Chinese AI builders trade doomsday fears for factory floors and small models
At Shanghai's AI Engineer event, practical deployment on factory floors and lightweight smartphone models took center stage over Silicon Valley's existential dread.
At the recent AI Engineer conference in Shanghai, developers and industrialists approached artificial intelligence with a markedly different tone from their Western counterparts. Rather than worrying about artificial general intelligence posing an existential threat to humanity, Chinese builders treated the technology primarily as an engine for marketing, practical workflows, and software development.
Practical Engineering Over Existential Dread
Major United States AI platforms like OpenAI, Anthropic, and Google remain mostly inaccessible behind China’s Great Firewall. In response, local developers rely on home-grown systems such as DeepSeek, Kimi, GLM, MiniMax, MiMo, and Alibaba’s Qwen. These models now handle everyday enterprise tasks effectively, without the apocalyptic framing common in American tech hubs.
Technical demonstrations highlighted a focus on efficiency and physical robotics. Developer AutoArk showed a 35 billion-parameter model running on an iPhone 16 Pro using only 2.5GB of RAM. Hardware builder DYNA demonstrated robotics designed to integrate with software models that receive monthly updates.
The Reality of the Factory Floor
Despite the fast pace of software, physical implementation faces sharp limits. Bottle cap manufacturer Zhu Shihui challenged consultant-driven hype, noting that many factory machines are completely pre-digital. His workforce has an average age in the late 40s and lacks advanced digital trainingTrainingTeaching an AI model by showing it very large amounts of data. Open in glossary. Because manufacturing margins are thin, full automation overhauls are often unaffordable. Zhu argued that AI builders must respect existing operational knowledge rather than trying to overhaul operations wholesale.
Speakers also tackled the role of the developer. Keynote speaker Ryo Lu questioned how engineers measure self-worth when AI writes code, arguing that builders should create tools that keep humans involved rather than rendering them obsolete.
How does this story land?
One tap per reader.
Quick poll
Can AI realistically modernize traditional factories that use pre-digital machinery?
One tap. No sign-up.
Ask this story
Still have a question?
Answers use only what is in this story. They are written by AI and can be wrong.