TRB Audience Summit is on Thursday. We have reached capacity for the event. I’m very excited to see many of you for a day of conversation and connection. We will be holding this event in London early next year. Get in touch with Richard Jamieson for information on partnerships: [email protected].
Today, a conversation with Swedish publishing executive Martin Schiori, appraising the existential AI doom narrative and on Field Notes our new AI-enabled podcast curation tool, Mixtakes.
How Daily Maverick runs on Momently
Daily Maverick publishes for over 3 million monthly readers, runs 20-plus newsletters, and sells sponsored content to major advertisers. All of it is measured in one place: Momently. Every editorial section has its own live dashboard, every newsletter its own tracker, every campaign its own client-ready report, so editors, audience teams, and sales all answer their own questions without filing a ticket with the data team. Read how they built it, and what it changed.
The mushy middle of AI adoption
The growing backlash to AI presents a dilemma for publishers: These tools hold the promise of deep productivity gains at a time when the math underpinning the business has changed, yet many workers have deep hostility to the tools.
To date, publishers have taken baby steps with AI, putting chatbots on webpages, adopting AI tools for back-office coordination work. For the most part, the news production process is largely unchanged, certainly not like how AI has upended web and app development.
According to recent research by The Rebooting and Wordpress VIP, most investment by newsrooms into AI is around workflow efficiency.
Much of that reticence stems from the specter of lost trust from pushing out something deemed AI slop. Stanley Druckenmiller opened the Overton window on AI writing by getting the Wall Street Journal to OK AI-created content from subject matter experts who contribute opinions to its editorial page. The reality is any AI screwup will be magnified far more than a run-of-the-mill human screwup. Right or wrong, many distrust AI output and the entire project.
Martin Schori, a former AI lead at leading Swedish publisher Aftonbladet, sees news businesses becoming either AI-powered information services or deeply human products built around personality, trust and relationships. Most are hedging, with the risk that incremental changes leave them without a compelling proposition on either side.
“Most media companies are ending up somewhere in the middle trying to do a little bit of both,” he told me on The Rebooting Show.
At Aftonbladet, Martin’s seven-person AI hub experimented with internal tools, training and audience products. The biggest internal success was transcription, something I’ve heard from several publishers. Many tools intended to produce substantial efficiency gains were less successful, partly because the technology wasn’t ready and partly because they began with what the technology could do.
“Everybody’s gonna have these tools if they don’t have them already,” Martin said. “That’s not gonna be the issue. The issue is gonna be the culture, the change management. How do we adapt to this whole new reality? Because AI is not just some new tools, it’s a totally new way of how people consume information and journalism, and what’s our role and value.”
The strategic question for publishers is less in adopting tools for efficiency – every company will do so – it’s whether they pursue a direct-to-consumer strategy or become an information supplier to AI interfaces. This is an issue we will discuss at TRB Audience Summit. Nick Thompson, CEO of The Atlantic, and Anna Bateson, CEO of The Guardian, will sketch out what they believe the industry and their businesses will look like in five years. I suspect the biggest change will be publishers becoming API providers.
“Some companies will definitely be some kind of supplier to AI interfaces or whatever it looks like in the future,” Martin said. “Some of them will take the total opposite way and really inject more human-centered journalism, relationships, community building.”
Rally Rd. is a marketplace where investors buy shares in rare collectible assets. Inside the app, compliance rules limit what they can say. So the real selling happens in their newsletter, Shiny Things, a weekly long-form essay that runs on beehiiv and never actually asks readers to open the app.
The strategy, when paired with beehiiv’s all-in-one growth platform, is clearly working: newsletter subscribers convert at 3X the rate of standard users, with a 20%+ lift in investment behavior and stronger portfolio retention.
And between Q4 2025 and Q1 2026, two major institutional partnerships opened tens of millions of dollars in tier-one supply, both attributed directly to the newsletter.
Owned audience, built right, doesn’t just drive engagement. It drives businesses to new levels of success.
On a holiday weekend edition of People vs Algorithms, we discussed how AI On the latest episode of People vs Algorithms, we discuss the return of the AI doom narrative and whether it will have resonance as a theoretical. We have proven ourselves time and again as mostly incapable of grasping exponential future downside. See climate change and Covid. I doubt that will change without a scarier incident than the OpenAI-Hugging Face hack. We recorded the episode prior to the call for “pacing the frontier,” which has been met by a counter reaction as Silicon Valley splits between the safetyists and the maximalists. The American spirit, the current administration and a fractured information space where trust is a scarce commodity means the smart money is on AI development plowing ahead without much of a slowdown.
Plus:
Cognitive struggle is now optional. The leverage of resisting the easy way out. Just as the shift to office work gave rise to the gym, the commodification of intelligence requires workout routines.
The half empty-half full AI jobs impact. The AI jobspocalypse didn’t happen. Yet labor’s share of GDP is in decline.
The analog backlash. From growing vinyl sales to new print magazines to packed stationary stores in Greenpoint, signs are afoot of an analog revival.
AI assistants could break through. The AI consumer product story has moved rom chatbots to agents to AI assistants. Meta’s Muse shows the breakout AI consumer product could be a dead-simple assistant.
Introducing Mixtakes
We added a new feature on PvA Field Notes: Mixtakes. These are reaggregated podcast segments on a particular topic. Soon FN users can make their own and comment on each segment.
Podcasts are interesting because they’re unlike other forms of easily remixed and aggregated digital content. The rise of clipping is a sign that that people want to break apart podcasts that are often a few good takes wrapped in a lot of yakking. (Not PvA, of course.)
What was interesting about Mixtakes is the feature was hashed out on the episode, sparked by a discussion on the platform about the era of mixtapes. People on FN contribute feature requests, which are captured by an AI concierge named Linda.
That’s pretty remarkable. And it points to how individuals and organizations can now create custom software.
Send me a note with your feedback by hitting reply. For sponsorship information, contact Richard Jamieson at [email protected].