Known unknowns of media in an AI era

Into the void

Quick reminder: For those of you who decamp to Cannes in June – rough life, it must be said – The Rebooting is partnering with Kerv on three days of programming. I’m also lining up other events including a cocktail gathering at a fancy villa and a dinner at a fancier restaurant. Let me know if you want more information on these events.

Make email the workhorse of your CRM

Email is more than a distribution channel. With solutions like LiveIntent’s, you can transform email into a revenue stream. Learn how you can connect advertisers to your most engaged customers by leveraging first-party data and custom ad formats with LiveIntent.

Known unknowns of media in an AI era

Much is still unknown about how AI will develop and the impact it will have across industries and society, which makes it catnip for newsletter writers. With the tech changing so fast already and Silicon Valley fully focused on this as the for-real New New Thing – meta-what? I don’t know what you’re talking about – the familiar feeling of another cycle of disruption and destruction awaits.

That said, AI comes up regularly in my conversations, with some seeing threats and others an opportunity to shake up an industry that’s become rather boring. As one exec I met with recently put it, “There’s not much interesting out there.” The models of the previous era have clearly broken down, but it’s unclear what, if anything, will replace them as growth opportunities. The excitement of newsletters and niche media is admittedly rather narrow in scope and ambition.

In the spirit of Donald Rumsfeld, here are the known unknowns of how the media business will fare in an AI era, with the caveat that it’s usually the unknown unknowns that prove the biggest sources of change.

Will good enough remain good enough?

The cliche is most tech disruptions are underwhelming in the near term and overwhelming in the long term. AI is clearly in the upslope of the hype era. It too will reach its trough of disillusionment. This always happens. Facebook was one of the big winners of the shift to mobile – despite not controlling the operating system, a remarkable feat in retrospect – and yet it faced doubts it could adapt.

Many AI chatbots and applications remain somewhat clunky. ChatGPT itself produces what I would consider good-enough copy, with a strange tendency to use nonsense filler like “cutting edge,” a habit that says something about human descriptions of tech. Midjourney can still be janky to use within Discord. (Anyone under 40 is at risk of seizure after 10 minutes in Discord.)

The errors and omissions of AI will inevitably decrease. Smart publishers will see opportunities to train AI on their own proprietary information to provide similar experiences to the chat-like interfaces that AI is establishing as an expectation.

How will the chess board line up?

The battle for AI supremacy is a fight that’s well above the pay grade of publishers. This is a familiar position. Much of the last 15 years has been like a Game of Thrones episode in which various egotists running fiefdoms began wars of conquest while the villagers suffered the most brutal of consequences. That’s why publishers will turn to a delay and obstruct strategy by using their political leverage to plead for relief from courts, regulators and governments. That’s unlikely to halt the medium-and long-term impact felt by AI.

Search models shifting to a chat context has bleak implications for publishers. It is simply hard at the moment to see how the removal of the familiar friction of search – comb through a bunch of paid links and SEO’d to the hills links – will not be a net negative to publishers whose models depend on search traffic. Microsoft has been admirably candid that it spend up deployment of AI in search to force Google to “dance.”

This presents another situation where publishers will begrudgingly find that Google might be its most reliable ally, if only because Google has a vested interest in maintaining a semblance of the status quo. But, as Chris Best said on a recent episode of The Rebooting Show, business model is destiny. Ads in AI chat – maybe a Vibrant Media-style double-underline approach? – are going to be tough to pull off, and the prevailing winds are gusting toward subscription models after tech’s anomalous obsession with ad-driven models during Web 2.0.

Where are the safe zones?

One exec recalled early in their career visiting the floor where the papers were printed and the giant plates that were put in place by specialized workers. When print modernized and then assumed a peripheral role, we didn’t spend much time hand wringing those occupations evaporated.

There will be job destruction. Tech is nearly always a deflationary force because it is the best tool we have for increasing human productivity. The internet oddly didn’t make workforces particularly more productive, possibly because communication tech gave us the world’s knowledge at our finger swipes but the apps drove us to distraction with all kinds of mindless scrolling.

AI tools will accelerate the trend of smaller organizations. The future publishing model will be a lot leaner than in the past and even the present. AI will erase many of the lasting advantages of scale, as small groups of individuals will be able to compete with the largest organizations by smart use of AI tools to compete. Being a small publisher meant being inefficient. You had to build an infrastructure – people included – that couldn’t be broadly applied and therefore was inefficient. Just as the shift to cloud computing made startups easier, this shift should benefit upstarts at the expense of incumbents, whose advantages in heft will increasingly be liabilities.

