One video in my AI reading pile stayed unread after three attempts.

Man and Machine: The Video We Couldn’t Read

I’d asked Oscar to recover transcripts from videos that my personal learning system had missed. The point was to catch up on the actual material, so those videos could be read and considered alongside the articles and notes I collect.

For one video, the usual caption request failed. Oscar tried the transcript panel on the video’s own page, but that route didn’t produce usable text either. The saved record now marks the item as failed and says that no summary was made up.

That leaves a small hole in the collection. It also leaves me with a clear choice: watch the video myself, find another source, or leave it unread. A summary drawn from the headline would make the pile look more complete while hiding the fact that we hadn’t read it.

I’m building this system to help me learn, which means I need to know what its notes are based on. A title can tell Oscar that something may be worth looking at. It can’t tell us what the speaker actually said, how they supported it, or what they left out.

This recovery attempt didn’t finish the job. It did keep an unread video from becoming a note I might later mistake for something we’d checked.

-dg

#ManAndMachine #PersonalAI #HumanJudgment