When big tech is saving democracy and the fourth estate
Accountability journalism in Minnesota, prediction markets funding newsrooms and fueling disinformation, Chat GPT replacing local news and 21 active calls.
Welcome!
This week on Media Finance Monitor
When big tech is saving democracy and the fourth estate
Prediction markets are funding good journalism (and fueling disinformation)
We are bestsellers on Substack 🥳 and about to launch a new vertical🎉
How Chat GPT is replacing local news
21 active calls (1 new)
Media Finance Monitor is where newsroom leaders, funders, and media entrepreneurs come to understand how money really moves through the information ecosystem.
We decode funding, business models, policy, and platform shifts and turn them into strategies you can actually use.
Paid subscribers get premium editions, full access to the archive and all our research reports, for about the cost of a pistachio éclair and a good espresso a month.
For deeper engagement, our Partnership Program offers one-on-one strategic advisory, curated briefings, and access to our internal databases and invite-only events.
Join as a free or paid subscriber.
When big tech is saving democracy and the fourth estate
(by Peter)
There is a specific kind of intellectual comfort in the “Fourth Estate” model of journalism. A very prestigious newsroom with serious journalists wearing serious suits looking at serious news and holding the powerful accountable. Back in the late 1990s I got into journalism in no small part because of this “All the President’s Men” branding. But the reality of how we now hold power accountable is shifting toward something far more decentralized.
After ICE agents killed Renee Good and Alex Pretti in Minnesota, the Trump administration’s response was blunt: the victims were agitators, domestic terrorists, they were out to harm federal agents who acted in self-defense.
These narratives collapsed entirely, within days. Not because reporters were embedded with immigration enforcement, not because of some well-placed source within ICE or the administration leaking insider information, not as a result of freedom of information requests by some newsroom. These false narratives collapsed because ordinary people standing nearby had smartphones.
We spend a lot of time in this newsletter cataloguing technology’s corrosive effects on the information ecosystem. Algorithmic amplification of garbage, platform dependency, the hollowing out of news economics, AI-generated slop flooding the zone. Maria’s piece below on OpenAI and local newsrooms adds to that ledger. The case against Big Tech is long and well-documented.
But Minnesota reminds us (okay, at least it reminds me) the equation is more complicated than that.
This particular act of accountability journalism required a confluence of technological developments that simply didn’t exist ten-fifteen years ago: phones with cameras good enough to capture usable evidence; batteries lasting long enough for extended recording; bandwidth making uploads instant and frictionless; platforms enabling rapid distribution. Each component was necessary. Together, they created conditions where anyone present at a newsworthy event can become a primary source of documentary evidence.
This is the fourth estate function in its most classical form: exposing state lies about state violence. But the production model underneath it is no longer classical at all.
The footage from Minnesota was damning, but it wasn’t self-explanatory. You could watch the videos and form impressions, but forming impressions and establishing facts are different things. This is where I think institutions might have a competitive edge. The New York Times’ frame-by-frame visual investigation (reconstructing the sequence of events, mapping movements, demonstrating exactly when and how the administration’s claims diverged from documented reality) required specialized expertise that most individuals simply don’t possess. It required time, methodology, and the kind of forensic attention that turns raw footage into evidence.
Verification, in other words, remains an institutional strength. The technology to capture newsworthy moments has been democratized; the technology and expertise to rigorously verify them has not, at least not yet. For publishers trying to figure out where they fit in an increasingly decentralized information ecosystem, this matters.
But a competitive edge and a sustainable business model are not the same thing. The New York Times’ visual investigations unit is extraordinary and it’s also a small part of a mega-machine sustained by Wirecutter affiliate revenue, The Athletic’s sports subscriptions, cooking apps, and games. The analysis that held the administration accountable is subsidized by product reviews and Wordle. Verification may be what makes institutional journalism irreplaceable; it’s unlikely to be what pays the bills alone.
Prediction markets are funding good journalism (and fueling disinformation)
(by Peter)
If you watch closely, there’s some fascinating word magic happening right now: gambling becomes “prediction markets,” bookmakers become “truth machines,” and suddenly it’s acceptable for CNN, CNBC, and the Wall Street Journal to partner with what are, functionally, betting platforms.
In December, CNN struck a deal with Kalshi to integrate prediction market data across its television, digital, and social channels. Days later, CNBC announced its own exclusive Kalshi partnership. Yahoo partnered with Polymarket, the other major player, in November and in January Dow Jones also announced an exclusive deal with them, bringing its data to the WSJ, Barron’s, and MarketWatch. The mainstreaming is complete. Gambling has put on a suit.
This isn’t entirely new territory. Sports betting has fueled sports media for years, with different countries drawing the line in different places. In Bulgaria, 2024 restrictions to betting and related advertising shook parts of the news sector quite significantly. The media market is brutal, and I understand, even if I don’t endorse, how publishers ended up accepting gambling money to stay afloat. But the power of the rebranding is amazing. “Prediction markets“ lets highbrow financial publications do what they might have hesitated to do with DraftKings.
The ambition is also remarkable. Kalshi CEO Tarek Mansour told a Citadel conference in December that the company wants to “financialize everything and create a tradable asset out of any difference in opinion.“ If that isn’t the kind of sentence you’d find embroidered on a late-capitalism throw pillow, I don’t know what is.
While I’m sure these deals bring new resources to important journalistic institutions (and also provide new relevance for some audiences), the entities behind them are also fueling disinformation.