What new models will emerge?

I was once told that in a crisis the leaderboard can change. Volatility is generally opportunity. AI is a great accelerator and will put further pressure on already stressed publisher cost bases.

Derek Thompson had a fascinating podcast recently with Dror Poleg, mostly on the future of cities. Poleg sees the changes happening with work driving a scenario in which a few big cities emerge as outsize winners – NYC will be one, sorry Jim Jordan – while many midsize cities struggle, and new winners emerge in Sunbelt areas now that companies have recalculated the cost-benefit analysis of only hiring people in a certain geographical location. Poleg sees AI having a similar effect, as “over the next 20 years, we’re going to go through the craziest time that humans have gone through.”

Poleg compares it to the music industry, where anyone can break out, the middle gets hollowed and nobody can feel very stable. “What technology is doing now is it’s making all occupations scalable,” Poleg said. “Some people are going to make more money than ever and service more customers than ever, produce more than ever before, but everyone will be competing with everyone else, which means to cruise through the middle will not be an option.”

Will brands matter?

The future Poleg sketches sounds in its outlines what was promised by Web3 and the creator economy, a vast leveling of the playing field and even the emergence of media companies as record labels.

You can make an argument that AI will further the demise of publishing brands. The atomization of content and loss of distribution and pricing power had already eroded the value of once-storied brands. And yet AI is going to lead to an explosion of synthetic content and fakery. Brands should become beacons of trust and reliability, and operate at a different end of the spectrum to soulless AI.

The nature of those brands will likely shift more in the human direction. The All-In podcast excitedly spoke of a future in which people imagine their own movies. I do not believe people want that. AI will lead to all kinds of bad personalization where it isn’t needed or wanted. Personalization has already become yet another optimization exercise that ignores the complexities and contradictions of humanity. The best of media is a connection on a human level.

Those without viewpoints – and many publishing brands I see fall into this category – will suffer mostly because they’re not unique and chose a safer middle path. AI will change a lot, but the middle always gets crunched.

The frictionless folly

In this week’s episode of People vs Algorithms, Troy, Alex and I discuss the role of friction in media, and how the onslaught of AI will just accelerate the quest of tech to eliminate all friction wherever it appears. I find this a fool's errand mostly. As Kev put it to me on Notes: “Most important things in life demand friction. No long-term satisfaction can happen following a ‘frictionless’ process, no true love or friendship can take place through a ‘frictionless’ environment.”

The publisher’s revenue playbook

Digital publishers have typically focused on one main source of revenue and it is time to consider diversifying!. The key is to make the right choices in your pricing and monetization strategies in order to see a big impact on revenue potential. Uncover eight strategies to boost revenue today from the experts at Zephr, a Zuora company.


The scale era of digital publishing has a book. Ben Smith’s “Traffic” frames the battle of the early 2010s around BuzzFeed and Gawker as publishers chased “eyeballs” in the mistaken belief that they would close the gap between time spent and budget spent that Mary Meeker foisted on everyone for a decade. Didn’t work out great. I hope to have Ben on The Rebooting Show soon to discuss the book (and get a Semafor update).

Twitter reminds me of Sebastian Sagmeister’s “Things I’ve Learned in My Life So Far,” where he says about drugs, which “feel great in the beginning but become a drag.” I don’t believe Twitter will die, but its influence and hold will continue to decline as people weary of the endless fighting and narcissism.

For my money, the original sin of internet publishing was the separation of the audience data from the media impression. Simon Owens sees “open programmatic” as the villain. One quibble is the shift to programmatic was driven by the buy side, so publishers didn’t have much leverage to dictate how its customers wanted to transact.

ICYMI: On The Rebooting Show, I spoke to Bullish Studio’s Brian Hanly, who walked me through how to turn to turn meme accounts into media businesses.

Human creativity will win in the long run. Here are the fonts Succession characters embody. Tom is very much Futura Bold.

Thanks to LiveIntent and Zephr for sponsoring this edition of The Rebooting. I encourage you to check them both out for how they help publishers build resilient business models. Get in touch if you want to discuss sponsorship options to reach over 15,300 publishing professionals. My email is