The Future of NFTs: Cryptographically Capturing Their Histories

Atomic Form is creating a full-stack solution for NFTs: from data management and provenance to hardware and exhibition. The topics in this article include media file store redundancy, metadata tagging beyond minted data, interchain record keeping, and exhibition data. Atomic Form is largely addressing the following:

• Creating superior data granularity to enhance record keeping, search, and discovery.
  A P2P economy for data availability, provenance, and exhibition with AF’s OS and Hardware.
  Creating hardware solutions that fully integrate with this superset of data.

Part 1: On Stories as Metadata

Pop Art was a pioneering artistic movement between 1950-1960. The works of that era are about 60-70 years old, and their records provide an important basis for contemporary cost estimates at auction. One famous artist from that era is Andy Warhol, who, throughout his career, has influenced the price of artifacts by simply once owning them and producing artifacts with illustrious stories, influencing their worth.

Let’s begin with Warhol’s ‘Shot Marilyns’

The “Shot Marilyns” was once just… the Marilyns. Andy Warhol printed five of them and stacked four against the wall at The Factory. Dorothy Podber, a performance artist, had asked Andy if she could shoot them. Andy thought she meant to take a photograph, so he obliged. Dorothy instead pulled out a revolver and shot them with a bullet, famously denting each print with varying degrees of damage. One of these shot Marilyns sold this year for $200,000,000.

This is a good and unseeable example of a small dataset from an influential artist undergoing a contextual modification that differentiates specific pieces from a larger body of work despite having similar subject matter and mediums. Sort of similar to baseball error cards. Today, we may consider “Shot Marilyns” to be five different token IDs from the same minting contract where four were exposed to some vulnerability or other unique event that influences their worth or makes them ultra-novel.

In 2090, people searching for iconic works of the early NFT movement will look to works minted in 2017-2022, just as pop art collectors look to the 1950s now. Their stories will influence the works, just as their owners will. Their stories can now be recorded through a series of decentralized blockchains and public centralized databases, influencing a new degree of provenance and information efficiency. By doing this, we can take advantage of owners' and creators' cryptographic signatures to make records on hashed data structures.

Consider in 10 years, you are presented with a variety of works from the 2018-22 era: would you value them differently if they were the first NFT sold at Sotheby's? Would you value them more or less if used for a marketing campaign? What if they were owned by somebody famous? Any collector, archivist, or historian knows the answer: there are unique circumstances that can create legendary artifacts.

Part 2: On Web2, Web3, and Distributed Filestore Data Combination.

The ability to own an NFT makes some assumptions about the reliability of the blockchain network's state and the off-chain storage locations where the media “lives.” NFT critics often note that you don’t “own” the media. This is kind of true. Practically speaking, you own a token that is a bearer bond to the data, as issued and assured by the contract's publisher with various addressable content and attributes. Figure 2 below helps illustrate the breakdown between an NFT’s on-chain and off-chain data.

Atomic Form seeks to maximize the redundancy of digital object indexing as well as enhance reliability and search through their categorization and provenance using IPLD blocks (see Figure 3). While Ethereum has blocks that must exist on its own, IPLD is a network-neutral settlement layer for interchain data. This means IPLD blocks can be distributed across any/all filestore systems. We do this with a number of partners who we can announce in the coming months.

There is a misnomer that Ethereum (or any public network) is immutable. In fact, the blockchain state is always at risk. Adding more trouble, tokenURI data is inconsistent, and its dependence on external databases for file storage (for media and other data) may not work out well in the long term. Finally, there is not yet a consistent methodology for adding arbitrary tags and incomplete data to NFT’s that go beyond publication (just like the story of the Shot Marilyns). The most consistent follow-up question to working with private collectors and galleries for exhibition solutions is “how do you know?”

This is a good question. In many ways, the Atomic Form Lore Framework is built around knowing. Built around matching IPFS/Arweave hashes with Ethereum hashes. Built around matching user signatures blockchain state. Built to provide proof and tying together shreds of data from various places for various purposes in IPLD Blocks that can link parent data to other Blocks from other networks.

As we have analyzed before, in some collections, there is a significant dependency on IPFS and some media links are simply dead. NFT owners hold the token in this instance, but the visual component (media) has expired only 1-3 years!  How will people trust their assets, will they still be there for a few decades later if we can’t sufficiently store them for a few years?. Despite all the dead links, you continue to hold that token, the bearer bond for data. The media files of these dead links can be stored and preserved by artists on other filestores like Filecoin or a renewed IPFS link.

For the avoidance of doubt on how NFTs work beneath the hood and on the fragmented nature of relatively high-value assets, let us consider Psychoprogeny - Mother Tree by Dustin Yellin:

The beautiful piece holds the data below as per the Christie’s website, which is a source of truth for this auction. Indeed as Christie's notes (see Figure 5 below), the contract address, ID, and address of the current owner are on chain. This data will not change without a 51% attack on ETH. The final sale price of $60,000 is off-chain and Christie's can become an oracle for that information on a blockchain, like the ones Atomic Form is making with the generation of yielded Metadata, Events, Multimedia, and other data including but not limited to analysis or tags. Furthermore, the tokeURI data is nestled inside the token ID and is not discoverable on any database that is not reading that. You would not be able to search for “physical+digital” releases, for instance.

Worded differently, summarizing off-chain, on-chain, and anecdotal yet provable data, in a single data structure is the greatest leap forward in information efficiency for digital art.

