Infinite Content, Finite Trust
What Barnes & Nobles' turnaround can tell us about the future of media
Hey team!
Hope you all survived a whole week without Uncredentialed! (kidding)
I’m pumped to be back and sharing my thoughts with y’all today. This week’s piece centers around a question that’s been rattling around the back of my head for awhile:
What happens when AI drops the marginal cost of content creation to 0 and volume goes through the roof?
Our feeds will be flooded with more and more content focused on whatever niches interest us. Always-on AI agents will be monitoring dialogues in the space and quickly producing content on it. As compute costs fall, the number of AI agents on any given platform, whether working on AI-generated media, AI-driven algorithms, or AI engagement, will begin to far outnumber human users, naturally driving a mean reversion of media feeds that struggles to reward authenticity and insight.
A mean reversion would limit the depth of niches, reducing advertising value, ultimately hurting the platforms content is being shared on. My proposed solution was inspired from reading about how Barnes & Nobles turned around its fortunes; something I’m calling the age of curation.
But first if this is your first time reading Uncredentialed, welcome! Subscribe below to receive weekly updates on my thoughts about tech, startups, and strategy!
The Barnes & Noble Turnaround
Walk into a Barnes & Noble (B&N) today and the air is distinctly different than a decade ago. In 2016 it felt like the business was going the way of Sears, Toys R Us, and many other previously great businesses that couldn’t survive the internet transition. Stores were closing left and right and would continue to dwindle until falling off a cliff during Covid.
But then, in the past couple of years, their footprint started growing, and growing dramatically at that. Since 2024, their number of storefronts have been growing by double digit percentages year-over-year and the expectation is that will continue into 2026. Anecdotally, my hometown’s store closed in May of 2020 and had been near-empty anytime I visited for a half-decade before that, but reopened last year and seems to regularly be packed.
So what happened? How did a well known brand like B&N get so close to the brink of death and what changed?
For years, B&N operated as a retailer. Using a model similar to grocery store slotting fees, publishers would pay for preferred placement in the store. Every shelf was on sale and store employees had little to no say in which books to keep stocked or where to place them. Each store felt more or less the same, essentially commoditizing the experience.
Customers recognized this and voted with their wallet with a barbell approach. If they wanted maximum ease and comfort and didn’t need a unique experience, they may as well just order off Amazon from the couch and if they did value personal touch, they’d shop at a small, independent bookstore where the selection is curated.
B&N’s CEO, James Daunt, was appointed in 2019 and sought to bring a small, independent bookstore feel to the company by giving stores their own autonomy. First, paid placement was dropped as a strategy, leaving stocking and placement decisions up to the local stores themselves, giving them a unique and personal touch. Next, they pivoted from a policy of hiring experienced retail managers to run their store to promoting from within. Not only were the stores given autonomy, but the leaders tasked with those decisions were now being selected based on passion for books and the business.
After bearing down through the retail whirlwind that was Covid, Daunt’s strategy began to shine, growing their retail footprint for the first time in nearly 2 decades in 2023 before the explosive growth of 2024 and beyond.
Media in the Age of AI
The B&N story is interesting on its own, but the reason I’ve been thinking about it so much lately is because I think it’s a preview of what’s about to happen to media, just on a larger and faster scale.
Every major social platform was built on an implicit promise of showing you things you’re actually interested in. Originally, this was done through chronological feeds of the accounts we chose to follow, with platforms making money on occasional interspersed ads.
After acquiring a lot of customers, though, these platforms had an LTV problem. Time spent on the platform was too low and ads didn’t convert enough. So out came algorithmic feeds, where content from accounts we follow were mixed with content from other accounts the platforms thought we might like, keeping us engaged on the platforms longer and providing better signal on what ads to serve us.
To understand how AI breaks this model, I think it might be instructive to consider the 3 pieces of the chain where AI will increasingly be integrated: creation and consumption.
Creation is the most obvious. When the marginal costs of producing content fall off a cliff, supply becomes functionally unlimited. The constraint that naturally kept feeds manageable disappears, replaced by an endless stream of fairly generic content optimized for surface-level signals.
Consumption, while less obvious, is much more destructive. The doomsday scenario of bot-mageddon – a world where all posts are boosted by hundreds or thousands of likes, comments, and reposts that cost the creator effectively nothing is obviously a bad one, but I would argue even the best case is problematic. I believe OpenClaw has given us a preview of the future agentic world. Most people will have always-on agents scrolling feeds to learn, interact, and share with you the most important info it sees. The issue is, for the vast majority of people, those AI agents won’t be converting on the ads they’re served.
Advertisers pay a premium to reach specific audiences with genuine interest in specific topics. That premium depends on 2 things being true: the audience is real, and the signal is clean. AI erodes both. Humans who will actually convert become a smaller and smaller fraction of total users, reducing what advertisers are willing to pay for a slot, and ultimately hitting these platforms’ cashflows.
