前Hugging Face高管王铁震谈中美AI开源与闭源策略差异及蒸馏争议
英文摘要
Former Hugging Face Asia-Pacific ecosystem head Tiezhen Wang commented on diverging AI release strategies: Chinese companies favor open-weight models while US firms like OpenAI and Anthropic keep models closed. He argued that distillation is a neutral technique, noting the irony that US companies train on public internet data yet try to prevent others from reusing that knowledge. Wang stated all AI-generated content should be free of copyright to avoid power abuse by compute-rich entities. He observed that China's plentiful open-weight models lead to drastically lower token costs, prompting Chinese internet companies to encourage maximum token usage, push employees to become AI-native developers, and even ban manual document writing.
中文摘要
前Hugging Face亚太生态系统高管王铁震评论中美AI发布策略差异：中国公司偏好开源权重模型，而OpenAI和Anthropic等美国公司采用闭源。他认为蒸馏是中性技术，讽刺的是美国公司用互联网公开数据训练模型却试图阻止他人复用知识。他主张所有AI生成内容不应享有版权，以免拥有算力者滥用权力。他指出中国丰富的开源权重模型使Token成本极低，促使公司鼓励员工最大化使用Token、成为AI原生开发者，甚至禁止手动撰写文档。
关键要点
China favors open-weight AI models; US companies like OpenAI and Anthropic remain closed.
中国偏好开源权重模型，美国OpenAI、Anthropic等保持闭源。
Wang calls distillation neutral and criticizes US firms for trying to block reuse of knowledge trained on public internet data.
王铁震称蒸馏是中性技术，批评美国公司试图阻止他人复用基于公开互联网数据训练的知识。
He argues AI-generated content should be copyright-free to prevent abuse by those with compute power.
他主张AI生成内容不应有版权，以防算力拥有者滥用权力。
Open-weight model availability in China drastically lowers token costs, leading companies to encourage maximum token usage and ban manual document writing.
中国开源权重模型使Token成本极低，企业鼓励最大化使用Token，甚至禁止手工撰写文档。