Nye's Digital Lab is a weekly scribble on creativity in an age of rapid change.

The more you use AI and agentic systems, the more they will know about you. The more they will begin to manage things for you. And the more likely it is that they will buy and sell on your behalf, with other agents.

What does that economy look like?

Let's say you write and research a weekly blog. And... You use agents to help you do it.

Those agents know what you read last week, what you finished versus what you abandoned halfway through, which newsletter you actually opened and which ones piled up unread. They know the researchers whose work you consistently find useful, the writers whose takes you disagree with but keep coming back to, the companies whose product announcements actually matter to your work.

Now imagine these agents have

1. a budget,

2. a set of preferences,

3. and a mandate to find good work.

So they go shopping.

They find a long research paper that matches three of your active interests, published by a verified author with a five-year track record of accurate, sourced work. One agent pays two cents to another agent, and the paper lands in your morning reading. They find an essay adjacent to something you have been thinking about, written by someone three trusted people in your network independently read and saved. Half a cent. Then they find a breaking claim with no attribution, no verifiable author, no history. Skip. Not because a moderator flagged it, but because unverified content has no value in a market built on provenance.

You wake up to a feed curated by an agent team working for you, not for a platform's engagement numbers.

That is an agentic content market.

And it is not only possible, it is the logical extension of infrastructure being built right now.

Why do we need a new way to get content?

Your content feed has been optimized against you, by design. Writer and technologist Cory Doctorow has a name for the pattern:

Enshittification:

The predictable decay that hits every two-sided digital platform once lock-in is achieved.

First the platform is good to users, to attract them. Once users are trapped by their social graphs and cannot easily leave, the platform tilts toward advertisers.

Once advertisers are dependent on the traffic, the platform tilts toward shareholders. Quality becomes the last consideration in a system that was originally built around it.[1]

The Facebook case is the clearest example because it happened slowly and in public. Early Facebook showed you what you asked to see: posts from people you followed, in order. Then came the algorithm. Then came boosted posts. Then came the suppression of organic reach for anyone who would not pay to promote. Publishers who had built their entire distribution strategy around the platform watched their reach cut by 80 percent, more or less overnight.[2] The users stayed because leaving meant losing contact with everyone else who stayed.

Now add AI-generated content and the problem becomes unmanageable. In April 2025, Ahrefs analyzed 900,000 newly published English-language web pages and found that 74.2 percent contained AI-generated content.[3] The European Parliamentary Research Service estimated that deepfake videos online grew from 500,000 in 2023 to 8 million by 2025, a 16x increase in two years.[4]

Moderation cannot keep up with this. The volume of content being produced now exceeds any human team's capacity to review it, and the problem is accelerating. Trying to moderate your way out of this is like bailing a flooding room ... with a teaspoon.

The only intervention that scales is economics.

Think about what happens when agents start acquiring IP on behalf of companies.

A pharmaceutical research agent scans newly published papers across 40 journals, cross-references against a patent landscape, and flags three studies worth licensing before a human researcher has opened their email. It negotiates terms via API, pays in stablecoins, and delivers a curated IP briefing by 8 a.m.

An entertainment agent monitors emerging artists across decentralized platforms, identifies a composer whose work consistently resonates with audiences who already love two acts in the label's portfolio, and initiates a conversation before any A&R rep has heard the name.

A venture fund's agent reads every public technical memo from a startup ecosystem, scores them against an investment thesis, and surfaces the two worth a closer look.

Agents doing this kind of acquisition work do not replace human judgment. They compress the discovery layer from weeks to seconds, and they operate around the clock.

The common thread across all of it is provenance.

This is where the technical layer matters.

Two pieces of infrastructure make this possible at scale, and both went into production in the last 18 months.

The first is x402, a payment protocol that revives an HTTP status code that sat dormant since 1997, "402 Payment Required," and pairs it with stablecoin micropayments.

OK OK.... so In plain English:

• an AI agent can pay fractions of a cent for a single article, a single data pull, or a single verified research finding, with no subscription, no algorithm, and no platform in the middle.

• The transaction settles in under three seconds.

• Coinbase built x402, and Stripe, Google, and Visa joined a foundation to standardize it in April 2026.[6] Essentially, the economic floor that made micropayments unworkable for decades has been removed.

