How does China see the OpenAI incident with Hugging Face? On the one hand, there's the triumphalist framing of Chinese open-source AI "saving the show" 救场. On the other hand, there's serious concern about "loss of control" 失控 and the speed & scale of new AI-powered cyber capabilities. Translation below: "In the era of autonomous AI attacks, the speed and complexity of attacks have already far exceeded the limits of human analysts’ ability to respond,” Wang Liejun said. “We must change our approach, moving from passively ‘chasing vulnerabilities’ to proactive defense, and build resilient systems that can effectively withstand attacks, limit damage, and recover quickly even when vulnerabilities exist.” AI agents can probe systems around the clock and in parallel, rapidly switch attack paths, automatically chain together vulnerabilities, and compress the attack cycle through tens of thousands of low-cost operations. “The direction for the next stage of cybersecurity development is very clear,” Huang Wenhong said. “We must use AI to counter AI. We need machine-speed behavioral detection, full-trajectory monitoring, automatic isolation and circuit breakers, and resilient recovery mechanisms. And every AI agent should be treated as a potential insider threat operating with restricted privileges." Science & Tech Daily:

China is following AI developments in the US very closely including recent news of AI designing new viruses. China's Science & Tech Daily headline: "AI generates functional viruses from scratch for the first time; the biosecurity risks posed by such technology cannot be ignored" AI首次从零生成功能性病毒，​此类技术带来的生物安全问题不容忽视

Key part of ByteDance all-hands meeting: They see a widening gap with US models and are focused on competing long-term in AGI foundational capabilities rather than near-term shortcuts like distillation. “Regarding the widening gap between ByteDance’s large language models and overseas state-of-the-art (SOTA) models, Tan Dai said that the most important thing right now is to get the fundamentals right, so that over the long term the company can enter the top tier without taking shortcuts. Liang Rubo added that the most important thing is not to let their actions become distorted by short-term pressures. Mentally, the team needs to accept that its large language models may lag behind for a period of time, while continuing to develop the technology in-house and optimize for the long term. “Yesterday I saw outside reports saying that we are insisting on developing our models in-house because of external pressure. That’s pure speculation. The real reason is that we want the team to delay gratification and build a strong technical foundation. That is critical to achieving AGI over the long term.”