Axios published a detailed account this week of how prediction markets have become engines of viral misinformation, pumping out false, misleading, and context-free claims optimized for engagement rather than accuracy. Polymarket falsely attributed quotes to Jeff Bezos. During the Minneapolis ICE chaos, Polymarket claimed deportations would cost Minnesota a congressional seat, a post amplified by the Department of Homeland Security before being deleted. Kalshi falsely suggested the US and Denmark were in “technical talks” to buy Greenland.
This is having a cultural moment partly because the current US administration has embraced crypto, and embracing crypto means embracing the prediction market ecosystem that emerged from it. The regulatory environment is permissive, the valuations are soaring (Kalshi just raised $1 billion at an $11 billion valuation), and media partnerships are how these platforms buy legitimacy and potentially new customers.
For Europe, the picture looks different, at least for now. The EU’s regulatory regimes (and perhaps the cultural traditions around gambling) are more rigid. This usually is my complaint, since it means slower adoption of new things. But here, maybe conservatism has an upside. Europe isn’t rushing to integrate betting odds into its news coverage (and it currently isn’t governed by a faction enthusiastically financializing every difference of opinion). I’m not usually the one making moral arguments about funding sources, but I do think there’s a hierarchy, and gambling money sits near the bottom. Calling it something else doesn’t change what it is.
We are bestsellers on Substack 🥳 and about to launch a new vertical🎉
(by Peter)
This is a media product about media products, so there is inherently a high level of navel-gazing in every edition. Because of this, we try to keep news about ourselves at a minimum. But this week is an exception to that rule because
we officially became “bestsellers” on Substack
we are launching a new vertical next week.
The first is more straightforward: we have been growing our paying subscriber base steadily since the end of last year (if you need me to convince you to pick up a paid sub, read this abstract piece from December: “I want your money”) I’m not sure if the bestseller badge comes with a lot more than a boost to our vanity, but it still feels very nice and it literally could not have been possible without you, so sincerely, thanks for your support.
If you’ve ever heard me speak in any context or read anything I’ve written about what makes media work, you know I’m obsessed with focus. Laser focus. Depth over scale. Pick your niche and serve it relentlessly.
So naturally, we’re launching a second newsletter for an entirely different industry.
Introducing Plancha, a newsletter for people who run restaurants, hotels, and hospitality businesses in Central and Eastern Europe.
Food has been my other obsession for as long as I can remember. I love eating it, thinking about it, writing about it. Some of my favorite pieces I’ve ever written have been about food, how ramen went from post-war survival food to global staple, how Franco-era politics accidentally turned San Sebastián into the culinary capital of the world. And through personal circumstances, I’ve ended up surrounded by people who do this professionally: chefs, restaurant owners, operators.
This newsletter, the Media Finance Monitor, started as a hobby. Now it has sponsors and well over a hundred paying subscribers and we take it quite seriously. That success made me think: what if we could build something similar for hospitality?
Plancha is the attempt. Utility first. Helping owners and operators make better decisions about their businesses, understand their customers, and stay ahead of what’s changing in the industry. No PR dressed up as journalism. No trend pieces disconnected from the reality of margins, staffing, and difficult suppliers.
I’m not doing this alone. I have excellent co-conspirators: Zsofi Bajor, Georgiana Ilie, Alexei Korolyov, Miklos Rozsa, María Paula Ángel Benavides - journalists and industry people who actually know what they’re talking about.
The site is live. The first issue goes out Wednesday, February 11. We’re starting with Austria, Hungary, and Romania, with plans to grow. It’s free to sign up. If you’re in hospitality, or know someone who is, I’d love for you to check it out.
(Don’t worry, this newsletter will still focus on media & money and honestly, I don’t expect many crossover editions.)
How ChatGPT is replacing local news
(by Maria)
So, as Peter explained above, we are launching Plancha next week, and for the first edition, I needed data from the Google Places API. A year ago I wouldn’t have known where to start. This time, I asked Gemini and got what I needed without really understanding the coding required. I did not go on Stack Overflow or any other website, just explained the task and AI gave me the information.
News consumption is following a similar pattern. According to a recent post by OpenAI, ChatGPT now handles about a million prompts a week asking for local news. People still care about what’s happening around them, but many of them no longer visit their local websites to get the information, they just ask their friendly neighborhood Large Language Model instead. There are no website clicks, no ad views, no news subscriptions. The interface keeps and monetizes the entire relationship. It’s not difficult to see how this makes local news production even more difficult than before.
Large publishers have more room to maneuver. Groups like News Corp, The Atlantic, Vox Media, and Condé Nast have signed licensing deals with OpenAI. Even if these deals don’t solve the underlying problem, they bring revenue and some influence over how their work shows up. Axios, while doing some local coverage, fits that profile. They’ve just signed a partnership with OpenAI to expand their work in nine new communities.
Smaller local outlets are in a different position. They produce a lot of the original information people are asking for, but the way people consume it no longer lines up with how that work is paid for.
The current situation resembles user-generated content on Facebook: some creators receive compensation through opaque, invitation-only deals, while the vast majority produce value and see nothing in return. OpenAI’s licensing agreements with major publishers follow this pattern, a few big names get paid, the system stays black-boxed, and everyone else is left guessing.
ProRata is betting on something closer to the YouTube model: a clearer, more systematic path to compensation for anyone willing to share their content. YouTube still concentrates enormous power relative to creators, Google sets the rules, controls the algorithm, takes its cut. But at least the mechanism is legible. Creators know how monetization works, even if they don’t control it.
That may be the best future local news can hope for: not ownership of the relationship, but a seat at the table and a visible path to payment. It’s not a solution, but it’s better than being free training data.
This last piece is part of a series focusing on local and community journalism and is supported by the LimeNet project and the European Union.
Here are the active calls, with the largest at the top: