Why crypto networks are the foundation for the next generation of content recommendation systems…

Back in 2017, I ran a server out of my college bedroom that had a thousand completely automated Instagram accounts reposting clips across sports, gaming, and comedy. Some days we’d do over 15 million impressions. I barely lifted a finger. More recently, i’ve been thinking about what the implications would be across creation, consumption, and curation if I were able to connect this system with some of the new A.I. enabled creation tools we’re seeing. In a world with programmatically generated media, we’ll need new tools for curation. Crypto networks might just be that…

One of my mixes for some background tunes if that’s your cup of tea.

So why even think about new systems for content recommendations?

When looking at the history of online media we’ve seen a consistent evolution in content creation tools that simultaneously required improved means of curation. Curation methods like social graphs were conceived of during the emergence of user generated content and mobile-first creation. Fast forward to today, we’re seeing the materialization and convergence of a variety of social, monetary, and technological trends that power new levels of content creation.

We have formats for virtually every media type, genre, and duration. Creators producing more content than ever to fill content gaps across different niche. Each piece of content recursively cascades into one another from long to short-form. And now, synthetically generated media emerging through products like Stable Diffusion; enabling anyone to generate hundreds of creations in minutes through a simple keystroke. We’re seeing the infrastructure and emergent toolsets lay the groundwork for content production at the rate and depth we’ve never seen before.

However, just like traditional social graphs quickly became outpaced as means of curation — today’s recommendation systems don’t properly account for a world of infinite content creation. Our attention is a physically constrained resource and as a result our ability to furnish value across an exponentially increasing amount of content begins to fade quite quickly. In the end, we’ll be again left without a proper means of curation for the next generation of media online.

All forms of speculation are the underlying fundamental mechanism for curation

Today almost every action we take on social platforms is a form of interest speculation. We bookmark photos in a speculation of our future selves liking it. We like our crush’s story in a speculation of her noticing us. We share videos with friends in a speculation on whether they’ll enjoy it. Each interaction is a form of speculation that powers connecting our non-explicit interests, gathering collective attention, and rewarding content in recommendation engines.

Meanwhile, many people claim that crypto needs to move past its only use case of explicit speculation. However, our transactions offer a similar look at our interests as our likes, comments, and shares do. Transactions like minting, selling, sending, and burning form a history of affinity to media types, social relationships, and usage patterns. In aggregate, we’re able to look at someone’s wallet and see what they’re most interested in.

While web2 interactions and web3 transactions may feel different, each are the basis for value determination and accrual on their respective networks. We’re simply early in the evolution of how consumers can interact with media in crypto products, the amount of capital required to use them at meaningful frequencies, and how the network of products interact/play off one another. Eventually, we’ll reach a point in which our actions generate enough on-chain data to synthesize meaningful insights into our interests.

Crypto offers a new mechanism for signaling value

Ethereum is a financial network first, media network second. By putting media on-chain, your media inherits the characteristics of the network. Meaning, media becomes an owned economic asset where value is signaled through transaction data. More importantly, transactions introduce an additional layer of scarcity for determining value in financial capital that have theoretical unbounded levels in value signaling (some people have a lot of money).

For example, look at SHL0MS NFT auction dubbed $CAR as an example for why financial capital creates important on-chain data in value signaling. Traditionally, when we mint an NFT the value of it is often a standard mint price + gas fees. However, in this auction the top 888 bids received an NFT where at anytime you could increase your bid. In the end, 1 through 888 NFTs were distributed to bidders highest to lowest. In practice, what we saw was a continuous rise in floor price where people paid for each NFT equal to their affinity + belief in its relevance to the world.

By introducing markets to media — our transactions types, price, frequency, and wallet history formed the on-chain signal of how valuable that media is within the broader network and to who. The additional introduction of capital and pricing allowed a degree of value signaling much broader and significant than a commoditized like, comment, or share. In other words, during that auction unique individuals expressed interest using a finite resource on a singular digital plane.

As a result, we can quickly start to see how crypto markets can serve an important function for curating media in the future. Just how we comment “first” on content we found early and love in web2, we’ll leverage our capital to signal the importance of media in the future. While it’s being foreshadowed now, we’ll need to see new inputs into on-chain content stores, products that create relational graphs between content, and new on-chain interaction methods. Luckily, many web3 organizations are well capitalized and understand the unique characteristics of on-chain media to unlock this future.

Closing up shop…

In summary, not only do I believe that web2 interactions are similar in purpose to web3 transactions, but crypto networks can be the infrastructure for the next generation of content recommendation systems. With a new network comes new signals for value. The introduction of capital creates the opportunity for scarcity and wider ranges of value significance to be applied in curation. So how should we think about structuring it? I’ll share more about this in my next post.

If you enjoyed reading this, feel free to reach out to me @sorenwrenn on Twitter if you wanna jam. I am one of the co-founders of TITLES, a new way to create, remix, and share crypto media. Thank you to @kneelingbus, @holyn, @marcusestes, @parkerseagren for early thoughts on the post. Please subscribe if you haven’t, I’ll be publishing a number of thoughts here over the remainder of the year!