Kudos to The Boston Globe for showing that there will always be a market for great local stories that are well told. Its viral story of a lesbian bar in western Massachusetts is out of a Woke 1 time capsule. It has now generated over 4,200 subscriptions, I’m told. The Globe put a $1 offer on the story for 26 weeks, which points to a smart way of using viral stories for sampling.
I’ve been Muse-pilled. This is the first AI product shipped that I can see becoming mainstream. I agree with Ben Thompson that there’s no mass market for productivity tools but there is for convenience. Meta smartly gave Muse suggestions that get around people not knowing what to do, and the early use cases revolve more around convenience than productivity.
On PvA: Breaker’s Lachlan Cartwright. We’ll discuss his reporting on Puck’s Redbird tie-up, 60 Minutes drama and fallout from the approval of Paramount’s Warner Bros acquisition. Get PvA.
On PvA Field Notes: Podcast breakdowns. We are using People vs Algorithms to not just connect the dots between media, tech and culture but also to incubate new media products for the rapidly changing information space. One of the features I find most promising is podcast breakdowns, where we disaggregate podcasts and allow them to be annotated at remixed. Check it out on Field Notes.
Trend I’m following: Gen Z maximalism. Younger generations always think their older siblings are dorks. Gen Z gravitating back to blinged-out MySpace design on Strawpages points to the open web living becoming far more interesting and creative. This page truly is bonkers.
Today, the view on what a good deal looks like as the open web gives way to the AI archipelago.
Shrink your list. Grow your impact.
Two publishing leaders cut their lists and boosted engagement, deliverability—and even advertiser value. This guide shows you how.
For years, we chased bigger lists. But bloated databases don’t drive clicks—they kill deliverability, inflate metrics, and mask real audience insights.
This guide dives into:
Two publishers who strategically shrank their email lists
How smarter segmentation boosted engagement and advertiser ROI
How to get internal buy-in and make the shift stick
If you’re questioning the value of your current list strategy, this is your playbook.
The final commoditization
The open web is being replaced by an AI archipelago of centralized AI chokepoints. If people get used to agents reading their email and shopping for them, they will use their agents to get most information. Selling access to a widely available information is a commodity business.
“If you need to have good AI, you need chips, you need power, and you need the information to run through them,” Jon Roberts, chief innovation officer at People Inc and point person for AI, in a session on the post-Google era at TRB Audience Summit last week. “The chips and the power give you a data center, but data center is the engine, the words and the information are the fuel. Two of these supply chains are well funded, one of them is not funded at all.”
The problem for publishers is scarcity. Chips and power have supply constraints. Enchilada recipes do not. AI companies already pay for specialized training data produced by doctors, lawyers and software engineers. Much of what publishers produce has readily available substitutes. An AI company might need a good enchilada recipe, but it doesn’t necessarily need yours. Information can be valuable without giving its producer much pricing power.
In the AI archipelago, publishers are commodity miners who pass along raw materials to AI agents to enrich with other raw materials, personalize, repackage and present as enriched products. The entity that controls the interface always extracts the majority of the value. The AI company gets the customer relationship, usage data, pricing power and ability to substitute one supplier for another.
To me, the question is what a “fair” economic bargain looks like, knowing that publishers will have neither political support for the next two years or much leverage. And just like every parent believes their children are remarkable, the reality is most are average. Publishers hav told me for 25 years that their content is premium despite mechanized ad systems tell them otherwise, at least from an economic standpoint.
Mike Reed, CEO of USA Today Co, has been the most bold in his public calls for publishers to consider exiting Google altogether. Such a move is drastic and Mike would only say on stage at TRB Audience Summit that Google-driven revenue is “getting small enough to where we can walk away from it soon.”
“My preference would be a combination of licensing and revenue share,” he said. “I think it's massive. And so I want to share in that revenue.”
Even bolder: “You could make an argument that for the companies that have built models on the back of our content we should own a piece of them as well.”
Needless to say, not highly likely to happen. This is tantamount to nationalizing these companies. I don’t see using a government-run Google in the cards. Maybe that’s Woke 2.
“We are not naive,” Mike said. “We understand consumers are gonna end up on other platforms and not ours. And and we're fine with the consumption of our content on other platforms as long as there's an opportunity for us to monetize that, whether it's licensing or advertising revenue shares.”
This is where the rub lies. I have no doubt we are in the midst of an interface shift. AI interfaces will emerge, just at GUI and web browsers emerged. Publishers are always downstream of these chokepoints and will need to adapt their models to pipe their information to a new interface that strips it of formatting. The issue I see is the rationalist Tech Brain sees publishing content as data. And much of that data is easily replicated and in the case of news has a steep decay curve.
“At the moment, it seems that the assumption is the internet is what's been written,” Jon said. “It's all that will ever be written. There's no new writing that ever needs to be done. So we'll just use it. But there's no ambition on the tech side to make a better product. They don't even seem to appreciate that there are new questions and that there will need to be new writing. It's wild.”
This points to the crux of the issue for publishers: Is there information commoditized or is it specialized? Right now, the AI economy sees much of what publishers produce as commodity information that isn’t that valuable. I don’t necessarily agree, but if what Jon is saying about the third leg of the AI stool being high-quality information, then these companies would have rushed to lock in agreements with publishers as suppliers. They haven’t.
My worry is there’s a fundamental misalignment. Publishers have been mostly in the business of running a full content supply chain. The news article, for example, is a refined information product. Many AI engines do not want refined products but the information stripped to its raw material. That way, they can reconfigure the content, mix it with other data sources and create their own finished product where they capture much, if not most of the value.
Nobody wants to be commoditized. Amazon can block Muse; I don’t think People Inc or USA Today Co has the same leverage.
That’s where I see the connection economy barbell coming into play. Licensing to AI interfaces will be a wholesale volume business. Most of these companies are shifting the weight of their efforts away from high volume information to customized products like events and communities.
Watch my discussion of the post-Google era with Jon, Mike and BBC chief subscription officer Melissa Chickering on YouTube. We will be publishing all main-stage sessions to the TRB YouTube channel.
Send me a note with your feedback by hitting reply. For sponsorship information, contact Richard Jamieson at [email protected].