The difference between fungible and non-fungible is often illustrated using banknotes and children. Apparently when dropping children off to school and later picking them up people prefer to pick up a very specific child (the same they dropped off).

On the other hand, banknotes are explained to be fungible, as it’s quite the same to you which banknote you will receive as payment or part from when buying something.

Yet, banknotes are not identical. Each one is actually unique, with its own serial number and different physical imperfections. Their fungibility is a social convention — we as society need them to be as fungible as possible so we squint a bit to pretend they are.

To help with that the process of producing banknotes is an industrial process with quality control. The machines that print the notes are made and maintained to a high standard with the goal of producing each banknote as similar to others as possible. No two ones are ever identical, of course, there are always micro imperfections.

If imperfections of physical objects could be accurately recorded they would comprise a very accurate digital fingerprint of the physical object.

Banknotes are easy to counterfeit given access to special equipment and materials, such as those that were used in production.

Yet if the producer could record and store a digital fingerprint of each banknote they produced it would not matter even if counterfeiters could rent and use the same machines on a different day!

The precondition for this to work is for resolution of scanning equipment to be ahead of the accuracy of manufacturing equipment. As long as it holds that

scanning resolution > manufacturing resolution

a digital fingerprint can be recorded (and later scanned) more easily than it can be forged. At the moment this probably means around a millimetre range. Of course, if there were only one millimeter-sized scratch on the surface of a physical object a manufacturer could forge the object given enough time and enough monetary incentive (which is there for very expensive goods).

So let’s introduce another variable: pixel resolution, and let’s clarify the terminology because resolution in common speak means both physical size of a pixel and number of pixels. The formula is then:

scanning resolution * number of pixels > manufacturing price for matching physical imperfections pixel per pixel

Interestingly, this means that unlike the manufacturing processes of old, where the manufacturers’ goal was uniformity of final products, now they will strive to make the differences between units as large as possible as long as the desired functionality and aesthetics are not affected.

The beautiful thing is that manufacturing with tiny imperfections does not cost more, it probably costs the same or even a bit less than when going for maximum uniformity.

On the other hand, forging a physical object does cost more the more imperfections there are to forge. In fact, given that objects are three-dimensional and each imperfection has three coordinates, forging probably becomes prohibitively expensive even at a rather small number of small imperfections.

We are there! The lidar technology is becoming cheaper and coming into mobile phones and it allows for scanning with millimetre resolution and recording a large number of 3D data points about physical objects. This means that for physical objects of sufficient complexity manufacturers can already record digital fingerprints that customers can verify on their phones and that are too expensive for forgers to forge.

A threshold has been crossed once 3D scanning came to mobile phones. As technology progresses it is likely that high-resolution scanning equipment will maintain an edge over high-resolution manufacturing equipment.

This is great news for consumers! When buying a physical product you mostly don’t care which exact one (it’s fungible to you), yet you care about provenance. That you get the original product will be ensured by leveraging its non-fungibility.

A great example of how intertwined the concepts of fungible and non-fungible are.

Let’s start with an example. According to the Court Examiner’s Report, the (now bankrupt) crypto lender Celsius spent over $500 million buying back its token CEL, for which it had only received $32 million during the ICO. According to the same source at least $75m of the $500m was upper management dumping their tokens. That would have been a horrible deal even if the company survived.

But while the CEL token was probably not the reason why the company failed (BlockFi failed too and didn’t have a token), the token was in many ways toxic for it. That is by no means specific to Celsius. Below is a list of ways having a token harms every project that has one:

1. The most obvious one — due to the US regulators’ overreach, token sales to US citizens puts the company into permanent danger of being investigated for breach of securities laws.
   
2. In the absence of tradable shares, company tokens assume some functions which shares normally have in the stock market. Namely, they are seen as a health gauge for the firm. As per the Efficient Market Hypothesis, share price is assumed to incorporate all the publicly available information about the company and its prospects. If there are good news, share price rises, if the share price is falling it must be because something bad happened, or is about to, and the insiders are dumping.
   When the company’s token is falling more than the market in general, the market assumes bad things are happening with the company. This happened with CEL and was one of the reasons why the company was spending its scarce resources defending the price of the token; it was aware of its signalling function. (Preventing a sharp drop of token price is especially important for the type of company that is vulnerable to bank runs, but not just those.)
   
3. The token pits insiders against the unsuspecting public (huge information asymmetries there). Insider trading is inevitable. Even when the odds of it being prosecuted are low, the setup is morally problematic.

4. The token allows investors to get out early by dumping on the public, and if the public is not there to buy, on the company itself. It thus reduces the incentives for investors to be long-term committed.
   
