微软首席执行官萨蒂亚·纳德拉在CNN节目中发出警告，指出完全依赖单一AI提供商的企业可能无法长期存活1。纳德拉表示，"任何没有这种控制权的公司，我敢说将不复为一个公司，因为你本质上已经把自己的思维外包出去了"1。他强调企业需要保有数据的自主权和多模型选择能力，以维护独立的竞争地位。
为了避免过度依赖AI实验室，纳德拉建议企业采取关键举措1。首先，企业应保留围绕数据的所有元数据，用于训练自有权重或开源模型1。其次，企业应避免过度依赖AI实验室内置的编码工具，如Claude Code和ChatGPT Codex等1。第三，企业可以通过将接口层与AI模型分离、将上下文和记忆机制从模型中独立出来的方式，灵活使用多个AI模型1。纳德拉指出，"通过让接口层独立于模型、让上下文和记忆独立于模型，你完全可以使用多个模型"1。
值得注意的是，微软本身是Anthropic和OpenAI的投资者1。纳德拉还强调，消费者与企业的情况有所不同，对消费者而言共享数据是使用服务所需付出的代价1。
Microsoft Chief Executive Satya Nadella has cautioned businesses against placing complete reliance on a single artificial intelligence model provider, arguing that such dependency could threaten their survival 1. Speaking on CNN, Nadella stated that "any firm that doesn't have this control, I will claim will not remain a firm because you've essentially outsourced your thinking" 1.
To maintain independence and competitive advantage, Nadella recommended that enterprises retain ownership of their usage data and metadata to train proprietary models or leverage open-source alternatives 1. He emphasized the importance of separating AI harnesses—built-in coding tools offered by providers such as Claude Code and ChatGPT Codex—from the underlying model itself 1. "By keeping the harness separate from the model and the context and memory separate from the model, you absolutely can use multiple models," Nadella explained 1. This architectural approach would enable organizations to adopt multiple AI providers simultaneously without vendor lock-in 1. Nadella distinguished between enterprise and consumer usage, noting that while consumers may share data as the price of service access, businesses should maintain strategic control over their AI infrastructure 1.
评论
还没有评论，欢迎留下第一条。