活动帧：面向智能体记忆与回放的确定性屏幕活动编译

摘要

计算机使用智能体需要付出完整的前沿模型推理成本，来重新推导其用户已经执行过的例行流程，因为当前智能体的记忆记录的是用户说过的话，而非用户做过的事。我们采用一条确定性、零模型的流水线，将被动捕获的屏幕活动编译为智能体记忆：该流水线将本地捕获流分割为类型化活动帧，即携带应用、站点、时间、输入量及指向原始行的证据指针的有界事件片段。流水线无需模型介入，因此输出字节一致、可缓存、可机械审计。在一位专业人士的单用户语料库上（51个活跃天，共128,756帧），该编译器将一天的原始捕获压缩为一个提示就绪的上下文块，体积缩小86倍，耗时68毫秒。读取该块的智能体在回答关于当天的问题时，以独立基准为参照达到98.4%的准确率（威尔逊95%置信区间：91.7-99.7%），而使用LLM对同一捕获生成的摘要仅能达到66-80%；中档模型读取该块的表现与前沿模型相当。同一编译器还可充当需求侧成本工具。它从被动的、委托前的人类活动中读取数据，而非智能体运行轨迹，从而提供两个参数——例行流程开销比R和例行流程重复率h——智能体成本模型一直假设它们存在，但据我们所知尚未对其进行测量。我们报告R的首批数值（作为建模上限）为60-343倍；可委托重复率在样本内为9.0%、样本外为7.7%；在这些数值下，现实的全量智能体令牌上限接近8%。编译后的例行流程可在模型不介入的情况下确定性重放，并在一次守卫匹配命中时以零模型令牌完成了现场演示。模式、编译器和评估框架均已开源。

English

Computer-use agents pay full frontier inference to re-derive routines their user has already performed, because an agent's memory today records what the user said, not what the user did. We compile passively captured screen activity into agent memory with a deterministic, zero-model pipeline: it segments a local capture stream into typed activity frames, bounded episodes carrying application, site, timing, input volume, and evidence pointers back to the raw rows, with no model in the loop, so the output is byte-identical, cacheable, and mechanically auditable. On one professional's single-user corpus of 128,756 frames over 51 active days, the compiler reduces a day of raw capture to a prompt-ready context block 86x smaller in 68 ms, and an agent reading that block answers questions about the day at 98.4% accuracy (Wilson 95% CI 91.7-99.7%) against an independent oracle, versus 66-80% for an LLM summary of the same capture, a mid-tier model reading the block matching a frontier one. The same compiler doubles as a demand-side cost instrument. Read off passive, pre-delegation human activity rather than agent rollouts, it supplies two parameters that agent-cost models assume but, to our knowledge, have not measured: the Routine Overhead Ratio R and the routine recurrence h. We report first values of R, a modeled upper bound, at 60-343x, and a delegable recurrence of 9.0% in-sample and 7.7% out-of-sample, for a realistic all-fleet token ceiling near 8%; a compiled routine replays deterministically with the model out of the loop, demonstrated live at zero model tokens on a guard-matched hit. Schema, compiler, and evaluation harness are open.