当通用对话模型的及格线被拉平，巨头与创企们不约而同地将目光投向了下一个战略高地，也就是垂直专业领域，尤其是代码（Coding）与智能体（Agent）赛道。与吟诗作对的闲聊不同，写代码和执行复杂任务，要求模型必须具备慢思考的能力。而这也是检验大模型是否能从“玩具”变成“生产力工具”的唯一标准。
Apple M5 MacBook Air: The specs。业内人士推荐夫子作为进阶阅读
In 2002, a team of researchers from NASA, the F.A.A., and six commercial airlines ran a series of experiments in Oklahoma City, in a decommissioned Boeing 747. They wanted to see how quickly a commercial jet could be secured in case of turbulence. They recruited a group of volunteer passengers, gave them all fake boarding passes and luggage, and had three of them carry around life-size baby dolls. Some of the volunteers were told to stay in their seats and act as if they were sleeping, reading, or working on a laptop. Others were told to stand in the aisles or sit in the lavatories. A crew of experienced flight attendants, drawn from the airlines participating in the study, made their way up and down the plane, serving fake food.，更多细节参见体育直播
A week before the release of Opus 4, which triggered ASL-3 protections for the first time, Anthropic changed the RSP so that ASL-3 no longer required being robust to employees trying to steal model weights if the employee has any access to "systems that process model weights".，推荐阅读51吃瓜获取更多信息