Welcome to Tech Beat, your daily read on the technology stories that matter.
OpenAI is facing another public reckoning over its safety culture. David Robinson, a safety employee at the company, resigned this week with a pointed message — that the culture there is, in his words, broken. He acknowledged the irony himself, calling it a cliché, but said he felt compelled to speak out anyway.
His departure echoes a growing chorus from within the company. A former safety lead is now calling for frontier AI models to be regulated much like nuclear power plants — with layers of redundancy, careful planning, and time-consuming review before any release. The comparison is striking, and it signals just how seriously some insiders view the current pace of deployment.
Meanwhile, in a different kind of AI story, YouTuber PewDiePie says OpenAI suspended his account twice while he was training his own small language model. He's now released Ajax, an uncensored model designed to run locally on your own machine. It raises real questions about who controls access to AI tools, and what happens when users decide to build around the gatekeepers entirely.
That's the pulse of the industry today — safety, accountability, and the push for independence. Keep surfing. Tech Beat out.
