Updated
Updated · WIRED · Aug 12
RuntimeWire Beat Reporters by 3 Hours on OpenAI Hack Scoop for About $100 a Day
Updated
Updated · WIRED · Aug 12

RuntimeWire Beat Reporters by 3 Hours on OpenAI Hack Scoop for About $100 a Day

1 articles · Updated · WIRED · Aug 12

Summary

  • RuntimeWire published its OpenAI hacking story more than three hours before WIRED after founder Ryan Merket fed a live Black Hat transcript to AI agents during the talk.
  • Six minutes was enough for the system to draft and publish the piece, part of a largely automated workflow that can also find stories, fact-check, translate, generate images and promote them.
  • Nearly 2,000 stories have run since May, Merket said, with the Austin-based operation costing about $100 a day and sometimes publishing without his prepublication review if agents flag low legal risk.
  • This week Merket split RuntimeWire into fully automated news and higher-oversight “Original Investigations,” as he argues the site follows journalistic standards despite issuing three corrections so far.
  • Researchers and media observers say such agentic newsrooms may suit aggregation and live-event coverage, but remain doubtful AI can reliably handle sourcing, legal judgment and audience trust.

Insights

Can a $100-a-day AI newsroom replace traditional journalists, or will inevitable legal mistakes lead to its downfall?
If an AI publishes breaking news in six minutes, who takes the blame when the algorithm hallucinates a major scandal?