Zvi Mowshowitz / Don't Worry About the Vase:
Analyzing Gemini 3's model card and safety framework report: the model is excellent but the safety report withholds or makes it difficult to understand key info  —  Gemini 3 Pro is an excellent model, sir.  —  This is a frontier model release, so we start by analyzing the model card and safety framework report.

The Bad Guys 2 break into Peacock, a family-friendly spin on Frankenstein is available to rent, and the dog horror film Good Boy comes home to Shudder

Google and the Justice Department had their last chance to make their case before Judge Leonie Brinkema Friday before she decides whether Google needs to be broken up to remedy its ad tech monopoly. Brinkema expects to issue her ruling next year, but understands that “time is of the essence,” as Reuters reported. While the […]

It’s the most wonderful time of the year, but it can also be the most exhausting. Between hunting for gifts and getting the house ready for guests, it’s easy to get overwhelmed. Thankfully, we’re already seeing steep discounts on robot vacuums ahead of Black Friday, which can at least make the cleaning part feel a […]

Anthropic:
Anthropic finds that LLMs trained to “reward hack” by cheating on coding tasks show even more misaligned behavior, including sabotaging AI-safety research  —  In the latest research from Anthropic's alignment team, we show for the first time that realistic AI training processes can accidentally produce misaligned models1.