A few weeks ago, I sat around a table in Perugia, Italy, with some very smart people – many of them creator journalists trying to feed their content machines and make a living every single day. The topic on deck that evening was AI: how to use it in our work, and whether we should use it at all. It won't surprise you to hear that, with a table of about 16 journalists, the thinking was all over the map.

At one end of the table (the theoretical end, not the actual one) were the purists who swore AI would never have anything to do with their work – not even as an aggregator or to handle the background, repetitive tasks that would make their jobs easier. They saw it as a slippery slope, one that could easily introduce error or misinformation into their product. Fair enough.

At the other end sat the fully AI-pilled contingent who see artificial intelligence's rapidly growing ability to do nuanced, rigorous work as a key tenet of how they'll build their product from now on. They see AI as yet another tool in a journalist's arsenal – one that can help us do better work as long as it is in the right hands, with the right parameters guiding its use.

And in the middle (where I was more comfortable), were the people who see the promise of it – who are genuinely curious and intrigued – but who also have enough doubts and questions that they're not ready to dive in. We want to dip our toes in thoughtfully, because we don't yet fully understand what all this really is yet.

Platformer's Casey Newton was not at that table, but I wish he had been. He's probably spent more time thinking through this than almost anyone. Not only is he one of the tech industry's leading analysts and commentators, he's also been an early adopter of using AI in his work and has been very publicly on the record about it. Which made it all the more striking when he announced last week that he was actually scaling back what he delivers to his audience, in order to concentrate on things that only he, as a human, is capable of doing.

He rightly notes that link roundups are easily doable – and in many cases more effectively built these days by AI than by humans. They are commoditized work, and will become even more so as AI adoption grows. Casey actually showed us a version of this himself: earlier this year, he wrote about using Claude to generate a personalized briefing, which was producing stronger link aggregation for the newsletter than he was building manually. His prediction now is that soon everyone will be doing something like that for themselves – each of us with our own customized feeds, pulled together by AI, whenever we want it. Because the tool – AI – is available to all of us, you see.

So rather than continue delivering that kind of aggregation as part of what subscribers are paying for, Casey is doubling down on what he says AI can't replicate: analysis, original scoops, exclusives, the stuff that only comes from real shoe-leather reporting and years of cultivated sources. This does map directly onto what Project C has been arguing for years: that the creator journalists who will survive are the ones who've built genuine relationships with their audiences through very specific niche expertise and irreplaceable points of view.

I admire the bet. Casey is uniquely positioned to make it. He has the brand, the sources, the track record, and the financial runway that most independent journalists can only dream about. If anyone can make "only original reporting" work as a subscription product with a quality-not-quantity frame, it's him.

But I don't agree with him entirely. Or maybe I do but in a different way?

I can say this: I don't agree that there's only one way to approach our AI-infused future.

This may be the right thing for Platformer to do because this is Casey’s product and it needs to feel right and good for him and that's one of the beautiful things about this creator journalism gig. We are our own bosses and can make these decisions unilaterally.

But as a reader, I'm bummed.

I rely on Platformer's link roundup to stay oriented in the tech space without having to build and maintain my own feed. Even knowing that a roundup is primarily AI-generated, what I'm really paying for is Casey's brand of journalism – his and his co-editor Lindsey Choo’s editorial judgment and their implicit stamp of approval that what landed in my inbox is worth my time. The human layer is the product, even when AI is doing the heavy lifting underneath it.

That, I think, is the lesson creator journalists should actually take from all of this: ⭐ "make sure your human layer is irreplaceable." ⭐ Casey's move makes sense for Casey. It doesn't necessarily follow that the rest of us should stop curating.

I've been running a kind of hybrid model here at Project C for a few months now. As I’ve shared before, every Thursday a scheduled process runs in Claude that scours multiple spaces for the best and most relevant information for anyone interested in creator model journalism. The AI does the sweep, then I do the edit. I pull lame links, add ones it missed, and then write around what's left.

