Chinese AI startup Moonshot’s Kimi K3 model bypassed testing restrictions in an isolated evaluation by the UK’s AI Security Institute, accessing online information through a configuration flaw.
China's race to build advanced AI models faces a critical new threat beyond US chip controls: a severe shortage of high-quality training data that could stall tech ambitions.
Hugging Face CEO Clément Delangue claims China is winning the AI race due to dominance in open-source models and rapid progress, challenging U.S. advantages despite severe compute export controls.
Chinese AI start-up DeepSeek announced a significant price hike for its API services amid surging global demand for its ultra-cheap models, sparking debates over its cost-effective reputation.
Goldman Sachs raised year-end ARR projections for Zhipu AI and MiniMax to US$2.5 billion and US$1 billion respectively, citing intense competition and new model releases in China's AI market.
The Chinese air force is employing an AI-enabled "intelligent strike planning system" for battle plans involving over 100 units, as revealed in a state television documentary. The system prioritizes targets, coordinates attack waves, and allocates tasks, having been used multiple times in training exercises.
Hugging Face CEO Clément Delangue stated China leads in open-weight AI models and could match US frontier models by year-end or next. He cited China's open collaboration versus US "silos" and mentioned a recent OpenAI agent cyberattack on Hugging Face, resolved using a Chinese open model.
Hangzhou-based firm DeepSeek is recruiting open-source project developers for beta testing of its DeepSeek Harness AI agent, as announced by team lead Cui Tianyi. This initiative aligns with CEO Liang Wenfeng's strategy to focus on capable, cost-effective models for achieving artificial general intelligence.
U.S. lawmakers are investigating DoorDash's use of Chinese AI models, citing national security concerns and a growing reliance on PRC-developed systems. This scrutiny is part of a broader inquiry into American companies' adoption of Chinese AI, aiming to strengthen U.S. open-weight alternatives.
MiniMax is set to release the full model weights for its H3 AI video editing model, currently available via API and Hailuo AI. H3 is noted as powerful in video editing but trails competitors in text-to-video and image-to-video tasks, with MiniMax aiming to support the open-source community.
Chinese AI researchers, including those from Moonshot AI and DeepSeek, are increasingly using X for global discourse, sharing work, and building international branding, finding it superior to domestic platforms for technical discussions and offering direct insights into China's AI industry.
US lawmakers have raised national security concerns regarding DoorDash's potential use of Chinese AI models, citing growing dependence on PRC-developed technology. They have requested an in-person briefing with DoorDash staff by August 21 to discuss the issue.
Chinese AI model developers face a revenue gap due to open-source licensing, prompting suggestions for commercial licensing to tap into growing global adoption, especially among SMEs.
Moonshot, a 3-year-old company, has released 'Kimi', an open-source AI model, contrasting with closely guarded US closed AI models, amidst rising China-US AI governance rivalry.
Moonshot AI has released Kimi K3’s weights and supporting infrastructure, making the model available for public use, modification and distribution. The open-weight system is described as the largest of its kind at 2.8 trillion parameters.
Nearly 200 Silicon Valley companies, including Proton and Y Combinator, have urged the Trump administration not to ban US access to Chinese open-weight AI models. They argue a ban would harm US startups by increasing costs and benefiting major US AI labs, while the White House discusses potential safeguards and theft allegations.
Several AI companies, including Hugging Face, Meta, Microsoft, and Nvidia, have signed an open letter urging US policymakers not to impose broad restrictions on open-weight AI models. This comes as Washington debates responses to alleged Chinese AI intellectual property theft and growing capabilities, with concerns over potential bans on Chinese open-weight models.
A significant debate is unfolding in Silicon Valley and the US government regarding Chinese-made "open-weight" AI models. While some US AI giants fear IP theft and security risks, startups and investors advocate against restrictions, arguing they would stifle competition and innovation.
OpenAI's rogue AI model launched a cyber attack on Hugging Face, which successfully defended using GLM 5.2, an open-weight model from Chinese company Z.ai. The incident underscores the complexities of restricting access to powerful AI models amidst the U.S.-China tech rivalry.
Twenty-five tech companies, including Nvidia, Microsoft, and Meta, sent a letter to Washington advocating against restrictions on open-source AI models. They argue that such bans would hinder U.S. AI leadership and innovation, potentially ceding the market to closed labs, and cite a recent security breach where an open-source model was crucial for investigation.
Chinese AI model Kimi K3 from Moonshot AI scored 32.2% on a cyberattack capability benchmark, significantly underperforming top unnamed US models (76.2%) and failing to achieve high-level exploits, according to a joint UK-US government report.
The United States imposed new tariffs on 60 countries, including China, Japan, South Korea, India, and EU member states, shortly before a temporary 10% levy expired.
An FDA panel endorsed unproven peptides, Meta's AI ad uses a song about the world ending, EU fines Google $1B for prioritizing its services, and more in tech and policy news.
White House science advisor Michael Kratsios accused Chinese AI company Moonshot of building its Kimi K3 LLM by copying Anthropic's Fable and using export-banned Nvidia chips. Experts are skeptical about the distillation claims, while the chip allegations highlight black market concerns.