5. The token distracts employees who are inevitably constantly checking and discussing the token price, and whose mood and job satisfaction can also fluctuate wildly with that price.
   
6. It distracts the company itself because it has to figure out things like market making (a paid service that is very expensive and extremely hard to get right), exchange listing (a large lump sum payment), and so on, instead of focusing on its core business.
   
7. It incentivizes the company to mark-to-market the tokens it has in treasury, artificially inflating its balance sheet, creating a moral hazard (the management can take on more risk because of the strong balance sheet).
   
8. It creates a bunch of fanatics in Telegram rooms that constantly complain about the token price, even if it is rising but not rising fast enough. These are dubbed “The Community” and are in reality as much a burden on the company as they are a marketing tool. 

The list just never ends. A token is like fire, a good servant but a terrible master. It can raise funds, spark interest and ignite early adopters, but it threatens to consume the whole enterprise and burn it down, which happens more often than not.

The invention described below has been inspired by the latest hare-brained scheme I have read for a project token. Again, it starts with incantations of Utility. Again, a New Vocabulary is invented, of Actions and Rituals that will be performed with the token, decreasing the dreaded Supply and providing the cherished Demand. You are excused to be confused — by that point to confuse you is the point. A religion which is understood without middlemen is a very primitive one.

What if… we cut the crap?

I am smart enough to understand that I am too stupid to really understand tokens. Tokens… have some magical properties, which are perhaps not even penetrable with reason alone.

And yet, with appropriate humbleness, we can reason about some of the underlying forces and emergent phenomena. And perhaps we can think of better ways to accomplish our goals.

Tokens are used by teams to raise funds for a project. They are sold for a price. This is happening even though the price should be zero because there is no legal obligation for the token issuer to ever share profits with the token investors if the project is successful.

Why is the price not zero? Because the market treats the tokens as pseudo-shares of stock. But unlike with shares, where the bet is on the success of the project, with tokens it’s a twofold bet: both that the project will succeed and that the founders will be honorable enough to share profits even though they don’t have to, legally.

Interestingly, when a project has both token and equity there should be a liquidity premium on the token, but a legal premium on the shares.

To summarize: tokens are, in effect, tokenized honor of the founders.

You raise funds. Your promise is: if we do well, you, our investors will do well too. There has to, eventually, be a hard link between profits and tokens. The cleanest endgame would be to convert the tokens into equity, in a jurisdiction, awarding the holders all the benefits and legal protections that come with shares. That may prove hard, even with good intentions, so other ways of paying back investors are being experimented with, most prominent of which is using profits to buy back tokens and burn them. It’s a rapidly evolving area in which we’ll see more innovation, no doubt.

But for each honorable project that tries to make sure their early investors have profited, there are dozens that have abandoned their token, that don’t care, or that run around in circles trying to prop up the token price artificially using mechanism design and supply/demand manipulation, instead of just calling it a share and letting the market price it as one. How do we solve this?

Two steps:

1. We declare project tokens “non/supra-jurisdictional shares of stock”, governed by common-sense understanding of profit sharing and the ancient rules of honor.
   
2. As mentioned, tokens are magical, so we use blockchain magic to enforce this, using Soul Vessel NFTs (ERC-721-SV).

At the beginning of the token offering, the project founders (who can remain anonymous), accept and declare that tokens are shares in the project and that they will fully partake in the fortunes of the project, one way or another. It is recognized that the project may be far from profitability, and far from incorporating and converting to real shares, but it is promised that if there is ever profitability, tokenholders will share in it in the same manner shareholders would. Rules of governance are also established (whether tokenholders can do other things as well).

A special NFT collection is minted. In it, each NFT represents a project founder. The NFT has two states, “locked” and “unlocked”, represented by appropriate graphics (say, a bird in a cage, and a free bird). The birds represent the souls of the founders. They are initially locked, and the only way to set them free is to send an appropriate amount of project token to the NFT smart contract’s release() function.

Effectively, this forces the project founders to buy back (and burn) almost all of the project token in order to set their souls free. To account for things like lost tokens, people who don’t wish to sell, and token price rise, there are two criteria. Either an amount of tokens needs to be burned, or an amount of ETH (USD?) at least equivalent to the amount initially raised needs to be sent to the contract, which will spend them on Uniswap to buy back the token and burn it. A Time Weighted Average Price (TWAP) mechanism is used to prevent manipulation.

What if the project has failed? Well, in that case the token will be cheap and the founders will not have a problem buying back the tokens and freeing themselves. This has an interesting side-effect… by not dumping the token of a rug-pull project, the investors can make it hard for dishonest founders to set their souls free.

