Hey friends,

It would seem that “AI” has been the buzzword of the day for approximately 1,259 days at the time of writing. ChatGPT made its way into the cultural zeitgeist on November 30, 2022, and ever since, we’ve been in a cross-industry, unilateral fervor over what we’re going to do about AI. 

From publicly traded SaaS companies to higher education, we’ve seemingly all lost our ever-loving minds made huge pivots away from what we’ve historically prioritized amidst our futile efforts to find clarity and actualized promises within this technology. 

Over the ensuing months, we’ve been sitting in the digital cuck chair, watching with a mix of horror and amusement as AI reshapes our online experiences. 

The “Dead Internet Theory” has evolved from being a discussion at the margins of obscure forums to being front and center with how we discuss digital platforms. Once treated as a conspiracy theory, the dead Internet theory asserts that the internet has become predominantly, if not entirely, populated by AI-generated content, bots, and algorithmic curation such that genuine human activity is minimal, with AI agents creating fake engagement to manipulate consumer behavior and social trends. 

And if you listen to this episode of The Interface about influencers being paid to peddle AI talking points via shady dark money exchanges… you might also become a dead Internet conspiracy theorist, if you’re not already.

Additionally, publications, bloggers, and writers have felt a huge shift in how people find their work. Among marketers and in SEO circles, we’ve talked a lot about the rise in “zero-click searches,” which is essentially where someone goes to Google, types in their query, and then gets back an answer in AI overviews or similar that prevents them from needing to click into any of the links on the search results page. Some estimates suggest that around 60% of all searches now result in zero-click behaviors, meaning that, across the board, publications that once depended on ad revenue that they earned by having people come to their sites and see the ads are now seeing their traffic drop off a cliff as people turn to LLMs and AI overviews for information and entertainment instead of their publications. (Insert GIF from The Devil Wears Prada 2 of Anne Hathaway shouting “Journalism still fucking matters!” here.) 

AI platforms haven’t just diverted traffic away from publications. Many have also been found guilty of copyright violations for training their models on protected works without compensation. Just last year,  in August 2025, Anthropic agreed to a $1.5 billion settlement to resolve a class-action lawsuit over its use of copyrighted publications in its training data. $1.5 billion is a huge sum of money! Except, not really. Anthropic was valued at $380 billion in February of this year and is currently eyeing raising another round of capital that would put their pre-IPO valuation in the ballpark of $1 Trillion. With a T. If $1.5 billion seems like an impossible sum of money (it is) consider that it is only 0.15% of $1 Trillion. 

New challenges equal new opportunities, right? Last year, Cloudflare– one of the largest cloud platforms and content delivery networks in the world– announced a new private beta, pay-per-crawl, that would allow site owners to charge AI crawlers and scrapers to access their content. Okay, that seems promising. If publications are losing on ad revenue but AI models need their content for training data, they can just charge AI models to access their content, right? Well… maybe. The pay-per-crawl Cloudflare beta was announced in July of 2025 and is currently in a closed private beta state, meaning that its initial testers still have access to it, but nobody else does. I haven’t seen any research or readouts on its success or how effective it has been. Speaking as a product manager, if a feature stays in a closed private beta for a year, that’s a pretty big sign that it hasn’t found product-market fit and needs to be rethought. 

It would seem that pay-per-crawl may be going the way of LLMs.txt files, a proposed web standard that would instruct LLMs how to interpret and access a site’s content… good in theory, but not adopted in practice. With LLMs.txt files, research has shown that creating one has zero impact on a site’s visibility in AI search or its overall traffic. It turns out, hyperadvanced AI models don’t need you to spell out instructions for how to access your content. They just do it. 

(“Just do it” may be Nike’s slogan, but even the shoe company Allbirds has made a pivot to AI. As of April of this year, Allbirds has pivoted from selling sustainable footwear to becoming an AI compute infrastructure provider. I’m not even joking.)

As I write the saga of AI fervor, I find myself staring into the face of systemic problems that can only be meaningfully curbed with systemic fixes. Legislation. Crackdowns. But I’m also aware of how unlikely it actually is for legislators in the United States to do anything (period) but especially that would be seen as curbing economic growth or yielding technological dominance to China. In a recent opinion piece for The Next Web about the value or preserving play in a culture obsessed with hyper-optimization, Zander Phelps remarks aptly, “So where does sovereignty begin? Not in Washington or Silicon Valley.” 

Instead of waiting for Cory Booker to do another 30-hour speech about nothing or for Chuck Grassley to release a statement about trans immigrants being to blame for economic disruption, I’ve been asking myself what the decay of digital spaces means to me and why I still choose to write on a digital platform if it feels like a losing battle. 

For one, I enjoy it. Writing is thinking, and slowing down to form my thoughts about how the cultural oeuvre intersects with creativity is meaningful for me, even if nobody reads it. Even if machines steal it. Even if it makes me a target in the future robot wars of tomorrow for being a dissident.  

But, beyond that, my hope is that the internet isn’t fully dead yet. That there are still readers who value slowing down, being thoughtful, and reflecting on what it means to live a creative life. And I want these conversations that start online or in your inbox to become seeds for conversations that happen in person (or, at least, face-to-face, as mediated by Zoom). I like to think that this newsletter will grow into an oasis in which like-minded folks can congregate– in which analog thinkers can communicate and collaborate with each other. I’ve been adding the tagline Reflection is resistance to the footer of this newsletter for a few weeks now (including a few weeks where I left in a typo and spelled it “resistence” but didn’t catch it because it’s an image from Canva, not inline text that’s easy to edit) because that is the ethos of the community I hope to cultivate by being an obstinate resistor of the dead Internet. 

I’m human and I’m still here. Are you? 

"The Information Diet Isn't Enough. What We Need Is an Information Village." by Hana at Card Catalog

The "information diet" has been the go-to framework for managing media overwhelm since 2012 — the idea being that what's broken is your personal consumption habits, and what's needed is better discipline. Hana pushes back on that with a reframe that's harder to dismiss: the problem isn't individual willpower. It's that we've lost the communal context that used to give information meaning in the first place. The village gave you trusted interpreters. The diet just gives you another thing to fail at.

New research published this year suggests that certain trauma-related biological changes in children can actually be reversed — not through medication or technology, but through consistent, face-to-face relational care. Child psychiatrist Pamela Cantor, writing at Biology of Becoming, draws on attachment science to make the case that human connection isn't just emotionally restorative; it's neurochemically necessary. The antidote to what's breaking us, it turns out, is each other.

Find something you made before the internet felt broken — a journal entry, an old email, a half-finished story, a forum post, a playlist, a photo you took because you wanted to, not because you were posting it. Sit with it for ten minutes without doing anything else. Then write one paragraph about who that person was and what they were paying attention to. Not as nostalgia. As evidence.

"The past is never dead. It's not even past." — William Faulkner