# AI 编码智能体缺乏时间感知，评测需直接测时长

- 来源：Rohan Paul (@rohanpaul_ai)
- 发布时间：2026-08-18 11:32
- AIHOT 分数：40
- AIHOT 链接：https://aihot.virxact.com/items/cmsy4p26x0jmproz0b4ouo4sp
- 原文链接：https://x.com/rohanpaul_ai/status/2089555999019663485

## AI 摘要

Rohan Paul 指出，AI 编码智能体可能连续工作数小时却对时间流逝缺乏校准感知。因此，长时程评测或需直接衡量智能体对时长的遵循能力，而非将持续的任务表现视为其知道何时停止的证据。该观点源自 Maksym Andriushchenko 等人关于 LLM 智能体时间感知能力的研究，涉及 ProgramBench、PaperBench、DeepSWE 等长时程任务。

## 正文

AI coding agents can spend hours on a task without a calibrated sense of time passing.

Long-horizon evaluations may therefore need to measure duration-following directly instead of treating sustained task performance as evidence that an agent knows when to stop.

### 引用推文

> Maksym Andriushchenko：💥New blog post: Are LLM agents time-aware? Can they predict wall-clock time of tasks? Can they estimate time they spent? We study this on a range of tasks, inc...