The power of this solution is that it uses magic. Typically, you want founders to be doxxed so that their reputations suffer if they behave dishonestly. But we wrote that founders can remain anonymous. The threat of their souls being captured on the blockchain forever unless they behave honestly is a terrifying threat indeed, and a strong incentive for honest behavior.

It’s a familiar thesis: blockchains are ideal property registers because they are fully open, easy to access, impossible to take down or censor. Also, transferring blockchain assets is trivial and independent of traditional third parties like banks and lawyers.

So why hasn’t it caught on? Multiple startups have been on the problem since (at least) 2017.

There are various UX challenges that need to be solved in clever ways but the most difficult of them is:

• Irrevocability of blockchain transactions

(Note: having-to-approve each transfer and being-able-to-revoke are flipsides of the same coin, but we argue here that the nuance is important and that the latter is the more important thing, which, if it works well, allows us to live without the former.)

States will never accept that transfers of property titles can’t be revoked. On Ethereum there are several existing standards for enforcing transactions. They solve the revocability problem from the technical perspective by allowing an authorized third party to force transfers whichever way (see ERC1400, ERC1404 and related). The approach taken there is technically sound.

But the problem of the authorized third party accessing the chain remains. The ultimate arbiter of property rights is the state (via courts, or, at a lower instance, regulators/agencies) and it has to remain the ultimate arbiter in such a way that it can override any state of the chain at any time, including its own decisions (for example a higher court deciding on appeal).

There are good reasons why it is not possible to simply deploy an on-chain register and expect the government to start using it.

• If the courts are expected to use Metamask to call forceTransfer(), it is simply not going to happen. For one, because courts are run by old people, secondly, they have zero incentive to do this.
• Even if they would be forced to do it (via statute change), private key security would be a hard problem. What if the court’s private key is stolen?
• Multisig schemes could help with the above, but still have their own issues with key loss, theft, etc.

No government would commit to having the only property register controlled by a very young technology without being able to ALWAYS override it, no matter what some blockchain or smart contract wallet thinks of it.

Proposed solution

What if we can ignore

• statute/regulatory changes
• court wallets
• multisigs
• old-judges’ onboarding altogether

for now and use Oracles to communicate real world events (court decisions) onto the chain permissionlessly, in a way that cannot be manipulated?

This would allow users to use the blockchain as a defacto register, in an optimistic way, not waiting for a positive court decision for accepting a transfer, but knowing that, if there is a need to revoke a transaction, this can/will still be done using the regular court system and reflected on the chain subsequently. This means that time can be used as a proxy. An asset that has been sitting at an address for a while, and hasn’t been challenged and reverted, is very likely legit and the owner most likely has full property rights to transfer it on to you as prospective buyer. Especially if the period defined in the statute of limitations has expired.

The government remains in charge and does not have to do anything.

How would it work?

A smart contract that controls the assets would respect a decision from an Oracle network that aggregates several reputable news sources to interpret court decisions. We are not interested in all decisions, just those that nullify, revert or force property transfers.

• Maybe just using the official court gazette would be enough, or combine with mainstream media.
• Mainstream media like NYT, WaPo, etc… do not have a standard for reporting court decisions… so this cannot be fully automated, Oracles would need human input.

The main issue is how to solve the game theory so that Oracles never collude with malicious players to censor from the chain a court decision that they do not like. There are also important questions about Oracle onboarding and offboarding, etc.

For that reason the Oracle network would not be built from scratch, but would utilize a known system such as Chainlink Oracles or Eigenlayer. The Oracles would be putting up a stake that would be slashed if they are dishonest and would be earning a fee on every property transfer. The game theory would borrow heavily from the experience of PoS systems like Ethereum.

Optional Attestations

Each property-representing NFT could have optional (equipable) attestations attached, as additional NFTs. These would be normal legal attestations, for example, by a legal office that claims to have verified ownership in the cadastre books. The number of these attestations would not be limited and anyone could request them (both the seller, and the purchaser). But they are optional in this system — the main confidence comes from time-as-proxy.

Why retail would love it?

The story of why tokenizing property is amazing can be found on the web page of any of the multitude of Security Tokens startups. But property-as-NFTs brings another very valuable innovation: Binding Permissionless Offers. This means that anyone from anywhere in the world can send a property owner an offer to buy the property, or offer a loan against it. The same infrastructure that is being built out for this right now for JPEGs can seamlessly be used for all other types of property, once the issue described above is solved.

“Monetary value is reflection of mindshare” — https://mobile.twitter.com/ash_bhimasani

Price is an objective expression of subjective preferences of the buyer and the seller. If prices are subjective, they can be anything. Yet, can we reason about them?

