Harness Continual Learning：冻结模型下的智能体持续学
An agent can improve without retraining the model.
Agent 架构中避免微调成本的新思路，用回归测试解决动态配置带来的稳定性问题，值得 Agent 开发者关注。
An agent can improve without retraining the model.
智能体可以在不重新训练模型的情况下实现改进。
Once agents rewrite their own prompts, memory, and routing, every update becomes a behavioral change.
一旦智能体重写自身的提示词、记忆和路由，每一次更新都会转化为行为上的改变。
As agents keep rewriting prompts, memories, skills, and routing rules, those updates become part of what the system has learned.
随着智能体不断重写提示词、记忆、技能和路由规则，这些更新将成为系统已学习内容的一部分。
Fixing today's failure can quietly break something that worked yesterday.
修复今天的故障可能会悄无声息地破坏昨天原本有效的功能。
The paper calls this harness-level forgetting: the model stays frozen, but the behavior around it keeps changing.
论文将这种现象称为“Harness 层面的遗忘”：模型本身保持冻结，但其周围的行为却持续发生变化。
Their Harness Continual Learning (HCL) framework puts every proposed harness update behind a gate.
他们的 Harness 持续学习（HCL）框架为每一项提议的 Harness 更新设置了一道关卡。
For evolving agents, prompt, memory, skill, and routing changes should be handled like code changes: regression-test them before they become persistent.
对于不断演化的智能体，提示词、记忆、技能和路由的变更应像代码变更一样处理：在它们变为持久状态之前进行回归测试。
– arxiv. org/abs/2608.19013
– arxiv.org/abs/2608.19013
Title: "Harness Continual Learning: Continual Adaptation Beyond Model Parameters"
标题：“Harness 持续学习：超越模型参数的持续适应”
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