Built for Curation
With advertising becoming less of a success driver for media platforms, they’ll have to find other paths to monetization. Essentially, instead of an audience of users that subconsciously open the app when they get bored, platforms should pursue users who consciously open the app because they see value in it. In my opinion, the best path to this is through building trust and loyalty between users and creators by allowing creators to curate and recommend content to their following.
Now let’s think about the range of platforms and who’s best suited to adapt to this.
On one end of the spectrum lies TikTok and Instagram. They are the epitome of platforms that aim to keep you scrolling and scrolling. You don’t open up either app seeking something in particular, you do it to scroll and be entertained. The high volume throughput allows them to have highly tailored algorithms that serve you exactly what you want to see.
Most importantly, in many ways their incentives work directly against building trust with individual creators. If you’re visiting a specific creator’s account, you’ll see what you came for and leave after, but if you keep scrolling through your never-ending “For You” page, the next thing you know it’s an hour later and you’ll still be scrolling, giving more surface area for ad revenue.
Next on the list comes YouTube. It’s a step up because content is longer and viewing tends to be more intentional. I’ll open the app sometimes just for entertainment but also a lot of times to see the new video an account I like posted. Thanks to the Google ecosystem, they also have a more diversified data moat that should better withstand noise from AI agents, meaning ad targeting should deteriorate more slowly.
The company seems to recognize and be adapting to the increasing importance of loyalty with the release of new features like the “hype” button and moving featured channels (essentially accounts curated by a creator for fans to go check out) to channel homepages. However, though they do offer premium subscriptions, at their core they’re still an ads business and want to keep viewers watching.
In my opinion, the next platform on the spectrum is Twitter/X. While they originally pursued an ads business model, the platform was never able to monetize ads the same way the previously mentioned platforms were. There’s a myriad of reasons for this ranging from the text-based format to the combative posture of users on the platform and more, but the important insight is that they’ve pivoted to leaning into a subscription model.
With a subscription model, they need to convince users that they’re getting enough value out of the platform to fork over their own money every month. This transition is already visible in how X is redesigning its toolkit. By doubling down on features like Lists, Articles, and the recently launched Starterpacks, the platform is essentially building a human-only layer on top of the algorithmic noise.
Starterpacks in particular represent a pivot toward human-vetted discovery. Instead of letting an AI guess what you might like based on a 2-second hover over a meme, these allow you to import an entire social graph of trusted voices in a specific niche, curated by editors or influential peers, in a single click. Now, I do think X still wants to keep you scrolling, it just is more okay with you getting what you want, instead of leaving you eternally in need of the dopamine hit of one more scroll like TikTok and Instagram.
Lastly comes Substack, where the rules were built for curation from day one.
Substack’s business model is perhaps the most perfectly aligned with trust. They’ve stayed out of the ads business and don’t even offer subscriptions directly to the platform. Instead, they only make money when a user trusts a creator enough to pay for their work.
This has led to the entire platform being built around trust. Restacks, default public likes, and, most importantly, their recommendations platform all serve as a mechanism for creators to lead their trusting subscribers towards other creators that they trust, building a positive, trust-based feedback loop.
In a world of infinite, free content, trust as a filter becomes in-demand. Substack has essentially built a platform where the filter is part of the product. By removing incentives to hack the algorithm for views, they’ve cleared the way for a reputation-based distribution system.
Curation Economy in Practice: The Turpentine Model
We’re already seeing the market place a massive premium on this “curation capital.” Just look at the rise of Turpentine, the podcast and media network founded by Erik Torenberg.
Torenberg recognized early on that the most valuable audiences are the high-signal ones, not the largest ones. In response, he built a network of deep-niche shows covering AI, corporate finance, nuclear energy, and more, hosted with practitioners and experts. By focusing on curated expertise over mass-market appeal, Turpentine became a primary vehicle for people in tech to stay informed.
The ultimate validation of this model came with its recent acquisition by a16z. Why would a top-tier VC firm buy a media network? Because in an era of zero marginal cost content, owning the curation is owning the funnel. a16z realized that if they own the most trusted voices in a niche, they don’t need to compete for attention on noisy, bot-filled social feeds; the talent and the deals will come to them because they provide the signal.
From Volume to Vetting
The Age of Curation is a return to form. Until recent decades, quality was key, but the internet moved the focus to volume. How many people can we reach, how frequently, and how cheaply? The issue is AI will be too good at that game.
As we move into 2026 and beyond, the value will shift from the creator to the curator. Just as James Daunt saved Barnes & Nobles by letting local booksellers pick what goes on the shelves, the next winners in media will be the platforms and individuals who act as high-fidelity filters.
The mean reversion of AI content is coming. When the feed becomes a mirror of itself, when it’s generic, optimized, and empty, we will go back to basics. Trust is on the rise.
Thanks for reading Uncredentialed! If you got this far, please click here to let me know so I can learn more about my read-through rates! And if you enjoyed the read, consider sharing with a friend and subscribing if you haven’t yet!
To get connected with me:
📧: matt [at] theuncredentialed [dot] com
LinkedIn: @matthewmjensen
Good insight 😃. Can i translate this article into Spanish with links to you and a description of your newsletter?