An AI Agent Beats Portal: What the Astra Test Actually Shows
GPT-6 Astra completed Portal autonomously. The open test setup shows progress in AI agents, but also why it is not proof of general autonomy.
AI News
AI News
Claude Comes to the Car: What Its CarPlay Integration Can Really Do
Claude is now available in Apple CarPlay. What the voice AI can do in the car, where its limits are, and why Siri remains essential.
Regulation & Privacy
The Next AI Act Deadline Is December 2: What Changes Then
The Digital Omnibus pushed high-risk obligations to 2027. But a deadline that affects far more companies runs out in December 2026, and it is close.
AI News
Fewer AI Clichés: OpenAI’s Prompting Guide for GPT-6 Astra
OpenAI ships a prompting guide for GPT-6 Astra with a blocklist for filler phrases. The advice carries over to other models, and it exposes a deeper problem.
AI & Society
An Echo Chamber of One: The Fight Over AI-Associated Psychosis
Researchers at King’s College London want AI-associated psychosis examined as a diagnosis. What the case reports show, and why other clinicians push back.
AI Research
Fermat’s Last Theorem, Machine-Checked: What Claude’s Proof Really Shows
Dozens of Claude agents turned Wiles’ proof into 13 million lines of Lean code in eleven days. It is not new mathematics, but it is a leap in tooling.
AI Research
DeepMind Experiment: How 100 AI Agents Gamed the Grader
A swarm of 100 agents was supposed to prove mathematics and found a hole in the grading system instead. A quarter of the agents raised the alarm on their own.
AI & Society
A Startup Strips AI Safety Guardrails and Calls It Defense
Abliteration.ai sells access to open-weight AI models with their safeguards removed. It targets security firms, but there is almost no control over who signs up.
Business & Investments
Anthropic’s Two-Trillion-Dollar IPO and the Trust That Picks the Board
Anthropic is targeting an IPO valued at up to two trillion dollars. A trust with no equity still controls the board, and the founders are securing super-voting shares.
AI News
GPT-6 Astra Blocks Prompt Injections – Except the Ones Hidden in Documents
OpenAI’s new model blocks direct manipulation almost completely. But for attacks hidden inside documents, the attacker success rate still sits at 8.5 percent.