Skild发布机器人基础模型S1，视频演示即可定义任务
Robotics is more and more getting close to get its version of in-context learnin…
Robotics is more and more getting close to get its version of in-context learning.
机器人技术正越来越接近实现其版本的上下文学习。
Skild just released S1, a robot foundation model that uses video demonstrations to define tasks instead of language instructions.
Skild 刚刚发布了 S1，这是一个机器人基础模型，它使用视频演示而非语言指令来定义任务。
Give S1 1 human video showing a long, multi-step task, and the robot executes it straight away. No retraining. No fine-tuning.
给 S1 一个展示长序列、多步骤任务的人类视频，机器人就能立即执行。无需重新训练，也无需微调。
If this scales, you pay the enormous data bill once during foundation-model training, then amortize it across thousands of new tasks through prompting. That could change the economics of robot learning completely.
如果这种模式能够扩展，你只需在基础模型训练阶段支付一次高昂的数据成本，然后通过提示（prompting）将其摊销到数千个新任务中。这可能彻底改变机器人学习的经济模式。
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力