M○C△ WEEKLY Round-UP 6/22

We got NEWS. We got THOUGHTs. We got ART! We got you covered.

Jun 22, 2025

Don’t miss an edition, hit this li’l ☝️☝️ subscribe button ☝️☝️ right up here. And catch up on every edition that passed you by.

In This Week’s Round-Up:

A Timely Reminder that Twitter is a Dangerous, Garbage Product
Is AI Going to Become Another Streaming Service?

Okay, Let’s Get On With It

The Crypto Art Side

Welcome to the new age, people, the age of freshly reconfigured Twitter content guidelines, sprung suddenly and without announcement, affecting individuals and platforms alike, reminding us all of the startling power a single social media site has over our entire art movement.

Where should we begin first? In that sense, what do you consider more important? That Shaw Walters, founder of AI-agent-spinup-service ElizaOS (formerly ai16z, though the project’s token still bears that title) was unexpectedly and outright banned from Twitter —along with the ElizaOS page too? Or that the much higher-profile Pump.Fun (launchpad for millions upon millions of Solana-based memecoins) was removed in the same flurry? While Pump.Fun’s page has since been reinstated, Shaw’s and ElizaOS’s have yet to be, displaying Twitter’s outright favoritism for users with substantial followerships. Nevertheless, we have the receipts to prove Pump.Fun’s ban, as well as many others in the greater crypto ecosystem, and while the pretense for the ban seems to be an attempt to pacify incoming crypto regulators who may be cracking down on memecoins and pump-and-dumps and other theoretically-predatory crypto-based activities, the warning is stark: Twitter’s overlords can —at any time and without pronouncement and for any reason and with no timeline for recourse— kick whomever they like off their platform. Perhaps we all know this by now. Perhaps it’s just another opportunity to look the truth in the face.

Whether via outright bans or the shadowbans (unannounced and only discoverable when an account fails to show-up in the search bar) that have affected artists like Natrix, OONA, PolyAnnie (all women whose work includes sexual content/commentary; OONA and Annie are currently shadowbanned), and even we at MOCA ourselves, the crypto ecosystem’s allegiance to Twitter as its central communication nexus has harmed its creative figures time and time again. It’s a shame that this kind of thing only seems to make waves through larger circles when it happens to larger commercial holdings like Pump.Fun or BloomTrading, but it’s a useful reminder to all of us that this art movement cannot sustain itself in any kind of decentralized or discoverable way without the willing participation of their host sites (be they Twitter, BlueSky, or Farcaster), and that their participation is always at risk of being yanked away.

I’ve written about Twitter’s influence on crypto art from an aesthetic standpoint before, but there’s been little —ironically— beyond Tweets themselves examining the feedback loops our presence on social media creates. It’s worth reanalyzing the risks to crypto art should Twitter decide to remove or silence artists en masse, especially now that there’s solidified offline a somewhat plutarchical class of artists and collectors around whom the movement’s narrative will only petrify further without the muchness that social media provides. We know these names, they don’t need rehashing any more. They the artists in represented gallery rosters, the collectors with hundreds of thousands of followers and brand sponsorships, they comingly at Christie’s and Sotheby’s sales, they’re all geniuses and innovators and believers of the highest order, according to many. To this point, few artists have been lucky enough to break out of crypto art’s staggering multitudes to see any kind of reliable success of the IRL variety: the shows, the invites, the talks, etc. Not to say that should be anyone’s goal, and it’s out of the range of possibility for many who live outside of major artistic urban centers, but it’s —especially now— a perhaps necessary bulwark against the social media rot, that which afflicts the too-unfollowed by way of bans and algorithmic disenchantment.

Shaky, shaky ground we’ve built this movement on. As ROBNESS explained to me once, Twitter was always going to be the hub of crypto art because it was always the hub of crypto, back in the days when founder Jack Dorsey’s crypto-tolerance made the platform a uniquely safe space for the discussion and proliferation of digital currencies. Twitter’s negatives are many, but its chief positive has always been in its ability to put art we didn’t know we needed in front of our eyes when we least expected it. The more that single positive is threatened, the less useful Twitter is as anything but an influence signaling device. Yet, without anywhere else that can host a grassroots movement, I ache for the masses who will be disenfranchised by enforced silence before they even had an opportunity to say their first words.

There’s no small amount of fatalism in my words today, I recognize that. After all, it seems insane that Twitter would invent any pretense for banning artists in great swaths, right? And yet, the women above were all banned for sexualized content or commentary, nothing harmful, just corporately distasteful. All the Nazi signaling and murder videos on Twitter are fine, it’s the tits that need to be covered-up. Artists from Iran, Cuba, and elsewhere have been banned for no other reason than their country of origin. Politically inflammatory art, risque art, anything that does not fit the oft-arbitrary guidelines of the host platform is in danger of being muted, the art and artist both. I’m certainly not the first to write about this. Others have done it more eloquently and with more emotional resonance. Nor am I coming to you with an holistic solution; I don’t see anything presently available. I wish I were more intelligent, more creative, could envision ways out of tight spots. But perhaps I’m doomed only to see problems without solutions.

