Discover more from Bryce Baker's Newsletter
When the cost of content creation hits zero
Blockchain will be integral for AI content creation tools to flourish
“Check out this picture of Tony Hawk doing drugs.”
“I thought Tony was cleaner than Clorox. Where’d you get that picture?”
“It’s not real, I made it using AI, Artificial intelligence, bro.”
“That seems dangerous.”
“Very dangerous things, theories.”
What is Dall-E 2?
Dall-E 2 is a new AI system from OpenAI that can take simple text descriptions like “A koala dunking a basketball” and turn them into photorealistic images that have never existed before. DALL-E 2 can also realistically edit and re-touch photos.
I highly recommend watching this short video, describing DALL-E 2:
For those who don’t watch it, the video explains:
DALL-E was created by training a neural network on images and their text descriptions. Through deep learning it not only understands individual objects like koala bears and motorcycles, but learns from relationships between objects, and when you ask DALL-E for an image of a “koala bear riding a motorcycle”, it knows how to create that or anything else with a relationship to another object or action.
Why is DALL-E useful?
Before DALL-E, if you wanted to generate an image of moderate complexity from scratch, you had to either hire a graphic designer or be one yourself. There are some exceptions to this, such as adding text to a pre-existing image, also known as meme-generation. However, as you’ll soon see, this will be much more powerful than meme generation.
The advent and acceleration of AI tools such as DALL-E will lead to advanced content generation, for free.
The wise child asks (felt jewish might delete later): How will this technology be used in practice?
The Nike swoosh is arguably the most iconic logo in history. Fun fact: “logo” was my first word as a baby. When I used to work at Nike HQ, one of the legendary stories was that the birth of the Nike swoosh was a crowdsourced logo.
True story: there was a $50 prize issued by Phil Knight to the winning designer who created the Nike logo, which ultimately was the swoosh.
It goes without saying, but Nike had great success with their search. However, a tool like DALL-E would significantly widen the pool of applicants able to participate, increasing chances of success of finding the right logo by a huge factor (probably by 1000x if you assume there are a few million graphic designers worldwide).
The impact of DALL-E on smaller businesses is significantly larger. With costs being the primary concern of small businesses, zero content creation cost is a lifeline.
For example, let’s say you want to run an Instagram ad to promote your brand. Instead of trying to hire an expensive agency or Tweeting out a crowdsourced contest, you can use AI image generation.
Take our brand, En Passant Digital, a brand which pays homage to chess (“En Passant”) and technology (“Digital”).
Next, I just have to figure out what to feed the algorithm to return the desired result. Even if my idea is inchoate, I can piece together a sentence based on what I want to portray about our brand, and adjust the inputs as necessary.
Let’s say I feed the algorithm:
An oil painting by Matisse (to show we are creatives)
A humanoid (to show we are humans)
Robot (to show we know technology)
Playing chess (to show we are strategic)
Input: An oil painting by Matisse of a humanoid robot playing chess. BOOM, I made an ad for our company in 30 seconds:
AI, the TikTok killer
The rise of TikTok demonstrated the power of an omniscient algorithm as it chipped away at the powers that be (i.e., Instagram, SnapChat). Some creators who make a large portion their living on Instagram, such as Kylie Jenner, have complained about Instagram’s attempt to copy TikTok by pushing Instagram reels:
The Kardashian fear is not misplaced. It’s a completely rational reaction to a landscape that’s shifting faster than Kim K’s relationship status (RIP Pete).
Former Facebook exec, Sam Lessin, explains the evolution of social networks in 5 phases, from the People Magazine era in the 1970s to now (I’ve added the platform responsible for the disruption):
The Pre-Internet ‘People Magazine’ Era
Content from ‘your friends’ kills People Magazine → Facebook
Kardashians/Professional ‘friends’ kill real friends → Instagram
Algorithmic everyone kills Kardashians → TikTok
Next is pure-AI content which beats ‘algorithmic everyone’ → Tools like DALL-E
We aren’t at #5… yet. AI content creation is still in its rudimentary phase. One of the biggest next steps will be AI generated video.
AI generated video
AI video is much more complex than AI imagery, as it’s 3D. However, as Ben Thompson explains, the progression from an AI text generator that helps you fill in your emails (GPT-3) to image generation (DALL-E) is a wide gap that was closed in just ~2 years. The text generator feeds on the inputs of the image generator, and the video generator will feed on both prior creations.
As our mediums of preferred engagement have progressed, our AI creation tools will progress alongside our preferences of consumption:
Eventually, we will be watching videos completely curated, and created, for our viewing pleasure, at zero marginal cost of creation.
The future I’ve described is coming faster than you think. FN Meka, a virtual TikTok character with 10M+ followers, just signed with Capitol Records. Pac Sun signed virtual influencer Lil Miquela to be the face of their new campaign.
The biggest moat that human social media influencers can hope to maintain is the strength of their personal connections to people, helping brands sell products that align with their persona.
