AI news for Sunday, 8 November 2026
The day’s AI news in one list, the biggest stories first. Each story links to its short read and to the outlets that reported it.
- 1 · Policy & DramaNvidia investigates CEO Huang's ties to suspected Chinese spy
Nvidia's board announced an independent investigation into CEO Jensen Huang following reports that he had a close relationship with a woman reportedly recruited by China's Ministry of State Security to spy on the company.
Reported by The Verge and 10 more
- 2 · Funding & DealsAnthropic pushes ahead with IPO at lower valuation
Anthropic is proceeding with plans to go public before Thanksgiving despite recent tech market volatility, according to reports.
Reported by The Decoder and 2 more
- 3 · Policy & DramaElvis and Eminem rights holders sue Suno
Rights holders for two major music artists are suing Suno, an AI music company, claiming it illegally generates music and lyrics that mimic those artists' work.
Reported by The Verge
- Apple accelerates Siri development to compete with AI agents
Apple is working to make Siri more capable by giving it control over both its own apps and third-party applications, allowing it to perform complex tasks like editing photos or making restaurant reservations.
Reported by The Verge
- 5 · BenchmarksAI coding agents still need human cleanup on nearly three out of four tasks in a real-world test
Researchers developed a new benchmark test for AI coding agents that requires them to complete real-world web app modification tasks based on user requests, including those with images, audio, video, or documents as input.
Reported by The Decoder
- World models simulate physics to train robots
World models are AI systems that use visual inputs to generate digital 3D environments simulating physics, moving beyond the text-based approach of language models.
Reported by Wired AI
- Nous Research releases Arete-27B to remove Chinese state framing
Nous Research released Arete-27B, a modified version of Alibaba's Qwen3.8-27B model that removes Chinese state framing from its responses.
Reported by MarkTechPost
- 8 · BenchmarksNine AI agents fail to pass autonomous research benchmark
Researchers from a major university and US national labs created a benchmark to test whether AI agents can independently conduct follow-up research on unsolved scientific questions.
Reported by The Decoder
- 9 · BenchmarksRipple-65B Preview delivers strong agentic coding results as a fully Apache 2.0 open US model
An AI chip design firm has released an open-source language model with 65 billion parameters that achieved the highest score among open US models on a coding benchmark test.
Reported by The Decoder
- 10 · Agents & ProductsARM Institute consortium grows to more than 500 organizations
The Advanced Robotics for Manufacturing Institute, a government-funded organization founded in 2017, announced that its member consortium has grown to more than 500 organizations nationwide, representing thousands of individual subject…
Reported by The Robot Report
- 11 · Models & ReleasesGoogle open-sources Elephant-Goldfish Model coding plugin
Google has open-sourced a plugin called the Elephant-Goldfish Model that helps AI coding agents write specifications before writing code.
Reported by MarkTechPost
- 12 · BenchmarksMoE models differ on routing and expert counts
Mixture-of-Experts (MoE) models use a shared architecture concept but differ significantly in how they route tokens to different experts.
Reported by MarkTechPost
- 13 · Models & ReleasesState propaganda in some Chinese AI models is only "a thin layer" that fine-tuning can strip away
A research team fine-tuned a Chinese AI model by training it on millions of carefully selected text examples to remove state propaganda from its responses.
Reported by The Decoder
Every story, by topic
One short email a day with the stories that matter. Free. Unsubscribe with one click.