Anthropic’s Secret Model 2 Beats Every Claude You Can Actually Use
Anthropic runs an unreleased internal model that beats every public Claude, and in the same report raised its own safety risk rating higher than before.
AI Research
Regulation & Privacy
The NEINhorn Lawsuit Against OpenAI: What It Could Really Decide
Carlsen, Marc-Uwe Kling, and Astrid Henn are suing OpenAI. The case concerns model outputs, not a simple ban on artificial intelligence.
Business & Investments
Unitree’s IPO Pop Reveals China’s Circular Robot Economy
Unitree’s stock surged 460 percent on its Shanghai debut. But nearly three-quarters of its revenue comes from a state-backed loop of robots and training data.
AI & Society
One Free Year of Google AI Plus: What Students Actually Get
Google is giving college students a free year of AI Plus. The important questions are eligibility, data, course rules, and what happens afterward.
Business & Investments
After Its $60 Billion Cursor Deal, SpaceX Also Tried to Buy Cognition
SpaceX bought AI coding assistant Cursor for $60 billion and reportedly approached rival Cognition too. What the AI-coding buying spree means for developers.
AI News
Claude Cowork: When the AI Sends Email Without Asking
Anthropic’s Claude Cowork now drafts and sends Gmail messages on its own, and the send-approval step can be switched off entirely.
AI & Society
Gemini Spark: What Google’s Always-On Agent Can Do and Why Germany Is Still Waiting
Google is expanding Gemini Spark into an always-on agent for Workspace. Here is what it can do, why oversight matters, and what availability means for Germany.
AI & Society
ChatGPT for Teens: What the New Safety Mode Can and Cannot Do
OpenAI is rolling out ChatGPT for Teens. Here is what the automatic safeguards cover, what parents can control, and where the limits remain.
AI Research
OpenAI Throttles Its Riskiest Model While Having Just Disbanded Its Safety Team
OpenAI paused parts of training on its Astra model over critical cyberattack capabilities, right after disbanding the team meant to assess such risks.
AI & Society
How People Really Use AI: The Data Gap Nobody Closes
OpenAI and Anthropic release only selective usage data on ChatGPT and Claude. Stanford researchers found real gaps. What that means for policy and business.