Why video is inefficient
- It's paced by the speaker. You can't slow down for the hard two minutes and speed through the easy twenty without constant manual intervention.
- It's not searchable. Finding the bit where they explained the thing means scrubbing.
- It produces no artefact. An hour watched leaves nothing you can review, unlike an hour reading with notes.
- It feels like learning. Video comprehension is fluent and immediate, which produces the same illusion of knowing as rereading, more strongly.
- Speeding it up doesn't fix it. 2x halves the time and keeps every other problem, and comprehension drops on unfamiliar material.
None of that means video is bad — for demonstrations, spatial explanations and worked procedures it's often the best medium available. It means video should be an input to note-making rather than a substitute for it.
The conversion
- 1
Get the transcript before watching
Most educational videos have captions, auto-generated or otherwise. A transcript is searchable, skimmable and about ten times faster to work through than the video. This single step is most of the efficiency gain.
- 2
Skim the transcript for structure
Two minutes. What are the sections, and which are relevant to you? A typical hour-long talk contains fifteen minutes you need, and finding them in text is trivial where finding them in video isn't.
- 3
Watch only the segments that need watching
Diagrams being drawn, procedures demonstrated, anything visual. Everything else you can read faster than they can say it.
- 4
Extract claims, not summaries
From the transcript: what is this video asserting that you didn't already know? Usually five to fifteen things. That's the actual content of an hour of video.
- 5
Convert each claim into a question
Now it's reviewable. This is the step that separates 'I watched a video' from 'I learned something', and it takes about ten minutes.
- 6
Note what to verify
Video content is unrefereed. Anything that contradicts your course material or sounds surprising goes on a check-later list rather than into your notes as fact.
Which videos are worth this
| Video type | Worth converting? | Note |
|---|---|---|
| Your own recorded lectures | Yes, always | Highest-value case. It's your examinable material — see how to use recorded lectures. |
| University course recordings from elsewhere | Usually | Good depth, but check it matches your syllabus before investing. |
| Explainer channels | Sometimes | Great for intuition on one hard concept; thin as primary material. |
| Exam-technique and revision channels | Rarely | Usually generic. Your department's past papers dominate these. |
| Demonstrations and procedures | Watch, don't convert | The visual is the content; a transcript loses it. |
| Conference talks and seminars | Yes if relevant | Dense and rarely written up elsewhere. |
The verification problem
Educational video is unreviewed. A confident, well-produced explanation can be subtly wrong, out of date, or right in general and wrong for your course's conventions — and there's no correction mechanism beyond a comments section you won't read.
So treat video as a source of understanding rather than a source of facts. If a video makes a mechanism click, excellent — that's what it's best at. But the facts, the definitions and the notation you'll be examined on should come from your own material, and where the two conflict, your lecturer wins. This is the same argument as AI study tools that cite sources, applied to a different medium.
When the video is the only source
Sometimes there's no textbook — a niche topic, a new tool, a technique your lecturer demonstrated once. Then the conversion matters more, not less, because there's nothing else to revise from.
- Write the procedure as numbered steps and then do it without the video. Watching a procedure is not learning it.
- Screenshot the diagrams, and redraw them yourself from memory afterwards.
- Note the timestamps of anything you'll need again, so you never have to scrub.
- Find one written source that agrees, even a brief one. A single unverified source is a risk you can usually cheaply remove.
The honest use of video
Video is unusually good at three things: making a hard concept click when the text isn't working, showing something spatial or dynamic, and getting you started on a topic you find intimidating. Those are real and worth an hour.
It's poor at building retrievable knowledge, which is why a term of watching lectures at 2x produces so little. If you use video, use it early — for understanding — and let the retention come from questions you generated afterwards, reviewed on a schedule. The medium changes; the mechanism doesn't.