It’s a Hanukkah miracle: I haven’t uttered the words “blockchain” or “crypto” yet. Not to worry. With zero marginal cost of creation, blockchain may actually have a use case ;)
Blockchain will be integral for AI content creation tools to flourish
As I spoke about in my last newsletter, the “real use cases” for crypto have often been misapplied. It feels like many are trying to fit a square peg in a round hole. In other words, “crypto-fying” things for the sake of crypto-fying them, as opposed to for the sake of greater efficiency:
That being said, as the cost of creating digital assets approaches zero, the importance of having a single source of truth (SSOT) and well-indexed search engines approaches infinity.
Single source of truth
Single source of truth, one of blockchain’s favorite calling cards, stands for the fact that when information is all over the place, we can point to one place for the most updated piece of that information. SSOT will be critical for:
Incentivizing great creation
When anyone on Earth can create realistic images and videos with the stroke of a keyboard, brands run the risk of losing control of their identity and being targeted by smear campaigns.
For example, in a subreddit, someone posts a hyper realistic AI-generated image of Hitler and Epstein, sitting by the fire, drinking Coca Cola, laughing and telling stories, with a big “ENJOY COCA COLA” logo plastered across the image (Godwin’s Law is undefeated).
If image generators automatically minted (AKA put on the blockchain) every image that was generated, the source creator would always be quickly identifiable. Coca Cola could register images created by their team from a Coca Cola wallet. Coca Cola will likely want stealth wallets to try out new ideas to A/B test new material, but anything they decide to release on public social accounts could be minted from their public wallet.
To drive this point home, AI generation could cause a major problem for brands due to the hyper-realistic nature of AI creator tools AND the fact that the amount of people that can be effective in advanced content creation is now anyone who can write (as opposed to just graphic designers). Most graphic designers have better things to do with their time, while trolls don’t.
Licensing digital assets
If high value digital art stays relevant, brands will license digital art in their marketing campaigns. With tools like DALL-E, you can see how brands might choose to leverage this.
Example: La Croix wants to put their can on top of a Fidenza by Tyler Hobbs, one of the most highly valued NFT art projects. La Croix finds a piece of his art they like, and upload their digital cans into DALL-E.
BOOM, they have a digital ad ready for Instagram built on top of NFT art culture. You can see the problem here: what’s in it for Hobbs? Hobbs needs to be sure a viewer can see who’s work they are looking at.
If you’ve ever used Adobe Photoshop, the concept you should have in your head is one of “layers.”
Layer 1 of the image: Fidenza
Layer 2: La Croix Cans
Layer 3: La Croix Logo… and so on
Each of these layers could be identified by a viewer. While La Croix will have to pay Hobbs for using his art to sell bubbly, derivative work will ultimately be created, and SSOT ensures that the O.G. La Croix ad shows who contributed IP to the image (a la song credits).
Incentivizing great creation
If there is no protection for the IP created using all of these tools and digital assets, the incentive to create really great things goes down significantly. It will be critical to provide incentives to this new generation of creators that can greatly accelerate the ability for people to connect with one another and their favorite brands.
Blockchain as an indexing tool
As the cost of content creation approaches zero, it will be impossible to find any of the content you’re looking for. It’s true that AI tools such as TikTok’s algorithm will help sort content based on preferences. However, when you’re looking for specific items, you will start to use tools such as DALL-E as your primary search action.
Instead of going from Google —> clicking a random blue link —> clicking another random link five more times —> Twitter —> finding the content (which wasn’t even the content you were looking for), you will just go straight to DALL-E, do not pass go do not collect $200.
Writing this article espoused a perfect example of this. I have seen so many web3-use-case-spoof videos over the past few months. Yet, finding some of the better ones for this newsletter took 15+ minutes, and I couldn’t even find the video I was looking for. As DALL-E tools get better, I’ll just type in the exact thing I want to see. If it’s on a blockchain index (e.g., Etherscan) and has already been created, I’ll have my result returned instantly. Or, if it hasn’t been created, I’ll have my content created, instantly.
Balaji, former CTO @ Coinbase, explains it well:
Anyone who works in crypto knows this is fact. On a block explorer, I can click on any action, see it’s entire history (when it was created, by who, where it’s been sent, etc). Google trying to keep up with infinite content creation and sources will be a Sisyphean task.
As the old adage goes, Google it. As the new adage goes, bitcoin solves this.
What should you do about it?
Considering harnessing the power of AI content creation tools?
Don’t get caught in the muck - let us help you wield the blockchain.
Or don’t, you missed everything anyway.
Thanks for reading En Passant Digital's Monthly NFT Newsletter ! Subscribe for free to receive new posts and support my work.
About the author: Bryce Baker is the co-founder of En Passant Digital, which designs and leads Web3 strategy for businesses.
Our other co-founder, John Tabatabai, has led investments for Crypto VCs and consulted for projects across the NFT space. He orchestrated one of the largest grossing NFT projects, kickstarting the first NFT bull run in early ’21, creating over $200M of value. John is designing one of the most influential Metaverse Community projects ($300M valuation at time of writing) which launches shortly.