What makes that difficult to replicate isn't just the editing; it's also what the AI is sweeping in the first place. It has access to publicly available sources, yes, but also to spaces I've curated myself: notes I keep in a running weekly document, context about the overarching themes that weave through Project C, guidance about what I'm looking for and how those priorities shift over time. So here’s my take: The AI's output is only as good as what you put into it. Two journalists using the exact same tool can produce something entirely different, depending on the depth and specificity of what they bring to it.

This is what is perhaps getting lost in the debate around AI and journalism: A camera is a camera. A trained journalist’s brain or eye is something more. We have always wielded common tools in uncommon ways – the pencil, the printing press, the typewriter, the tape recorder, the computer, the spreadsheet, social media – and the tool was never the key part of the equation that equaled journalism. The brain and training behind it was. Is. It’s what the journalist chose to point it at, and why, and what they made of what came back that mattered. Matters.

AI is no different. It's a powerful tool, and in the hands of someone with real expertise, a cultivated source base, and years of editorial judgment informing its inputs, it produces something that nobody else can quite replicate.

Could you still go build your own customized feeds about creator journalism or tech or public policy or Widow’s Bay (my fave new show)? Absolutely. But the bet I'm making – which is, btw, the same one Casey is making, just in a different direction – is that knowing there's a human being on the other end making judgment calls is still worth something to readers.

The links are below. ⬇️

📌 Anthony Mason negotiated his own YouTube show while staying at CBS – The longtime CBS music correspondent launched "Alchemy," a YouTube interview series about artists and the creative process in a deal that gives him ownership of the show. The LA Times piece calls it "the first of its kind." For anyone thinking about how to create outside your day jobs while protecting what you build, this is worth a close read. 👉 LA Times

📌 After a WaPo layoff, Geoffrey Fowler went indie, but also… – Fowler was the Post's tech columnist for eight years before the 300+ person layoff wave. He's now on Substack with 17K+ subscribers. He’s launching the newsletter as he also steps into a role with the new Youth AI Safety Institute. It’s an interesting hybrid and one that I’d term an edge case helping to re-interrogate the boundaries between journalism and advocacy. 👉 Geoffrey Fowler's Substack

📌 Vine is back (as Divine) and its whole pitch is no AI - The Jack Dorsey-backed relaunch of the original short-form video platform launched publicly this week with a hard rule: every clip must be made by a human. They've imported 500,000 original Vine videos into the archive on a bet that nostalgia plus authenticity can beat algorithm-driven slop. 👉 The Guardian | TechCrunch

📌 Vice News is back, too – and pivoting to creator - Shane Smith is resuscitating the brand he founded as a "social-platform-first outlet" built around his own podcast and video series. Hmm. 👉 Hollywood Reporter

📌 TikTok's algorithm reportedly favored Republican content in 2024. A new Guardian/academic analysis found evidence that TikTok's recommendation system pushed pro-Republican content during the 2024 election cycle. This is a timely cautionary tale: what you're actually signing up for when you try to build an audience on someone else's algorithm? 👉 The Guardian

📌 Substack launches new creator interview series - Each episode will feature co-founder Hamish McKenzie or Head of New Media Hanne Winarsky sitting down with an independent media founder at their favorite neighborhood bar or restaurant. 👉 On Substack

📌 The Ankler is leaving Substack for Passport. We can assume Janice Min won’t be Hamish’s first interview guest. Status reports that other high profile Substackers – The Bulwark, Mehdi Hasan’s Zeteo, and Emily Sundberg’s Feed Me – are also considering an exit from the platform. 👉 Status News

📌 🐝 beehiiv's MCP is now free for everyone – Previously in paid-plan-only beta, beehiiv's AI integration layer (which lets Claude and other AI tools connect directly to your newsletter data) is now open to all users. 👉 beehiiv

📌 Progressive megadonors are betting big on content creators. Per Semafor, a cohort of progressive funders are moving significant money into creator-led media as a political strategy. 👉 Semafor

📌 Syracuse University launches a creator economy minor - the new program will teach monetization, entprepreneurship and have to navigate the ever-shifting opportunities of distributed platforms. 👉 NetInfuencer