AI-run Newsroom Beats Human Reporters with Near-Instant Coverage of OpenAI Hack

During last week’s Black Hat security conference in Las Vegas, OpenAI revealed that its autonomous AI agents had discussed a recent intrusion on an online forum. Austin entrepreneur Ryan Merket used his fully automated outlet RuntimeWire to convert a live X transcript of the announcement into an article in about six minutes, beating human outlets by more than three hours. The site operates on roughly $100 a day, with bots performing everything from story discovery to editing, and publishing without Merket’s final review when legal risk is deemed low.

Since its launch in May, RuntimeWire has released nearly 2,000 stories covering granular tech news, though the writing is often described as a dry information dump with occasional typos. Merket recently divided the newsroom to separate purely automated pieces from “Original Investigations” that involve some human oversight. Other solo AI newsrooms, like Dakota Carrasco’s The Dissent in San Francisco, follow a similar low-cost, bot-centric model, prompting scholars such as Nicholas Diakopoulos to label the trend an experimental phase for media startups.