(Re)Structured News | Tow-Knight Center at CUNY
(Re)Structured News argues the "defensible moat" framing is already obsolete — AI is filling those moats faster than anyone can dig them. The smarter question isn't what humans do that machines can't, but what you need to learn to govern systems that may soon outperform you at nearly everything. A useful reframe for anyone building editorial workflows around AI right now.
The Media Copilot
Kris Krüg's piece in The Media Copilot makes a practical case for vibe coding as a journalism tool: describe the interactive experience you want, let the AI build it, ship something readers can actually explore. The argument is that the 3,000-word investigation is only one container for reporting — and often not the most useful one for the data sitting behind it.
Luiza Jarovsky, PhD
Luiza Jarovsky's newsletter makes the structural case that AI hype isn't just marketing — it's the latest iteration of a centuries-old cultural reflex that treats technological progress as inherently good. The argument lands because it names the incentive architecture holding it in place: IP law, capital, and the story modern societies tell about what it means to be advanced.
Shae O. (Substack)
Shae's piece makes a practical case for local AI models — the kind that run on your phone or home server, no data center required. The examples are personal and concrete: a Zotero plugin that reads research papers aloud, an iPhone app that captures stray thoughts without breaking focus. The argument underneath is simple but useful: most frontier-model use cases don't actually need frontier models.