We propose that NFT prices are narrative-driven, that these narratives can be very personal, but also shared, and while the number of possible narratives is infinite, the number of shared ones is necessarily low. In this document we attempt to enumerate various shared narratives we’ve spotted in the wild. The thesis is that the more narratives an NFT belongs to, and the better it fits them, the more expensive the NFT is likely to be.

As a way of contrast, here is an example of a personal narrative: pink-haired PFPs. Such narratives are amazing and important, but they are the long tail of demand. They mean there will always be some buyers, but they’re not what is driving prices up.

Below is the (work in progress) list of shared narratives, with some examples for each. You will notice that some are post-hoc, and may seem tautological. But that is fine. If one has paid a lot for a jpeg based on fear of missing out, they then have to come up with a justification. The justification becomes the narrative.

1. Absolute age

This refers to how early the NFT collection was minted. The older the better.

Examples:

• Crypto Punks (the oldest collection, by absolute age)

2. Relative age

When the collection belongs to a category (there are many similar collections), it pays to be the earliest / first-of, irrespective of absolute age. First collection that does this or that.

Examples:

• Crypto Punks (the oldest profile picture collection)

3. Dedicated to an identity group

A collection can find value by being dedicated to a group of people, defined by some characteristic (sex, ethnicity, occupation, hobby…). The narrative is that it is dedicated to, celebrative of, etc.

Examples:

• World of Women (first collection dedicated to women)
• Crypto Chicks (collection for women in crypto)

4. Establishing an art style

This is a post-hoc narrative. Collections that are seeing their art massively copied are thereby proving they have memetic potential and staying power.

Examples:

• Doodles (a simple, but memorable and massively copied style)

5. Ground breaking technology

Similar to establishing an art style, an NFT collection can be innovative on a technical level, for example by inventing a new way to generate art from data, or store art on chain.

Examples:

• Terraforms (fully on chain art)

6. Convincing roadmap and credible team

Evidently, many NFT collections are in fact fundraising for a business. In such cases the market can value the NFTs in the collection highly if it believes 1) that the team is capable enough to execute and 2) that the team will find a mechanism to share profits with the NFT holders. Like any other business, but with added uncertainty about the payout mechanism.

Examples:

• Any team building a game

7. Famous artist

If an already famous artist produces a collection, it will have value because of the artist’s fame.

Examples:

• Snoop Doggy Dog’s collection

8. Ownership history

If an NFT has been owned by a succession of interesting people, its ownership history can be the dominant part of its narrative.

Examples:

• None at the moment

9. Strong meme

If the collection is built around a meme or a joke, and it’s one that resonates with many people.

Examples:

• Crypto Dick Butts

10. Useful for virtue signalling

A collection can have value because owning an item from the collection signals virtue. Examples:

• Unicef NFTs

11. Useful for sophistication signalling

A collection can have value because owning an item from the collection signals sophistication.

Examples:

• Art Blocks (and generative art in general)

12. Useful for signalling wealth

A collection where items are very expensive can derive value from the fact that rich people will buy it to show off wealth.

Examples:

• Bored Ape Yacht Club

13. Was already famous once

A collection can derive value from the fact that it was already present in the collective imagination once, so it has a foothold, and it is plausible that it will be able to stay there or regain position if it loses it temporarily. It is plausible to think that “slots” in the collective imagination are limited, there there is a certain (very low) number of collections most people in the space (or most people in general) can be aware of, and that once a slot is won by a collection it will have inertia to stay there. See this tweet thread for an elaboration of this view.

Examples:

• Pudgy Penguins (they were written about in the mainstream media)

14. On chain permanence

A collection can derive value from the fact that the artwork is stored completely on chain, which is technologically challenging, expensive, and rare. Typically, the more complex the art, the harder it was to put it on chain, the bigger the flex.

Examples:

• Terraforms (fully on chain art)

Private/owned/controlled blockchains come under many names on many different platforms but a good working definition is: any blockchain where there is an entity or a small set of entities who can plausibly censor the network.

Examples are corporate internal blockchains, as well as inter-corporate (say, a blockchain for banks) or government-controlled blockchains.

Such blockchains defeat the purpose of blockchains because they can be censored, so whether they will survive has remained a legitimate question since they were first proposed.

We propose that they might, but that the reason why is not related at all to the reasons why they are typically introduced and how they are marketed.

They are marketed as bestowing on the (usually corporate) entities that introduce them some of the aura of public blockchains and the character traits they are associated with: trustworthiness, clarity and transparency, immutability, solidity and robustness. 

It is extremely unlikely that anyone is being fooled and that any of that matters at all.

