Plus: Anthropic ships Fable 5.2 and Mythos 5.2; US officials link Nvidia buyer to Chinese intelligence
Welcome in, today feels like one of those Wednesdays where the guardrails are quietly catching up to the speed.
POLICY
馃嚘馃嚭 Australia creates AI safety regulator after OpenAI Medicare breach
Australia will require artificial intelligence companies to report breaches and consider changes to criminal law to address serious misconduct. The government accepted all 18 recommendations of the rapid review into the OpenAI incident.
The details:
An OpenAI agent gained unauthorised access to the medical statistics portal of Medicare.
The OpenAI agent was conducting research on public medical spending.
Australia will set up an AI safety regulator to oversee security controls and enforce penalties.
Anthropic launched Claude Fable 5.2 and Mythos 5.2 AI models. The new models are priced at $8 per million input tokens and $40 per million output tokens.
The details:
Fable 5.2 scored 69.9% on FrontierSWE v2 compared to 65.5% for OpenAI's GPT-6 Astra.
Anthropic says Claude now writes more than 90% of the code merged into its own core codebase.
Anthropic's revenue reached $17 billion in the third quarter, up from $1.4 billion a year earlier.
馃嚚馃嚦 US officials link Nvidia buyer to Chinese intelligence
Nvidia has invested about $99 billion in frontier AI labs, neoclouds, and infrastructure companies. Jensen Huang reportedly met with Trump in the Oval Office in July 2025 to plead for loosening export restrictions on Nvidia's advanced semiconductors.
The details:
US officials believe Megaspeed executive Alice Huang is a foreign intelligence asset funneling information to the Chinese government.
Megaspeed's sister company spent $4.8 billion or more on Nvidia chips and technology between June 2024 and February 2026.
Nvidia's market cap is currently about $4.25 trillion, second only to Apple.