The future of media is a bank

A world without subscriptions.

*Hi, my name is Daisy Alioto. I have worked with some of the most prestigious media companies in the world and I am currently the co-founder and CEO of Dirt Media.

In this paper I will argue that spontaneous payment systems are the future of subscriptions. Rather than subscribe to media at a set rate, users will curate their information diet through a fleet of autonomous AI agents trained on their taste and empowered to tip for it.*

The new desirable consumer will have high agentic capital, meaning that they are the most skilled at using AI to optimize their monthly entertainment spend. Attention will no longer be measured by inflated engagement metrics on social media, but in stablecoin transactions signed by both agents and the humans that deploy them. Programmable wallets will be the norm.

Welcome to the future, where paying attention means to pay.

But how did we get here? First, I will explain why the media industry is so broken from the perspective of someone who loves both magazines and technology. Then I will explain how a combination of AI and digital wallets can unlock new audiences and spend. Finally, I will argue why this is better for both the creator and the consumer. ⟢

CONDÉ NAST HAUNTOLOGY

It has been a summer for media nostalgia. Back in April, former Vanity Fair editor and Air Mail co-founder Graydon Carter released his memoir, When the Going Was Good. The press cycle around the book was an opportunity to reminisce about $166,000 per story rates and expensed town cars. The Condé Nast nostalgia flames were fanned again this month with the release of Empire of the Elite by Michael M. Grynbaum.

Do we miss the glamor? Oh yes. But Christian Lorentzen's review of Grynbaum's book gets at the real crux of this nostalgia: "The best editors I've known have always been talent spotters, the sort who prefer saying 'yes' to saying 'no,' even if their jobs require them to say the latter more often."

Call it The Meaning Economy, friction, The Taste Economy, gatekeeping, curation as a service, Slow Media, or the value of attention. All of these frameworks are actually describing the same thing—the exchange value of media. What the magazine really represents is a system of value. A marketplace for real status, and by extension dollars, if you're lucky.

This is the reason every tech cycle since the release of the personal computer seems to reinvent the concept of the magazine. You can stand onstage and gesture at a magnesium block or wave your arms up and down and cry "we are the media now" but at the end of the day you still need a slot. A slot to put your attention in, a slot to put your money in, a slot to put your dreams in, and out comes status.

That slot, sorry to say, is not the computer. Yes, "there is money in the computer," as the old programming joke goes. But there's only status in the magazine. Why is there money in the computer? Because there are magazines on there. All sorts!

Some of them live on Substack, which is officially a unicorn. Others are on YouTube, like Joshua Citarella's Doom Scroll. Though I've argued elsewhere that solo creators are more like vessels for taste than containers, I don't think that a "magazine" strictly has to be text. If a media brand can come to signify a distinct point of view, it has the makings to be a status distributor. Regardless of your opinion on traditional media, three things have become abundantly clear:

Media is no longer the core product of any media company
Subscriptions as we know them will soon be obsolete
The rise of AI agents will create net new audiences for media

The magazine, preferred buzzwords aside, exists because of the fundamental nature of the digital media ecosystem. The first creator is always a human, the final curator is always a human. In between is a vast bazaar of supplies and demands. This bazaar is currently incredibly inefficient.

It is inefficient because the internet hasn't solved paying for media. It solved media distribution, immediately and irrefutably, but the times when the internet seemed to be a sustainable source of income for media brands were an illusion propped up by cheap venture dollars. (If you want to know more about that, check out Eddie Huang's documentary Vice is Broke.)

Of course people pay for media, in the same way that Americans vote. But in order to accomplish anything worth doing, you have to motivate the ones that don't. Getting people to pay for media is, for me, the only problem worth solving. I think about it when I wake up and when I fall asleep. I am so interested in solving this problem that I spend more time on it than I spend on the thing I love most in the world, writing, because everything I care about is downstream of it.

And what a stream it is. For every chic memoir reminiscing about the Time & Life building bar cart I'll show you a millennial sick with nostalgia for $60,000 blogger salaries and free seltzer. The digital media ecosystem as we know it is getting much, much worse. The works are gummed up with AI (and, sadly, human) generated slop, filler and Flubber. You're trying to jam your coin in the slot to get some damn status and it's just not going in.

We could throw up our hands and sink into the traffic apocalypse, allowing the delta between small, print projects and viral short form video ecosystems to grow, a population split between insufferable book clubs and slobbering iPad babies.

Larger digital media companies could limp along, throwing events for influencers and licensing their content to OpenAI. You could be one of the few hundred people to get rich on Substack. We could all become livestreamers for an audience of bots, political podcasters without real free speech, or sell feet pictures to fund graduate degrees that won't get us teaching jobs. We could turn every blog post into a meme coin and then start a blog to explain the meme coins.

We could use AI to write our own songs, about ourselves, cast virtual avatars in the music videos, and watch our own hyperpersonalized ads starring synthetic influencers—Klarna-ing some athleisure we don't need and won't wear outside the home.

But then, what would we talk about?

