AI News, Sep 8, 2026: OpenAI claims major math breakthrough amid academic controversy

Today's updates highlight major leaps in scientific AI alongside high-stakes corporate competition, new consumer agents, and security challenges.

1. 01
   Hackers are stealing Claude subscription tokens from usersTechCrunch

   Anthropic warned subscribers to secure their accounts after users reported hackers stealing Claude API tokens to run unauthorized queries.

   Why it matters

   If you subscribe to Claude, your account could be silently drained of usage limits, costing you money or locking you out of the service.
2. 02
   Meta launches Muse, an autonomous personal AI assistantMeta AI

   Meta announced Muse, a highly integrated personal AI agent designed to proactively help users accomplish daily tasks and manage personal schedules.

   Why it matters

   The agent can perform complex tasks like booking flights or selling items by directly accessing your email, calendar, and payment systems.
3. 03
   OpenAI claims major math breakthrough amid academic controversyWIRED AI

   OpenAI announced an AI-generated solution to the Navier-Stokes Millennium Prize problem, but faced accusations of pressuring a researcher to exclude a rival co-author.

   Why it matters

   Solving this 90-year-old physics problem could improve fluid dynamics simulations, but academic disputes raise questions about AI research ethics.
4. 04
   Meta stops grading engineers on how much they use AIThe Decoder

   Meta removed AI tool usage metrics from software engineer performance reviews after employees began artificially inflating their activity, a practice dubbed "tokenmaxxing."

   Why it matters

   This reversal highlights the limits of forcing AI adoption, showing that measuring tool usage can lead to low-quality, automated work.
5. 05
   Google DeepMind maps 9 billion human DNA variantsGoogle DeepMind

   Google DeepMind released AlphaGenome Atlas, an AI-powered map detailing the molecular effects of nearly every possible single-letter DNA change in the human genome.

   Why it matters

   This massive predictive map of genetic changes could help scientists understand diseases better and accelerate the development of personalized medicine.
6. 06
   Mistral AI raises €3 billion to expand sovereign AI modelsMistral AI

   French AI startup Mistral raised €3 billion in a Series D funding round, valuing the company at more than €21 billion.

   Why it matters

   This massive funding round secures Europe's leading AI startup as a major independent competitor to US giants like OpenAI and Google.
7. 07
   OpenAI releases ChatGPT Images 2.5 with new drawing toolOpenAI

   OpenAI announced ChatGPT Images 2.5, introducing a "Sketch" feature that lets users turn hand-drawn doodles into polished AI-generated images.

   Why it matters

   Users can now guide AI image generation by drawing rough sketches directly in the chat, making visual editing more precise and intuitive.