TopicUpdated Nov 6, 2026Issue #00686 stories today
China AI
Chinese AI labs, models and chips: DeepSeek, Qwen, Kimi, MiniMax, Huawei and the US–China race. 86 stories in today’s issue from 37 outlets, ranked by heat.
Voters showed up in historic fashion for the Democrats on Tuesday night, flipping control of both houses and leaving Trump a lame duck for his final two years. These races were won mostly by campaigning on affordability and corruption. But you’re ChinaTalk…
China’s pure-play artificial intelligence chipmakers crossed an important milestone in the third quarter as several firms swung into profit, but severe supply chain hurdles and soaring memory costs cloud their long-term growth prospects. Domestic graphics…
China’s stocks pared weekly gains and the benchmark 10-year bond hit a record low as the global artificial intelligence (AI) stock bubble burst following the US Democrats’ landslide win in midterm elections. The CSI 300 equity index fell 1.4 per cent,…
Just goofing around with the stock ref2vid workflow, character swap Lora, and some character sheets. Did all of the shots separately and reassembled in my NLE. Almost all of the clips were first attempts. I was floored…
The New York Times reported last week that Mr. Huang had a close relationship with an executive that U.S. officials determined was funneling information to Chinese intelligence.
Yesterday's hot paper was Sparks of Weight-Space Programming: LLMs can natively read & write neural network weights by Rikiya Takehi, Jayden Teoh, Alex Zhang, and Omar Khattab. The paper (OK, the authors classify is as a "blog post", but it's paper-shaped)…
Transcript: - What about the frontier where companies are pushing forward? - We need to limit that. Full-on pause. Full-on pause. Okay. And what happens then? The instant question, what happens if we pause and China…
Turns out MiniMaxH3 can make subjects invisible leaving only floating clothes. It's a useful feature if you want to isolate character clothes before doing outfit swap. Here's the prompt and single-frame edit workflow…
Uni-robot-agent enables open-weight robot-use agents to surpass proprietary models in zero-shot manipulation and suggests joint development of four core capabilities for future training.
Kimi K3.1 and Tencent’s Hy4 are both gutting out cybersecurity, reverse engineering, and vulnerability-patching capabilities from their open-source models. This is becoming a really bad pattern across these latest…
Nvidia CEO Jensen Huang has been asked by three House Republican committee chairs to brief Congress behind closed doors about his reported ties to a woman accused of spying for China. Why it matters: The request escalates the scrutiny of the Nvidia CEO over…
We’ve released Qwen-Image-2.1 Uncensored Omni , the most capable unrestricted Qwen-Image-2.1 model we’ve built so far. The rank-128 adapter was trained on roughly 1 million images , merged into the base transformer at…
Big update to the Delta Inference Engine, you guys asked for more features, more benchmarks and less clickbait. I'm here to deliver The short version: I've added almost everything requested and more. Delta Inference…
The ChinaTalk team is going to CoRL (the Conference on Robot Learning) in Austin! Reply to this email if you’d like to hang out between November 9th and 12th. We’re also doing karaoke on the 11th — RSVP here if you’d like to join! Ukraine and Russia have…