MEDIA IS NOT THE PRODUCT

In 2025, every company is a media company. Unless you're a media company, then your core product is elsewhere—community, events, data. As much as media alone is not a viable product, it's an essential wrapper around everything from protein bars, to venture firms, to fitness equipment.

AI underscores the importance of taste in brand differentiation. This could take the form of curation, or friction between the customer and the experience they are hoping to have. Clothing brands have newsletters, decentralized states throw parties, Ethereum layer-2 networks show up in Erewhon.

Media is still the best taste signal we have, which is why MUBI just raised $100M from Sequoia and A24 is valued at $3.5B. If the last ten years have been about exploiting—and later, escaping—the algorithm, smart brands are already trying to beat the LLM, rapidly disrupting our understanding of online real estate with very real, energy-guzzling, data centers.

There are different units of real estate on the consumer-facing internet, some more notable than others. In 1996 we got page rank, the hierarchy of reputation and authority that orders results from Google.

With Satoshi Nakamoto's famous 2008 white paper, blockspace emerged as a new way to organize the internet, a form of storage that could not only store reputation but also receipts. But it's expensive and annoying. And statusless. While people want borderless money in principle, what they really want is to be a part of something.

LLMs, the various chat interfaces that have replaced the search query process (and I'm including those slippery little Gemini summaries in this), mark the end of page rank supremacy. And most consumers will not interact with blockspace through bitcoin, Satoshi's original use case, rather stablecoins—making receipts far less interesting when $1 is $1.

"The hunger for something shared is going to increase," writes Stratechery's Ben Thompson about the solipsism of the AI era. Any new system for monetizing media needs to combine the provability of blockspace with the social dynamics of taste to create a new hierarchy for content that both humans and AI agents can participate in.

Taste + Blockspace = The new page rank

Paying attention will increasingly include paying something—even a very micro amount of money—as an attestation that something was viewed. We'll see novel combinations of taste, limited digital space, and reputation as the read receipts of the post-AI era, when friction doesn't just signify a barrier to consumption but, rather, the presence of friction indicates that something was consumed.

Currency isn't universal. Engagement, as currently measured by social networks, isn't stable. But we can stabilize it by getting people to pay for media.

SUBSCRIPTIONS ARE NOT THE ANSWER

Search is not the only part of the digital media ecosystem that is obsolete. We are also in the denouement of the ad-supported internet. Referral revenue from SEO optimization is headed out. Ad impressions for dollars? Also out. Subscriptions, which formed a substantial part of this economic pie, are also changing.

And they should change. Because they aren't efficient. Why does a subscription to a magazine with multiple fact checkers cost the same as my friend's newsletter? Why is it normalized to pay $0 or else choose from a flexible, but ultimately rigid, set of subscription tiers? As a newsletter proprietor, it makes no difference to me whether someone pays $5 a month for two months before cancelling or gives me $10 once a year because they really liked reading about underhand serves, midwestern perfume, or skateboarding over bricks.

What I am calling spontaneous payment systems (SPS for short) are normalized on OnlyFans, TikTok and at your local coffee shop (the tablet is just going to ask you a quick question). Not so much in the world of digital media, where you pay to enter rather than when you're in a good mood on your way out.

There have been multiple attempts at micropayments over the years. Initially, they were impractical because banks don't like a lot of small transactions and publishers don't like a lot of small transaction fees. However, there have been workarounds (including the blockchain) to this for years.

The real reason micropayments have never taken off is that there is no status in paying 5 cents to read an article in The Guardian, and no parasocial impetus. Most people don't subscribe to media because they can't get the information anywhere else, rather, they want to support the creator(s) and have access to their POV.

As the internet fills up with slop, the real opportunity is for micropayments to replace subscriptions as a proxy for attention, and for these transactions to be as public and socially oriented as Venmo. Money can be waiting for creators, and claimable at their email address, before they've even onboarded to the payment platform.

I'll stop short of saying the largest publishers could be stablecoin issuers, but why not? Starbucks is essentially a bank, storing and earning money on the float of the world's coffee spend. We need infrastructure to efficiently store and distribute the world's entertainment spend as well, a marketplace where you can both give attention in the form of micropayments and earn the attention of others as well.

THE RISE OF AGENTIC CAPITAL

Well, that sounds terrible. I hate making decisions. In my purest fantasy a faceless man orders off a restaurant menu for me and bankrolls my writing practice for the rest of my life. Not going to happen!

Besides, I do like making decisions based on my taste. It's just overwhelming sometimes. If I could delegate some of those decisions to agents based on my taste my life would be much easier.

Oh, like an algorithm? No, not like an algorithm at all. I can tell you're confused so let's take a step back.

During the scale era of the internet, traffic reigned supreme. Ad-supported websites like Buzzfeed, Mic and Refinery29 chased impressions in order to stay in business. Going viral meant more impressions. In this scenario, any reader had the same value to the company—one reader, one impression per page. With the caveat that the closer someone seemed to the median reader the more valuable they were as a signpost. That is what going viral meant, catching fire with the masses. These websites had normie capital.

