: Right, I mean, that's essentially a futures contract.
Wolfram: Yes, and those are very stylized contracts. But when it comes to delivery, the question is, were they really grade-A bananas? Or were they rotten?
Today one would write a contract that says, "According to the USDA classification of bananas, they need to meet these criteria …" But when the contract becomes computable, a machine-vision system can be set up that’ll look at these bananas and decide if are they grade-A or not.
ETF.com: So computable contracts could mean greater accountability for both sides of that contract.
Wolfram: Yes. Also, automatic executability. Now, a human inspector has to show up and say, "OK, that banana works; that one doesn't work," and so on; as opposed to just a device that's watching these things. So the efficiency of commerce gets higher, if the contracts are computational and can be executed automatically by machines, rather than having to involve lots of humans in the loop.
Of course, not every computational contract will need a blockchain.
ETF.com: However, it's not too hard to see how this could be used for, say, crude oil futures, or any sort of hard asset that needs to be assessed and warehoused. Settlement takes two or three days now. What I'm hearing from you is that computational contracts could potentially speed that up.
Wolfram: Yes, I think so. But in each different industry, there's a different set of things that take time and that involve humans, etc. The trick is how to encode those things in such a way that computers can execute them automatically. But I think we can actually start to do this stuff now.
Except, if you look at what actually exists on blockchains right now, the big missing link is "smart oracles."
ETF.com: What’s a "smart oracle"?
Wolfram: It answers the question, "What actually happened in the world?" for the blockchain. It's that link between the real world and the blockchain.
For example, I've got this parameter in my abstract contract, and I need to know, "Was the temperature in Timbuktu above 90 degrees yesterday?" That's something the blockchain doesn't know. So how do you find that out? That's our thing.
The Wolfram|Alpha system, which empowers Siri and so on, is basically the only example right now of a broad, multidomain source of computable knowledge. The people who try to build these oracles for blockchains basically all use our technology stack for that.
ETF.com: What kind of input questions could a smart oracle handle?
Wolfram: "What was the weather yesterday?" But there are also things like "Did this IOT [internet of things] device record this particular measurement?" Or "Was that plant watered yesterday based on our soil moisture measurement device?" And so on.
ETF.com: Let's get back to the financial markets. Do computable contracts have trading applications, beyond just improving the behind-the-scenes for futures markets?
Wolfram: In terms of financial instruments, once you’ve made sort of the execution of a thing computational, you can take buckets of those things and start to have markets of them, which is what happened in the case of options.
ETF.com: Is cost the limiting factor, though?
Wolfram: Well, take weather insurance. You buy a $100 contract for weather insurance. If you actually want a global blockchain for that, if you really want tens of thousands of people to stick your data on their computers, it's not going to be cheap. And it's also pointless, because nobody cares. Just have some cloud-based provider transact those things, right?
The other end of the spectrum, though, is people doing $10 million transactions. These people are, realistically, not going to want to put their trust into the trust-less world of the blockchain, so to speak. They're going to want the legal system and insurance and all those kinds of things around their contracts.
ETF.com: So where is the middle ground?
Wolfram: I don't know. But I do think there is a middle ground. Maybe it will be permissioned blockchains within individual industries or within groups of companies.
ETF.com: What’s a "permissioned blockchain"?
Wolfram: When you talk about bitcoin or Ethereum or something, those are public blockchains that anybody in the world can connect to and put things on. But if you have a consortium of companies that are all, say, involved in the same particular supply chain, then you can have a blockchain those companies can put things onto. It's a private, not public, blockchain.
And if you do something with, say, 30 companies all contributing to this one blockchain, they can all have a copy of it. It's a really pretty decent software solution to the problem of maintaining a ledger for lots of people.
ETF.com: Seems like this could have application in, say, clearinghouses. The Depository Trust & Clearing Corporation and stock exchanges could use it to track who owns what, when and where.
Wolfram: Yes, that's a place where people are looking at this. Banks, too. But the ecosystem of computable contracts doesn't exist yet. We are kind of in this strange state, because it's not worth it to build [the ecosystem], either, if it isn't being accelerated by lots of money being in that space.
ETF.com: What sort of limitations do smart contracts introduce that don't exist in the status quo right now?
Wolfram: You lose wiggle room. Instead, it's basically, "Well, there was this piece of code we set up, and it ran, and these are the consequences." In a sense, it's more like dealing with the natural world. You can ask, "Why did the tornado turn left?" Well, ultimately it's the laws of physics that determined that. It's not like a human decided that.
Contact Lara Crigger at [email protected]