Continuing down the rabbit hole, the token ID published by Christie's brings us to the Arweave link holding the data below (see Figure 6). The MP4 is also stored on Arweave but realistically can have redundancies within an IPLD block. This data is the end-of-the-line for discoverability and attributes. It is possible to remint tokenURI fields over time, but these changes should be event's that add new blocks to the Lore IPLD chain.

However, these changes may not be properly logged on a time-series database and may fail to exist if a contract owner loses their keys or in 2090, the artist fails to survive. This data may also not survive bridging to other networks and is subject to various risks.

These are costly problems to overcome. Data that is fragmented, lost, incomplete, difficult to obtain, or untrustworthy will undermine the potential value of these pieces at time of sale. How would you know, by buying Mother Tree that there was a sculpture? It’s original price?

The solution to this issue is not necessarily imposing standards, which won’t always be adopted or adhered to. Rather, we can solve this by providing tools for normalization and contribution by other actors.

This is why we have such high conviction regarding the Atomic Form’s Lore Framework. It will allow us to create a singular blockchain based on all on-chain and off-chain data. We will have the flexibility to create public, private, and hybrid APIs based on the permission levels of the data pulled into these datastores. Atomic Form may also provide this service alongside artists, collectors, game developers and/or anyone else interested in hashing token data or non-token data for canonical reference. AF’s IPLD chains can use any number of storage solutions (e.g.IPFS, Filecoin, Arweave or others as blockstore solutions) and in the future will be audited by independent actors in the way ETH is now.

As we will soon announce, we are contributing to a network to prove blockchain state data.

Part 3: On Exhibition as a Lost Piece of Metadata and the P2P Exhibition Economy

Atomic Form has also found itself as a centralized Web2 source of truth regarding the exhibition for Christie's x MAPS sale (discussed in the previous section); Atomic Form used custom display hardware to support the show and powered it entirely through a hardware device called the Photon. Well, 27 Photons as seen below:

Each Photon is independently registered with a single exhibition device which includes but not limited to: a High DPI Projector, 1:1 custom display or a store-bought 65" OLED. The Photon is also tied to an AF user profile or wallet address. In this instance, Christie's essentially had a user account on AF that received exhibition rights.

The Photon is a small computers that are optimized for GPU, CPU and Memory. Photons do not cast digital art. Instead, AF’s software parses the tokenURI fields for the media links and downloads them locally. The owner of said NFT must authenticate the local downloading of the media. In this process, users sign messages authenticating display and exhibition to specific devices.

Users may also sign said authentication remotely, allowing users to lend exhibition rights without losing custody of the token, from my photon to yours. For example, a user in New York City may wish to loan an exhibition to a gallery in Europe or a friend in Hong Kong. Through the Atomic Form OS, the signature that authenticates ownership of an NFT and its exhibition on a specific photon or user can be stored in the same database that combines the Ethereum, Christie's, and Arweave data mentioned in the section above.

Bringing It All Together

Imagine going to an NFT auction in 2025 and you have no idea about NFT’s and digital art but you have recently sold your laundry share start-up and know a lot about memes. A QR code at the door lets you bring up an app with a map of all the digital collections on display along with a search function tied to Lore. You approach a high-resolution particle whirlpool that looks familiar. You snap a QR for the piece and are redirected to a provenance page that uses the tools mentioned in this article to add a myriad of securities and context to this particular sale.

By taking a quick look at the blockchain data associated with one piece you know it was minted and transferred in 2018 but the receiving wallet has used 100’s of ENS names and the history of registrants has not been captured or indexed easily. By reviewing annotations based off of Lore, you discover that this piece was somehow on exhibition both at the Venice Bienale and a Snoop Dogg concert in 2019. You feel happy because signatures that attest exhibition purposes for the concert have been logged and even if they have not been logged on Ethereum, sources of truth such as Snoop Dogg’s Twitter or Rolling Stone verify this provenance. You rest easier knowing the original 4k IPFS associated with the token now has an 8k render on four different filestores on a special blockchain.

You wonder about the history of cryptopunks.

You search historical records on Lore for punks that fit 3 criteria: have been licensed, are V2, and have received high value loans. You find punk #2499 fits these classifications. The punk has over 50 contract interactions with Defi protocols generating solid yield. You scan a QR code and Lore has been used to hash data around vital points of the punk.

You discover that this punk, call it #2499, was licensed to a chain of restaurants for a laughable annual sum, making its loan-to-value ratio more rational. You now confirm the IP of this punk was considered interesting. You also know for a fact that the CC0 rights for the punk are intact because Yuga has been Oraclized into Lore as a source of truth (like Mother Tree having a physical sculpture on the Christies page, or Snoop Dogg’s Twitter showing him performing in front of the same piece you are considering buying). You are also astutely aware of the fact this was the first punk to go through a bridge from ETH to DOT as well (even if you don’t know what DOT is). Finally, you have not only ETH pricing history for the punk, but Web2 auction house pricing, and a series of loans. Your information set is vast.

Finally, you have made a considerable sum from P2P exhibition economies based on AF hardware and software products. You, as an individual, are so integral to the future of cryptographic royalty generation that you seek pieces that have high rates of signature generation for exhibition purposes as a part of your collection. You loan your new works to an auction house using photons in Hong Kong to generate revenue from exhibition while you sleep. The data is recorded. You become the Lore.