But there is an avenue through which controlled blockchains might still survive and prove useful, and it is because blockchains have travel companions, and the foremost among them is an ethos of sharing and collaboration, and quality, often open-source, tooling that is borne out of that.

Corporate-owned databases like Oracle and IBM DB2 suffer from astonishingly lousy software stacks. Those behemoths are known to buy up companies that build apps on top of their databases, which their corporate culture subsequently kills, strangling them slowly and draining them of vigor until the founders leave and they’re effectively dead. Big corporations are where software goes to die.

That’s it, that’s the thesis. Blockchains, even though in almost every way inferior to legacy databases, might still win out simply because the accompanying tools and apps ecosystem is growing and evolving rapidly and has probably already eclipsed legacy-database-oriented tooling in both quantity and quality.

NFTs are mostly short-term investments, with buyers hoping for a quick buck, but to explain why NFTs have a price we have to dig deeper and visit the weird subset of NFT buyers who are not investors.

To say that NFT price is market-determined, and driven by narratives, is not controversial. But there is something very peculiar about it, very different from how coin and token prices are also driven by narratives.

Bitcoin, for example, is mainly driven by only two narratives: means-of-payment and store-of-value. An ERC20 token issued by a team is driven by the story of why the enterprise will be a success. That is but one narrative, or, at best, several pitches for the same thing.

NFTs are different because the number of narratives an NFT can belong to is… infinite. Here are two examples of people ignoring the main/official narratives of the NFT collections (the ones provided by the issuers) and creating their own narratives instead.

The first example is a (or the?) pink hair vault. As the name says, the collector seems to be collecting pink-haired NFTs across collections.

https://opensea.io/pink-hair-vault

The second example is a “GSR blue vault”. Not as focused and specific as the collection above, here the collector seems to be buying blue-chip NFTs with something blue in them, in order to match the name/theme of their company. In this case it seems a bit of a flex, as it’s an investment fund. But the goal does not matter, it’s still a personalized narrative.

https://opensea.io/GSRblue

These are the first two found on a quick search, but there are many others.

How does this relate to price? Well, it’s not a direct, but a roundabout connection. The speculators and the investors, the shared-narrative buyers, are providing the upward pressure on prices, but it is the private-narrative buyers that provide the bottom support. The price of an NFT will almost never go to zero, because to someone, somewhere — it tells a story.

Here is an attempt to define what is an NFT.

Arabs have a saying about everything coming and going but the pyramids being there forever, defying time. The blockchains are the pyramids of our time. The internet was built to survive nuclear war by using stateless protocols that enable the network to survive even with many of its nodes being down. Blockchain is a step further… humanity’s attempt to create a verifiable state (of information)… such that will survive even the internet being down. For it only relies on the internet for convenience, but does not depend on it for consensus.

NFTs live on blockchains, thus they are graffiti on the pyramids, scratches and scars on something that was designed to live forever.

In the most abstract terms, they are just messages from the originator to the world, inheriting all the qualities of digital signatures (authentication, integrity and non-repudiation) plus the unique innovation of the blockchain (prevention of double-spend via a shared ledger).

Such messages are a very strong foundation for property transfer, and indeed that is why blockchain was invented in the first place, for Bitcoin transfers are just random messages, carved into the everlasting blocks of Pyramids of Time, which we choose to interpret as reassigning ownership of Bitcoin. Such is the power of those messages… that people are willing to pay real-world cash to receive a message saying that they now own a part of an imaginary thing… that does not exist anywhere except in the messages talking about it.

Strong Messages indeed. The price of Bitcoin is the current consensus price for being mentioned as owner of a worthless thing on the Forever Ledger, the most durable store of records ever to exist.

Soon after it was understood that, while writing on the Time Pyramids might be hard and expensive, they are also indestructible, all sorts of conventions sprung up about how to word messages for a particular purpose and how to interpret them. NFTs are messages that say:

“To the last recipient of this message, the following belongs: ________”

Anything can be on the dashed line. An image. An amount of currency. A car. A house. Technology cannot enforce how a message will be interpreted, nor whether it will be honored. That is up to humans. But it can, and does, enforce that the message itself cannot be forged and that only the last recipient can send it onward.

Thus the definition of NFTs is simple. They are conventions about messages, carved into time itself, that are meant to let them behave like pointers to bearer assets (assets that can be transferred onward by their last recipient, which we dub the owner).

Clearly it is all imaginary and a matter of convention. Yet the durability of the messaging medium gives the property records on it a certain quality that other property registers will never be able to match. Human life is short and he is a feeble creature that changes opinions and beliefs as winds blow. But what is written on blockchain will remain written — and it will anchor and shape our future.