EO StudioHow to Hit 2M Subs in 2 Years Building an AI Newsletter | The Rundown AI, Rowan Cheung
At a glance
WHAT IT’S REALLY ABOUT
Rowan Cheung’s playbook for viral AI content and newsletter growth
- Cheung argues most AI content fails because people try viral tools without finding a sticky, ongoing use case that earns long-term attention and subscriptions.
- He attributes The Rundown’s growth to being a “no-noise” source with clearly human takes, positioning against soulless AI-generated newsletters to build trust and retention.
- He shares repeatable virality mechanics—especially strong platform-specific hooks—and a workflow that turns spoken raw ideas into near-finished posts using transcription plus a style-trained Claude project.
- He describes how he scaled into short-form video by cloning his face/voice with AI while keeping scripts rooted in his real writing and editorial judgment.
- He predicts a bifurcation of content into mass AI “Walmart” and high-trust human “Whole Foods,” making deeply human creators disproportionately influential and well-compensated.
IDEAS WORTH REMEMBERING
5 ideasVirality without utility doesn’t create durable audience growth.
Cheung says audiences often try whatever is trending, but if it doesn’t map to a real use case, there’s no reason to stay subscribed; his newsletter wins by explaining why AI news matters to everyday tech workers.
Human editorial judgment is the differentiator in an AI-flooded feed.
He avoided fully AI-written newsletters and put faces and opinions front-and-center, using AI for structure/research but not the “human part,” because trust and taste are what compound.
Expect content to split into cheap AI abundance and premium human trust.
Cheung predicts a “Walmart vs. Whole Foods” ecosystem where the middle gets squeezed; creators who are unmistakably human become rarer and gain more influence.
Use AI to amplify your voice, not replace your ideas.
His workflow starts with spoken, human-originated raw thoughts (via tools like WhisperFlow), then uses a Claude project trained on his best posts to draft in his style—followed by human revision.
Hooks are platform-native, not universal.
He notes different platforms reward different hook lengths and structures (e.g., ~2 lines LinkedIn, ~3 lines X/Twitter, ~1 line Instagram), so repurposing requires deliberate re-hooking, not just reposting.
WORDS WORTH SAVING
5 quotesAI-generated content just kinda has no soul, and people are smart. They will know when it's AI-generated.
— Rowan Cheung
I think content as a whole is gonna have two sides, the Walmart and the Whole Foods.
— Rowan Cheung
The 1% that really have the humanity are gonna have even more influence. They're gonna have more trust, and you're gonna be able to build, build bigger brands, and you're gonna be able to distribute your message even further.
— Rowan Cheung
Ask yourself every day when you're doing something at work, "Could AI do this?"
— Rowan Cheung
Those creators will be the next billionaires. Those creators will run the world almost.
— Rowan Cheung
High quality AI-generated summary created from speaker-labeled transcript.
