More scoops, less aggregation and analysis: How Casey Newton is revamping his newsletter to compete with AI

Original reporting, news analysis, and a roundup of links. Those have been the three pillars of journalist Casey Newton‘s technology newsletter, Platformer, since it launched in 2017. But, Newton wrote Monday, two of them — link roundups and news analysis — may no longer work so well for his audience in a time of AI automation. So he’s experimenting with changes to Platformer’s offerings, spending more time on original reporting and scoops, less on aggregation and analysis. “We’re betting that the value in tech journalism is moving away from aggregation and predictability,” Newton wrote, “and toward original reporting and surprise.” There’s a lot to think about here for anybody who runs a small publication or sends out a daily newsletter. To be sure, Newton’s case is unique: Platformer is a paid newsletter whose tech-savvy readers are more likely to be using AI than the audiences of more general-interest publications. But the concerns he has now will become relevant to other beats and topics — politics and business, to name just a couple — sooner rather than later. So I asked Newton a few questions via email. Here’s our conversation, lightly edited for clarity and with a bunch of links added for context. By the way, Newton said readers have responded positively to his proposed changes: Monday was Platformer’s largest day for new paid subscriptions this year. Laura Hazard Owen: In your post, you wrote, “The world of link roundups feels much more crowded…but due to a half-decade of layoffs and shuttered publications, there is less and less journalism to make sense of.” Could you talk a little bit about how you’ve seen this play out as you compile the section of links for your newsletter (or, well, used to — as you said in the post, that link roundup is going away because “Techmeme does this particular job better than we can, and does it 24/7.”) Like, are you seeing fewer publications and sources out there? Do you think the broken-ness of X has contributed to the problem? Casey Newton: The main dynamic I’ve noticed here is not that there are fewer sources to draw on, although that’s absolutely true. (It’s depressing to think about how many good tech publications have come and gone just since I started Platformer — BuzzFeed News, Vice, Protocol, OneZero, and most recently, almost the entire tech section of The Washington Post.) The larger issue is that the press corps now feels too small to really swarm a story. When Cambridge Analytica broke in The Guardian and New York Times, the entire press corps got to work identifying their own angles of attack and amplified the story into an international scandal. It’s extremely hard for me to imagine that happening today — Jeff Horwitz has been on an all-time run discovering scandals at Meta over the past two years, and they’ve gotten shockingly little attention. The press corps is too small, the business models have changed (a big part of fast-following other reporters’ scoops was the hunt for Google traffic), and (to your last point) distribution is broken. One of the best things about X was the way that tech reporters would amplify each other’s scoops; that’s gone now and has shown no real signs of re-emerging anywhere else. Anyway, this has basically killed off one of my old jobs, which was that if there were 30 stories about Cambridge Analytica on a Tuesday, I could pick out the most important details across all of them and give you a sense of where things were headed. That felt really useful, for a time. But when it’s just me writing “here’s the news that Jeff Horwitz broke,” it’s much less valuable. Owen: I’m intrigued by what you said about chatbots increasingly being able to provide good-enough news analysis that it could cut into what humans are providing. You wrote, “It doesn’t require much of a leap in imagination on my part to imagine a day where your current lineup of morning and afternoon newsletters is largely replaced by an agent-written briefing that has been exquisitely tuned to your professional concerns — and, unlike this newsletter, instantly respond to your questions about its findings.”1 Is this concern specific to tech journalism, do you think, or does it apply to other areas of journalism, too? Newton: As I said in my piece, I know I’m out on a limb here. Most people would still much rather get their news analysis from a trusted domain expert than from a chatbot. But I am betting this will change as the models improve and (crucially) the products people build around them improve as well. At first, only a particular kind of person will try this — but I think this sort of person will be overrepresented in my readership. And it will expand from there. So if you write a newsletter about, say, national politics, and your stock in trade is explaining what the latest poll numbers mean for Democrats, I absolutely think a bot is going to overtake people in its ability to interpret those numbers someday. I can also see it happening across various business journalism domains, as well as in sports. There are lots of reasons I could be wrong. Chatbots have no moral authority, which makes their writing about tech policy (my beat) feel pretty bloodless and slop-py. Some writers excel at being entertaining (Matt Levine) or useful (Emily Sundberg) or building community (Anne Helen Petersen), and all of these make them less resistant to being replaced by NewsAnalysisBot 5000. But if you’re not the very best in your field, and don’t already have some degree of renown, I think all of this is going to become more difficult. “What kinds of editorial businesses can only be built around a human being” feels like it is going to become a more and more important question. Owen: I agree the news analysis will probably have to be much much better to compete with the chatbots — more scoops of perception, but those are really hard and require a lot of experience! It goes back to what you said about the best-of-the-best writers continuing to stand out while a lot of the middle kind of just fades out. O.K., last thing. In your post, you talk a lot about the importance of scoops to your new business model. More than a decade ago I worked for a tech news site and a big part of what we did was covering company and product announcements, embargoed news, etc. What happens to all that in this environment? I know it’s never been a huge part of what Platformer covers, but it remains a key component of what the remaining big tech news sites cover. How does this type of journalism continue to work and where does it work? Or does it not work anymore? Newton: One thing is that companies will continue to go direct and release news through their own owned-and-operated channel. Look at the way OpenAI now announces everything first in their Discord and on a YouTube livestream; that’s the model. Tech sites today continue to cover it because the stories are fast and easy to write and there’s still some traffic to be chased, but I’m not sure whether that bargain lasts another five years. The big players like OpenAI will be fine, but startups have a real challenge here. I’ve found that there is very little appetite among readers to learn about a new tech company they’ve never heard of. And in a world where Google isn’t feeding traffic to publications for covering them, it can feel like there’s no incentive to pay attention. The flip side is that this creates room for new publications (like Alex Konrad’s Upstarts, which writes the sort of profiles that TechCrunch used to.) - I have to add here: If these briefings are written by AI, that obviously must be disclosed to readers up front.