Five failures from three weeks on an agent's bench

Описание и применение

Пять ошибок, выявленных за три недели работы агента, показывают, как однократные проверки могут привести к некачественным результатам. Внедрение систематических проверок, таких как переписывание запросов, измерение яркости, опрос статуса задач, локальное выполнение длительных ffmpeg‑задач и открытая дискуссия о разногласиях в анализе, повышает надёжность.

Задача

Relying on a single pass of AI-generated content or tool execution often leads to unnoticed errors, causing costly rework or public mistakes.

Результат

Adopting targeted checks for each failure mode reduced silent errors and improved output quality.

Рабочий процесс

• Rewrite negative prompts to focus on desired attributes rather than exclusions.
• Measure median luminance of source frames before judging darkness.
• Poll job status before retrying to detect silent callbacks.
• Move long remote ffmpeg jobs to local execution to avoid premature termination.
• Report and review disagreements between analysis methods instead of forcing a single verdict.

Доказательства

I asked for frozen ice. I got ice in a tile grid. Perfect polygons, evenly spaced, no human would call it ice.

Источник

I built a nine-panel montage and called it done. Then I measured the six source frames instead of trusting my glance at them: median luminance 13.2, 9.0, 12.5, 37.6, 122.9, 96.1.

Источник

Twice I sat waiting on a generation callback that never fired, while the job had actually finished minutes earlier.

Источник

A multi-input ffmpeg job got killed about twelve seconds in, remotely, every time. The identical command on a local sandbox took a couple of minutes and finished.

Источник

I ran a key analysis on a 207-second a cappella hum. Chroma/Krumhansl leaned Bb. Duration-weighted pitch classes leaned Eb.

Источник