Meta is paying to peek at how you use their latest AI model
Meta is paying to peek at how you use their latest AI model
Meta 付费“窥探”你如何使用其最新 AI 模型
Most AI tools allow you to opt out of sharing your usage with the model provider to improve future versions. Meta has taken that idea and put a price tag on it. For its new Muse Spark model, intended for operating coding and other agents, Meta is offering an explicit discount averaging out to about 95% for users who “contribute” to the development of future models by sharing their prompts and model outputs. 大多数 AI 工具都允许用户选择不与模型提供商共享使用数据，以避免这些数据被用于改进未来版本。Meta 将这一理念付诸实践，并为其标上了价格。针对其旨在运行编程及其他智能体（Agent）的新模型 Muse Spark，Meta 提供了一项明确的折扣——如果用户通过共享提示词（Prompts）和模型输出来“贡献”未来模型的开发，平均可获得约 95% 的折扣。
While 1 million input tokens under a standard agreement costs $1.25, under the contributor pricing model they cost just 10 cents. For output tokens, the standard price is $4.25 per million, but that same million costs just 20 cents under the contributor model. 在标准协议下，100 万个输入 Token 的成本为 1.25 美元，而在“贡献者”定价模式下，成本仅为 10 美分。对于输出 Token，标准价格为每百万个 4.25 美元，但在贡献者模式下，同样的数量仅需 20 美分。
Meta has had a rough time trying to obtain training data: An initiative to track the computer usage of its employees, launched earlier this year, attracted wide internal criticism and was paused in June. The company didn’t respond to a question from TechCrunch about its new pricing model. Meta 在获取训练数据方面一直困难重重：今年早些时候发起的一项追踪员工电脑使用情况的计划引发了广泛的内部批评，并于 6 月被迫暂停。对于 TechCrunch 关于其新定价模式的询问，该公司未予置评。
This kind of user data is vital for making agentic tools work better. “The reason we saw a big jump in [coding agent] capabilities between April 2025 and October 2025 was that Claude Code, by default, would store all your coding agent sessions and use them for reinforcement learning training,” Mario Zechner, the developer behind the open source harness Pi, told TechCrunch last month. 此类用户数据对于提升智能体工具的性能至关重要。开源工具 Pi 的开发者 Mario Zechner 上个月告诉 TechCrunch：“我们之所以看到 2025 年 4 月至 10 月间 [编程智能体] 能力的大幅飞跃，是因为 Claude Code 默认会存储你所有的编程智能体会话，并将其用于强化学习训练。”
But even as the imperative for model builders increasingly becomes deploying agentic tools for use outside of software engineering, their ability to evaluate and improve those tools is blocked by the complexity and lack of digital traces for many professional workflows. 尽管模型构建者越来越迫切地需要将智能体工具部署到软件工程之外的领域，但由于许多专业工作流程的复杂性以及缺乏数字足迹，他们评估和改进这些工具的能力受到了阻碍。
Arvind Narayanan, a Princeton computer science professor, noted that there is good evidence that large companies don’t want their data to be used for model training. “They stick with token-billed Enterprise plans even though the subscription-based consumer plans like Claude Max and ChatGPT Pro are discounted by 10x-20x or even more! (The main difference between the plans is data retention + enterprise IT governance),” he wrote on social media. 普林斯顿大学计算机科学教授 Arvind Narayanan 指出，有充分证据表明大型企业不希望其数据被用于模型训练。他在社交媒体上写道：“尽管像 Claude Max 和 ChatGPT Pro 这样的订阅制消费者计划价格便宜了 10 到 20 倍甚至更多，但企业仍坚持使用按 Token 计费的企业版计划！（这些计划的主要区别在于数据保留和企业 IT 治理）。”
Perhaps in recognition of those dynamics, Meta is offering companies explicit compensation to obtain that information. Its pricing guide notes that the contributor tier “lowers the barrier to entry for prototyping, testing integrations, and scaling experiments where training on your data is acceptable.” 或许是意识到了这些动态，Meta 正在通过向企业提供明确的补偿来获取这些信息。其定价指南指出，贡献者层级“降低了原型设计、测试集成以及在允许使用数据进行训练的情况下扩展实验的准入门槛。”
That, Narayanan suggested, could in turn incentivize large companies to be more diligent about which data is truly proprietary and which could be shared with model providers. The framework could also play into growing price competition between the frontier labs. Anthropic’s newest Fable and Mythos models, released yesterday, came with lowered costs for processing cached tokens, while OpenAI’s latest models got major price cuts at the end of July. Narayanan 认为，这反过来可能会激励大型企业更审慎地甄别哪些数据是真正的专有数据，哪些可以与模型提供商共享。这一框架也可能加剧前沿实验室之间日益激烈的价格竞争。Anthropic 昨日发布的最新 Fable 和 Mythos 模型降低了缓存 Token 的处理成本，而 OpenAI 的最新模型也在 7 月底进行了大幅降价。