Hello from rainy London; some patterns are reliable. Tonight, we are kicking off TRB’s Media Mixer series because, despite the many challenges, Media People need to find each other. Thanks to Beehiiv for supporting the mixer. We’ll be holding a NYC version on 3/19.
Today, some notes on the limits of pattern matching.
In my conversation with Omeda CEO James Capo, we discussed the ways publishers are attempting to incorporate AI into their products. James warned that too many are following familiar patterns of imitating big tech, in this case Open AI, by adding chatbots to their sites.
“I would challenge the folks in the media companies to think about the utility that AI can bring to your audience and less about trying to just copy what you see OpenAI doing,” he told me.
Pattern matching
I map cities by New York City neighborhoods. South Kensington feels like the Upper West Side, Knightsbridge the lower Upper East Side along Madison Avenue, Hackney as Bushwick, and Shoreditch as London’s Williamsburg. Most cities have a Williamsburg.
Patterns always emerge. They’re a handy way to make sense of the unfamiliar as we try, mostly fruitlessly, to predict the future. Now, at the dawn of an AI age, we again fall back to familiar patterns, only when we are confronted with something that is heralded as far bigger and unlike previous dislocations.
AI exec Matt Shumer added the latest entry into this growing self-interested oeuvre with a viral post ominously titled Something Big Is Happening. Shumer’s thesis is that we are back in early February 2020, when only a smattering of people recognized what was coming and raided the toilet-paper shelves while the rest were going about their lives without bother.
Shumer pattern matches what is happening to coding to other fields, including marketing and journalism. His reasoning: “AI isn't replacing one specific skill. It's a general substitute for cognitive work. It gets better at everything simultaneously.”
AI alarmism is spreading beyond the X fever swamps. Chris Hayes frames the coming disruption as part of a larger class war pattern. The oligarchic elite used globalization and deindustrialization to obliterate blue-collar jobs and will now do the same to the laptop class of white-collar workers. This lines up with the AImaxxing view that you have 18 months to acquire as much capital as possible before the AI hunger games truly kick in.
An emergent pattern in AI is the yawning gap between the true believers and, well, the rest of us. The AI Super Bowl – nearly a quarter of ads featured AI – included ads by both Anthropic for Claude and OpenAI for its Codex AI coding platform. Anthropic has smartly cast itself as the Good Guy in the AI race. CEO Dario Amodei regularly sounds the alarm that we are unprepared for the tidal wave of change coming. The ad itself was misleading. Without naming ChatGPT, it hinted ads would corrupt interactions with chatbots.
Regular people were unimpressed. According to iSpot, the ad ranked as one of the worst it's measured in the past five years of Super Bowl spots. The general public is still confused by AI while the TBPN crowd is jabbering about Elon putting data centers in space, spinning up OpenClaw and the End of Work.
I can see why OpenAI’s “You Can Just Build Things” ad didn’t land quite the same as Apple’s 1984 ad. AI’s “marketing crisis” is that its boosters have scared the shit out of regular people who do not want their lives overturned for the prospect of Amazon developing a mass surveillance system through its Ring doorbells in order to find lost pets. Doesn’t strike regular people as a great tradeoff. And people have caught onto the pattern of Big Tech promising innovations that turn dystopian.
AI-generated ads will pattern match on what’s working. The excitement ad executives have about AI-generated advertising – surely not shared by their ad creatives – is in disconnect with how consumers perceive AI-created ads, according to research by the IAB and Sonata Insights.
This pattern has long existed. Digital ad executives were always far more breathless about the benefits of targeted advertising than regular people who reliably found the practice creepy.
Silicon Valley’s many boosters have a now-familiar pattern of blaming media gatekeepers whenever their plans are unpopular. This go-to line is losing salience as the legacy media keeps losing power. The Washington Post’s deep cuts and turmoil has resonated across the industry. The Post isn’t Mashable. It solidified modern journalism’s identity as a firm check on institutional power through its Watergate investigation.
Gerald Baker, a Wall Street Journal columnist, detects a pattern of the type of journalism practiced not just at the Post but throughout the field. In his telling, legacy media has established a embraced a specific type of social activism that focuses on opposing Trump while ignoring the issues like mass immigration that elected him because they matter to regular people. The Covid era has never ended for these people. I always find kernels of truth in these pieces, only they’re all intellectualized versions of “go woke, go broke” X posts from anonymous accounts based in India. The woke horse has been flogged enough.
The real pattern is that picking a side in politics is a sensible business strategy to a point. Audience capture is a hack. The Post’s adoption of “Democracy Dies in Darkness” is painted by Baker as woke overreach, only it can also be seen as following an established pattern.
After all, The New York Times has grown more successful as it has grown more progressive over the years. It is closing in on 13 million paid subscribers. Its ad business is expanding, with digital ad revenue growing 25% year over year. Investors hate the Times spending big on video that CRO Joy Robbins told me it is deliberately under-monetizing, but it is laying the groundwork for expanding its key average revenue per user beyond its current $10 if it can crack video.
The Times' success is increasingly coming from its non-news products. That’s hardly surprising. News has long served as a habit product, and it is becoming often a front for other more lucrative businesses. The success of Baker’s own Wall Street Journal is impressive. It is also part of a Dow Jones portfolio that’s increasingly reliant on its B2B data businesses as a growth engine. In its latest results, Dow Jones as a whole had 8% revenue growth while its risk and compliance unit had 20% growth. The Journal is a hero product in that group. This is somewhat akin to the role Bloomberg Media plays in the Bloomberg LP portfolio that’s powered by The Terminal.
Those dynamics do not exist at the Post. It has a typical subs-and-ads business model that has a ceiling, particularly without choosing a side in partisan food fights. Ask The Messenger about how lucrative the lane is for centrist general news. The biggest missed opportunity was not taking an expansive view of the Post as Washington’s hometown paper by developing the prosumer model of Politico that spawned Axios, Punchbowl, Semafor, Puck and others.
The Boston Globe is a contrast here. Also owned by a billionaire, it stomached years of losses to build Stat, a B2B business for the life sciences industry that lines up with Boston’s strength in the industry. It added Boston Magazine to solidify its hold on the area. Maybe Washington’s too transient of a city for that model to work, or maybe Bezos is bored by that kind of sensible small ball. My read is he wants the influence of owning the Post without the risk to cloud contracts with the government. It’s one thing to have an advertiser pull an ad campaign, it’s another for Trump to scotch billions in government contracts.
We discussed all of these issues in this week’s episode of People vs Algorithms, out tomorrow. Get PvA.
Send me your feedback by hitting reply or emailing me at [email protected]