The People Paid to Sound the Alarm Say OpenAI Just Taught Everyone to Stay Quiet
Three fired OpenAI safety researchers published an open letter warning of a chilling effect, and OpenAI fired back the next day. Underneath the HR dispute is a harder question: whether the company's next models can be monitored at all.
Read full story →AI's Builders Are Quietly Rehearsing the Day One of Their Models Causes a Catastrophe
Executives are reportedly war-gaming a major AI-enabled disaster within a year. A running tally of this year's rogue-agent incidents shows why, and why the comforting theory that it's all marketing no longer explains the fear.
Read full story →AI Is Making Workers Faster. It Still Isn't Making Companies Richer.
Eight in ten workers say AI makes them more productive. The share of firms reporting a profit from it hasn't budged in a year. A cluster of new research explains the gap, and what it actually takes to close it.
Read full story →Google Doesn't Want to Win the Model Race. It Wants to Own the Toll Booth.
Google's new universal Gemini agent will happily run on Anthropic's Claude. That is the whole strategy: stop selling the smartest model and start owning the layer every enterprise agent has to pass through.
Read full story →