EVENT DOSSIER
Amazon is purchasing a large number of rare books and scanning them at its facility codenamed VGT3 in Las Vegas. According to tracking devices, the company uses the data from these texts that cannot be accessed via the internet by cutting the book spines and scanning them for training large language models. Amazon responded that this initiative aims to purchase books through commercial channels to improve products and services for customers. Due to limited internet content, rare books provide valuable training data sources.
亚马逊正在购买大量珍稀书籍，切断书脊并扫描以用于 AI 训练。根据 404 Media 的报道，该公司在拉斯维加斯的一处设施（代号 VGT3）内放置了追踪设备，确认了该过程。亚马逊方面回应称，其通过商业渠道采购书籍是为了改善客户使用的产品和服务。由于大型语言模型需要海量文本进行训练，且互联网上可获取的内容有限，这些无法在互联网找到的珍稀书籍提供了宝贵的训练数据源。