What if there was, in the future, a network that is bigger than all the networks we know today? A network online 24/7 , 365 days a year. A network making what was impossible possible. A network securing trillions. A network powering some of the most useful and used applications. A network competing against itself for a share of the future. A network of truth.
What if there was, a network of networks? A monopole paradox where no network can compete but everyone can join and compete.
How much would you value this network?
The Oracle Problem
The oracle problem has been discussed numerous times in the last years. Often oversimplified to how you get real-world data into a smart contract, the oracle problem itself is a lot more complex. Sending data isn't going to cut it or solve the problem, as the problem resolve around the truth (real world data) and a user consent to the context (the smart contract preprogramed terms and conditions).
Truth is the most complex element of the oracle as it's hard to define and agree on. It's been one of the largest central subject of philosophy for thousands of years. Is it a belief, a proposition, a declarative, a state, a representation? Perhaps a fact or a reality that everyone agrees to it? Can the truth change in time? Can it include something that is unknown? Does it need to be both sound and complete at the same time?
Can an oracle be aware it isn't providing the truth before providing it? Can an oracle learn from another oracle it lied? Can an oracle without possessing complete known be sure it hold the truth? Is it possible for an oracle to trust itself it possesses the known? Can an oracle just copy answers from another oracle?
The problem isn't around hardware or software, consensus or trust, inbound or outbound, human or machine. The problem is about the algorithmic complexity, randomness, and probability revolving around the truth. Laying in a grey zone between philosophy, social agreement and algorithmic information theory, one could argue that the truth is in itself, an unsolvable problem.
How do you then, come to solve the unsolvable? The answer is at the pages 10 and 11 of the 2017 Chainlink Whitepaper: Why the ideal oracle (ORACLE) is hard to achieve. Achieving properties as close as possible as to those of oracle by approximating the truth under realistic trust assumptions, thanks to a game theory inside a distributed system, aka an Oracle Network.
If the, at present, best representation of truth is an overlap in between reality and social agreement, then a game theory between oracle themselves to approximate and agree to this truth is the best that can be achieved and agreed on inside a given context. Since an approximation is derived from a game theory, then the more competitive the game theory is, the better the chance to hold the truth. Therefore the more competitive the network is, the more valuable it is.
If an oracle network that comes to the best approximation at a specific time for a specific context hold the truth (e.g. 1), then anything else that doesn't provide 1 is worthless.
This is the most important part. The truth, can't be improved, iterated, or have substitutes. Only duplicated.
If you don't hold this truth at a specific time for a specific context, then the rest doesn't matter. The product, the technology, the team, the community, the memes, the CEO , the investors, the partnerships, etc. Nothing matters, as you don't hold or get to decide the truth.
One could argue that a variation in the truth could still be agreed on. Yes if the context (e.g. smart contract) allows it and is socially (e.g. users) agreed. But as we move forward in time and oracles start to be used for larger applications, agreed by exponentially growing user base, and secure trillions of dollars in value, a difference of tiny fraction could be too impactful to handle.
Simply put, the toughest game theory always wins, and the rest is useless.
Oracles are an extension of a context (e.g. preprogramed terms and conditions).
For dApps, they belong to the trigger layer and the oracles and play a crucial part in the execution.
A faulty Oracle trigger a false positive or negative vis-à-vis the preprogramed terms and conditions that were expected and agreed on by the parties to be the truth, (e.g. developers and users)
The key here is agreed on by the parties. If the truth-settlement layer of an oracle (e.g. aggregation) was to change, then it should be clearly stated to all parties and re-agreed.
As oracle solution get adopted, the real problem will arise when, a tiny change in the aggregation or a fraction of difference in one source would have meant a totally disparate outcome for a user (due to a different truth, as subtle as it could be).
To this end, used oracle and (and by extension their aggregation contract if existing) should be clearly stated to end users to make sure they agree to this truth/trigger layer.
Ultimately, the aggregation-settlement layer is the trustworthy-agreed experience, the bridge toward real adoption.
The toughest game theory
There can be many oracles. But only one ground truth. If one says 1 and the other 1.00000001, for a specific time and context, one is right and the other is wrong. But both can say one or the other and be right.
Everything else equal, two oracles providing the truth 1 will both hold the ground truth. But since the trustworthiness comes from the toughness of the game theory, choosing the less trustworthy oracle always involve more risk. The period of time it can provide the truth is and will always be undefined in a timeline that isn't finite.
The nature of the output (the truth) makes it important to have the toughest game theory. To best way to do so is to have competing pillars that upside themselves.
Example, number of users, number of oracles (competitors), reputation, collateral, accuracy, uptime, hosting coverage, etc...
One (maybe the only) way to outpace the toughness of a game theory would be to add a pillar or replace one by a superior one; providing we can weigh the pillars into pro-rata of toughness. Toughness then becomes a subjective territory but the number of existing pillars is finite thus at some point you exhaust the toughness possibilities and reach a tipping point.
The more tough the game is, the less we have to fundamentally assuming that oracles are going to fail.
Network effect as a moat
Network effects and the Metcalfe's law aren't to be explained anymore. Network effect is the source of success of some of the largest and most impactful business, products and projects. A straight line up building a larger moat as time goes by.
An oracle network is that. Oracles incentivized to compete against themselves to form/confirm the most approximate truth there is. And the most approximate truth attracts incentives.
One could argue that as market gets bigger, it becomes harder for the winner to stay in a winner-take-all mode and keep their moat. That large markets are now being split among multiple players. This is true for various reasons. The most common explanation is, as the bigger the market, the more niches opens up to competitors and customers/users needs vary increasingly with it.
The thing is, as we stated before, there is nothing that can replace the truth, it isn't something you can adapt, change, improve to get new users/customers or tap into a niche.
Again, the truth, at a specific time for a specific context, can't be improved, iterated, or have substitutes.
What's interesting here is the irreplaceable nature of the denominator in the network effect and the self-improvement of it. This mean, as an oracle network becomes the first mover to secure incentives and build the truth, it can't be replaced in time.
An oracle network reaching the tipping point in toughness would form an ever-growing monopole. Something that can't stop improving itself and can't be replaced. A network of oracle networks push things farther, keeping the truth away from being a centralized point of failure that could be targeted by regulators and lawmakers. Making sure truth from approximation is always available.
If everything is worthless without the truth, then what is the truth worth? Can supply and demand equilibrate to a fair value? How do you measure both?
At face value you can use the cashflow/reputation, providing reputation can be denominated in something tangible. A substitute would be cashflow/collateral, but does trustworthiness absolutely require a collateral?
Then, what multiple would be fair?
Just as the oracle problem is unsolvable, the oracle network might be unvaluable.