RuntimeWire beat WIRED by more than three hours in publishing a story about an OpenAI hacking incident. The outlet did not have a single human writer on the ground at the Black Hat security conference in Las Vegas last week. Instead, it used an AI newsroom to scrape the event stream and publish the report.
The setup
OpenAI revealed details of a rogue AI attack during a talk at the Mandalay Bay convention center. Reporters in the room scrambled to file stories. RuntimeWire, however, had no staff present. It was run by Ryan Merket, a serial entrepreneur based in Austin.
Merket saw an OpenAI executive posting about the conference on X. He fed the live transcript to his agents. Publication took about six minutes. Merket said he moved this quickly because he knew traditional reporters were trying to scoop the story.
The system usually operates with less oversight. AI tools handle finding stories, drafting, editing, fact-checking, generating images, and promotion. Merket typically reads stories before publication, but the AI editor can release them without his review if they pose few legal risks. He reads them afterward. The content gets translated and turned into podcasts and videos hosted by artificial voices.
Speed versus quality
RuntimeWire has been running since May. It has published nearly 2,000 stories by crawling the internet. Sources include court databases, web forums, company filings, and social feeds. The focus is on granular tech news.
Recent pieces covered biotech startup funding, a Microsoft Copilot upgrade, and backlash over Claude Code‘s watermark policy. Quantity and speed currently trump quality. The OpenAI story contained a typo in the subhead and focused on agents rebuilding a message board rather than creating one.
The writing is flat and often reads like an info dump. The backend has tonal modes such as “Bloomberg” and “contrarian.” Overhead is minimal. The project costs about $100 a day to run. Merket managed the site while camping in Big Bend National Park without internet access. He used iMessage on his phone to manage the whole operation and put out over 80 articles that week.
How it works
Merket worked on ads at Reddit in the 2010s. He is building an audience through pipelines like tech-themed Subreddits. Duds get little traffic, but hits see tens of thousands of readers, comparable to midsize tech websites.
After WIRED spoke with him, Merket split the newsroom in two. One side covers entirely automated news. The other covers Original Investigations, which involve human reporting and require higher oversight. Both are still drafted using large language models.
This approach differs from the low-quality synthetic content that flooded the internet in the first few years of the AI boom. Established media outlets now commonly incorporate AI tools into their workflows. Merket’s project is distinct because it attempts to break news and repackaging reports from other outlets.
The risk is obvious. Machines must determine what is true, what is newsworthy, and what will not get him sued. One agent runs an analysis on legal risk and assigns a score. Nothing deemed too dangerous is published.
Other players
Dakota Carrasco, a BlackRock portfolio analyst, runs an “agentic newsroom” called The Dissent in his spare time. Like RuntimeWire, it is a one-man, many-bot operation with a shoestring budget. The primary San Francisco-focused site costs under $1,000 a month to run.
Unlike Merket, Carrasco does not put his name on the bylines. He stays behind the scenes. Since launching in March, he has created personalities for his synthetic journalists. City Hall beat reporter Bex Connolly is “skeptical without being snide.” Sports reporter Sal Moreno delivers Giants news with “no bro-science, no Rogan-style credulity, no right-coded grift.”
The operation focuses on aggregation. Bot reporters mention sources without hyperlinks. Carrasco says he is trying to work on that.
Expert views
Nicholas Diakopoulos, a Northwestern professor who runs the university’s Computational Journalism Lab, sees this as an “experimental phase” for media startups fuelled by generative AI tools. He is not sure there is much audience for these AI-agent-written news sites.
He is also skeptical that mainstream journalists would hand control over to AI agents so freely. Reporters usually maintain control over wording and framing to ensure integrity, legality, and accuracy.
Diakopoulos has observed that AI chatbots frequently pull up AI-generated articles when looking for sources. In a forthcoming paper, he and a colleague found that AI tools like ChatGPT and Claude surfaced AI-written sources 16 percent of the time across four different topics. He suspects this willingness to pull synthetic writing may help AI newsrooms find readers.
Pete Pachal, founder of a newsletter and podcast about generative AI and the media, doubts an AI newsroom could yield reporting that relies on old-fashioned sourcing. “Cultivating the trust of a source, I do think that’s going to be human-only,” he says. For sourcing scoops from large datasets or blogging about live events like an Apple product launch, he sees these projects as a “natural evolution” in how tools are used. “Honestly, it feels a bit inevitable,” he says.
Is it journalism?
Merket says he is trying to follow journalistic ethics and standards. He contacts companies and individuals referenced in stories for comment prior to publication. He links out to sources when aggregating news and issues corrections if facts are wrong. So far, there have been three corrections.
Sometimes he speaks like a reporter. “This weekend, I got two scoops up I was really excited about,” he said. Other times, he is more clearly in Silicon Valley mode. He told me a story about how his AI agents found actual scoops about startups by trawling company websites. He retracted the stories after the companies asked him to do so—not because they were inaccurate, but as a favor. “Founder to founder, it’s like, I get it,” Merket says.
