Chinese lab Alibaba Qwen announced Alibaba Qwen 3.8 Max today, a model that matched and beat GPT 5.6 Sol and Fable 5 on many benchmarks at a much lower price point with open weights. Its launch video showed humans relaxing while AI agents completed work tasks, including designing chips for 12 hours while an employee fished, verifying protein sources while another played tennis, and autonomously reconciling finances while another went rock climbing. The upbeat video featured nature sounds. This contrasts with Anthropic’s video a month ago, titled “There’s hope in hard questions,” which used a focus group’s voices about AI apprehensions, jarring music, and images of fires, graveyards, mass surveillance, social upheaval, and young children in hospitals. That video was heavily panned on social media. The two approaches reflect broad societal attitudes: young people in China are more optimistic about AI, while young people in the US are among the most skeptical. The videos may have a self-fulfilling aspect, cementing public views. In the US, opposition to building datacenters could be fueled by negative AI information, making it harder for US labs to compete with Chinese counterparts as the AI race tightens.