Then came the era of niche, B2C-bordering-B2B subscriptions. Newsletter-forward brands with boutique audiences that made a lot of money and were willing to spend it for fresh insights about their industry. I'm talking about Axios, The Information, Puck etc. It was still one person, one impression, but these people had more money. They had capital capital.

In the first scenario, companies grew and monetized by reverting to the median reader over and over again. In the second, they created new audiences by tapping deeper into the spending power of industries like entertainment (Puck, The Ankler), technology (The Information, Semafor) and political lobbying (Axios, Punchbowl).

In a variation on this trend, media startups like the ones in The Chernin Group's portfolio (Food52, Hodinkee) aimed to get people with average capital to spend more on their publication through ecommerce—content, community and commerce combining for mixed results. TCG's latest investment? Substack.

And that's where the history of traditional digital media stops, basically. Excluding creator economy breakouts like Joe Budden and Call Her Daddy, there have been no new business models for almost a decade. Part of the problem is that you need money to test new things and it's hard to make a case for investing in media unless you can bring in net new audiences. Normie capital has pivoted to short form video and capital capital can only accommodate so many paid fireside chats with Bob Iger.

Cue the rise of agentic capital. A new class of audience using autonomous AI agents trained on their taste to extend their own attention and spending. Suddenly, one person can be worth more than one impression. The average entertainment budget can be more efficiently split across creators, leading to all sorts of opportunities. How do we make it easy for them?

Jason has $200 a month to spend on all online entertainment. He deposits it in an app where it is converted into a stablecoin and delegated to a wallet controlled by an AI agent trained on Jason's taste.

The agent begins to spread it out across Netflix, Spotify, individual podcasters, newsletter operators and magazines. It makes no difference whether the article is on Substack or an independent blog—Jason's autonomous AI agent finds it, taking into account what Jason's friends and connections have recently paid for. Some of these transactions the agent completes itself to get around paywalls, others it brings back to Jason for approval—like suggested tips for content that creators have uploaded for free.

Because Jason is part of a network of entertainment spend, his Netflix subscription is liquid. Some months he pays $0, other months he pays $20, depending on his anticipated usage. Jason's AI agent is now part of the total addressable market for any new media startup.

But wait, Jason isn't just some guy. He has his own following through a podcast he started with his friend. A lot of listeners like and trust Jason's taste. It's easier for their agents to subscribe to his agent, to get the same recommendations that Jason gets. A secondary marketplace for agent subscriptions arises on top of the primary one—the one for Jason's attention. And yours. And mine.

Maybe Jason can eventually earn yield on the amount of money he keeps in the app. Or one of the creators he follows will launch a token. That doesn't really matter in the beginning. What matters is that Jason is increasing his status by paying for media, something that has never happened before at scale.

The people with the most agentic capital will be the people with the most taste. Most of the focus (and panic) around media and AI has been on AI as a media creator, when it can create far more value as net new audience spending in the market.

Remember, a human is always the first creator and final curator, magazine without end.

THE SLOW CANCELLATION OF THE PRESENT

Any protocol that can harness AI agents to trade the taste of real people, with their input and the input of their social circle, has the potential to become the world's entertainment banking system. Think of it as setting a bounty on your attention for as little as $5.

This will create new "emergent status games" as my friend Trevor McFedries likes to call them. Data broker, content liberator, taste mapper, superfan—all of these actions, when viewed through the non-custodial algorithms of the past ten years, will seem like fumbling blindly in the dark.

At last, your attention is worth something to the media you love most. There is money in the status machine after all—let's get it out. ⟢

Dirt has been building a technology product that can meet the media moment. Interested in following along? Find us on X.

RELATED READING

The Most Valuable Commodity in the World is Friction by Kyla Scanlon · In the AI Age, Making Things Difficult Is Deliberate by Willem Van Lancker · Attention Is the Last Scarce Resource by Rex Woodbury · Attention is what makes us human by Daisy Alioto · "The Taste Economy" by Daisy Alioto at FWB FEST24 · The Taste Economy by Daisy Alioto (2023) · Computer Scientists Figure Out How To Prove Lies by Erica Klarreich · Cloudflare introduces "pay per crawl" · Micropayments that support revenue sharing by Liz Harkavy · Flynet operating statement by Blackbird · Stripe introduces Stablecoin Financial Accounts · WaPo eyes micropayments by Sara Fischer · Every marketing channel sucks right now by Andrew Chen · Is it finally time for media companies to adopt a common publishing platform? By Jesse Knight (2018) · How and where will agents ship software? by Stepan Parunashvili and Nikita Prokopov · The Agentic Web and Original Sin by Ben Thompson · Content and Community by Ben Thompson · We're all in it together ft. Defector's Jasper Wang, 404 Media's Sam Cole, and Hell Gate's Max Rivlin-Nadler on Tasteland · Taste Is Eating Silicon Valley by Anu Atluru · Persuasion Markets: A Manifesto by David Phelps · Economics of Love by David Phelps · Against Platforms ft. Mike Pepi on Tasteland · The Print Magazine Review of 2024 by Amanda Mull · What if Subscription Isn't the Right Business Model for the Internet? by Monica Villar ⟢