This may be an answer as to why so much crypto art today feels anesthetized, riskless, corporate in the way it refuses real confrontation, is placating and anti-inflammatory. Perhaps it’s more than Twitter not showing us this kind of art. Perhaps its creators have already long been locked in the attic, locked out of our digital kingdom outside our attention. Would it surprise you? If Shaw and ElizaOS and Pump.Fun didn’t have many millions of dollars locked into their ecosystems, would we know about their bans at all?

At least Pump.Fun is back online. All’s well that ends well, huh.

The Tech Side

I’ve been meaning to write something on this topic for weeks now, after reading the (now old) news that The New York Times signed a licensing agreement with Amazon of all companies “to use its editorial content for artificial intelligence products such as Alexa, marking the publisher's first licensing deal tied to generative AI,” according to Jaspreet Singh at Reuters. Meanwhile, back in April, Jeff Bezos’ Washington Post signed a similar deal with OpenAI, wherein “ChatGPT will display summaries, quotes and links to original reporting from the Washington Post in response to relevant search queries. ChatGPT will feature the Post’s content across politics, global affairs, business, technology and more,” as per Todd Spangler at Variety. Both agreements are shifts away from the henceforth litigious attitude taken by publishing companies (of music, art, writing, etc.) regarding the generative AI giants who many claim unlawfully trained models using their copyrighted work. Not to say that the litigiousness is going to end (given the blatant copyright violations across almost every major AI training-house), but now that a few powerful publishing organizations have begun choosing their AI fighter of choice, a fascinating future may be at hand, one in which the legally-included content within a model becomes a prime selling point, a differentiator for models based quite literally on what’s inside them.

The number of these lawsuits, by the way, is staggering. There’s Disney and Universal’s claim against Midjourney. There’s Getty Images vs. Stability AI. There are 12 individual claims by authors including George Saunders, Junot Diaz, Michael Chabon, and John Grisham, and also the aforementioned New York Times, filed against OpenAI, so many across multiple U.S. states that they were federally consolidated into a class-action. However you yourself feel about the rights of AI companies to use this data as it freely it exists on the internet, there will be rulings on the subject. Likely, there will be rulings in different directions in different governing regions, and given that American judges at least have already begun siding with plaintiffs in similar (albeit lesser) cases, there’s a fair chance that the age of unbridled knows-everything-because-it-was-trained-on-everything models is soon to end, given the sheer amount of expected recompense in forthcoming payments to unlicensed sources.

Point being, the legality of these models, as they currently exist, is under great threat. Thus, reliance on licensing agreements —with publishers, music labels, estates, individuals— is not only going to increase, it’s going to become extremely competitive. Noncompete clauses, first-look advantages, preferential licensure, the same mechanisms that govern the world of streaming television are naturally coming to generative AI models, because that’s how this industry works. It will begin as a way to avoid lawsuits —for example, buying a license The New York Times for the rights to continue using their information in the models— but will quickly become a differentiating factor. Amazon signing a licensing deal with the Times means nobody else can sign a licensing deal with the Times; only Amazon is exempt from the ensuing litigious fire; it’s both a defense mechanism and a perk. Use Amazon AI, get The New York Times! Use OpenAI, get the Washington Post! Netflix has Seinfeld; Hulu has It’s Always Sunny in Philadelphia; Anthropic will have the Wall Street Journal (not actually, just as a point of comparison). Imagine the selling point of every model being the information inside of it, that it is permitted to pull from and summarize and put in front of your face. Right-leaning, left-leaning, parochial, globalist. Which models can create what kinds of music based on what internal processes, which models can reproduce certain styles of art. I believe we’re moving not towards a singularity but towards a momentous and stultifying split, a way of packaging Generative AI that is akin to a cable package, a bundle, three channels for one low monthly price.

Perhaps this is the nature of all entertainment-adjacent products, and make no mistake, for all its uses, out-of-the-box generative AI is an entertainment product, competing for our attention alongside sports, streaming services, gaming consoles. AI giants want your eyeballs, and they want your attention to turn into paid subscriptions. They want massive companies to license specialized instances of their products, and they know that these companies expect their employees to actually use these instances they’re paying so much for, and so once again, attention metrics are paramount.

I’m waiting to see exactly how these licensing agreements are evoked within an AI’s interface. Will Amazon AI’s future UI/UX include news summaries drawn from the New York Times, and will that differ in content/style from OpenAI’s WaPo offerings? Let’s take this to its logical conclusion, where all created content across mediums is specially licensed to one or another AI model, and it’s easy to envision a world where we use each of these massive models for different things, and different demographics split on where their loyalties lie. Here’s your model used for the news, and your political leanings, your country of origin, or ethnicity will dictate which is most applicable to your life. Here’s another smattering for research with the same individualized offerings, others for art-making, some for music, as many AI models as there are apps, because each of these giants will need to choose content masters to compensate and thus placate. The same way Netflix’s offerings in the U.S. differ from those in Europe, we might have ChatGPT textboxes providing fundamentally different answers depending on where users are accessing the model and what courts have ruled therein.

It would be an insipid, overly-economic, frustrating kind of future indeed, even if it’s ethically correct. But that’s probably how this ends up. Either complete AGI takeover…or a generative AI bundle for a new low low price, in ad-supported tiers or not.