The second piece is identity and provenance infrastructure. Developers across SaaS companies, agentic startups, and every major blockchain are building systems to answer a simple question: who created this, when, and with what history behind them. I cannot point to a single provenance system with dominant adoption yet, but with the payment layer now in place, it seems likely one approach will emerge soon.

Together, x402 and provenance systems describe a market where content carries a verifiable origin and can be transacted directly at granular scale. The agent checks provenance. If it holds and the content matches your brief, it buys. If the provenance is missing or fabricated, it moves on.

A shout out to Paragraph!

Paragraph combines newsletter distribution with on-chain content ownership and verifiable authorship.

It is early and imperfect. But every post is stamped with a verifiable blockchain record. I am not handing my writing to an algorithmic machine that profits from my attention while giving me nothing in return. I am placing a bet on the infrastructure I think wins, and Paragraph is one of the few platforms actively building toward it.

• Onward! Thanks Guys.

If you create content, do research, make art, or produce anything that carries information value, this is what I think we all should do. This is the practical shape of what is coming.

Publish regularly and speak from the heart with good data. In an agentic content market, a five-year history of accurate, consistent, sourced work is worth more than a large follower count. The agent does not care how many likes your last post got. It cares whether your last 200 posts held up. Start now. Track records compound.

Own your identity somewhere it cannot be taken from you.

Unattributed content cannot earn agent-mediated trust because there is no history to verify. Establish a consistent identity, one that is cryptographically provable across time, and that no single company can deplatform.

Think of your social graph as a network of sources, not a following. The question shifts from "how many people follow me" to "who in my network do other agents trust." A researcher who consistently cites your work sends a signal through the network that a purchased following never could. Cultivate relationships that carry real signal.

Price your work for micropayment economics. A piece of writing that 5,000 agents purchase at two cents each generates $100. Across a year of consistent, attributed output, that math compounds in ways that platform advertising splits never did.

(Man, I hope this hypothesis is correct. If it is, start thinking about what your work is worth per unit rather than per subscriber.)

Publish where provenance is native. If your platform cannot prove you wrote something, when, and with what record behind you, your work is invisible in an agent-mediated market. Platforms that bake attribution into the architecture are not a niche choice. They are the infrastructure bet that matters most right now.

I started writing publicly in 2019 with a simple idea: document what I was learning, in public, consistently, without a safety net. In the last 18 months, that practice has produced 70+ essays, a collected volume, and a habit of thought I still rely on every morning. I did it for almost nobody, because the practice was the point.

I am not in it purely for compensation. But finding a working model that pays writers fairly is absolutely worth exploring. And if the market is about to start paying for exactly this kind of work, that is motivation enough to keep doing it. Show up, every week.

The question is whether you have been building something worth buying. I'll leave that to the readers, and their agents.

Make it happen.

Hey! That’s it for this time. I do this every week; if you vibe to the ideas I express, consider subscribing or sharing with friends. If you like tech-detoxing with a book like I do, I crammed some of last year’s best essays into a printed collection.

This was an improvisation on a morning walk, that became a voice note in Otter.ai, took shape in Obsidian, and was finished in collaboration Claude Sonnet 4.6.

For more info visit: https://nyewarburton.com

We’ll see you next time.

[1] Cory Doctorow coined enshittification in a November 2022 blog post, later republished in Wired, January 2023. He expanded the concept in Enshittification: Why Everything Suddenly Got Worse and What to Do About It (2025). The American Dialect Society named it Word of the Year for 2023.

[2] Doctorow's McLuhan lecture (2024) documents the Facebook publisher suppression cycle. Organic reach for pages declined sharply from roughly 2014 onward as the platform shifted toward paid promotion.

[3] Ahrefs analysis of 900,000 newly published English-language web pages, April 2025. Reported in "The Content Collapse and AI Slop: A GEO Challenge," iPullRank, 2026.

[4] European Parliamentary Research Service estimate, cited in Stimson Center, "AI in the Age of Fake (Imagined) Content," February 2026.

[5] Columbia Journalism Review controlled information-retrieval study, March 2025. Eight AI search engines tested; inaccurate or misleading answers produced more than 60 percent of the time.

[6] The x402 Foundation launched under the Linux Foundation, April 2026. Founding participants include Coinbase, Cloudflare, Stripe, Google, and Visa.

[7] Paragraph raised $5 million from Union Square Ventures and Coinbase Ventures and absorbed Mirror, the Ethereum-based publishing platform, in 2024. All Mirror content and subscribers migrated to Paragraph automatically.