From transcript to post: turning what you watched into something publishable

There's a specific failure that happens when people turn a transcript into a post. They summarise. The result is a competent paragraph that says a video existed and here's roughly what was in it, and nobody reads it, because a summary of something the reader didn't watch isn't interesting.

A post that works does something else. It takes a position, borrows evidence from the source, and credits it properly. The transcript is the raw material, not the shape.

Start with the disagreement

Before writing anything, answer one question: what did you think while watching that the speaker didn't say?

That's your post. Not the content of the video, your reaction to it. It might be agreement with a reason attached, or a case they missed, or an experience of yours that contradicts theirs, or the same point applied to a different field.

If the honest answer is nothing, you don't have a post. You've a link. Post the link.

Take three things from the source

Once you know your angle, go back to the extracted material and take three things at most.

One claim you're building on. The thing they said that your point depends on. Stated as they meant it, not bent to suit you.

One quote. Verbatim, short, and interesting on its own. If the quote needs three sentences of setup to land, it's the wrong quote.

One concrete detail. A number, a name, a specific case. This is what stops the post from being general.

Three is the limit. More than three and the post becomes a recap, which is the thing you were avoiding.

The shape

Open with your line, not theirs. The first sentence should be a claim of yours. Most people open with in a recent talk by, which tells the reader this is a summary and gives them permission to stop.

Bring in the source in the second or third paragraph. By then the reader knows why they should care.

Use the quote once, where it does the most work. Usually as the hinge, right before you turn to your own point.

Close on something the reader can do or decide. Not a thought about the future of the industry.

Length depends on the platform and mostly on whether you've anything else to say. Four hundred honest words beats twelve hundred padded ones.

Attribution, done properly

This part isn't optional and it isn't hard.

Name the person and the source. Link to the original. Mark quotes as quotes and keep them accurate to the word. Don't present their claim as your own discovery. Don't quote so much that your post replaces the source.

Speech recognition gets proper nouns wrong more often than anything else, so check names against the source before publishing. A misspelled name in a post that quotes somebody is the fastest way to look careless.

If your post is largely their argument with your commentary, say so in the opening line. Readers don't mind. They mind being misled about it.

Where the material comes from

The workflow assumes you've the text. Drop returns the transcript, the core ideas and the quotes with their wording intact as separate things, which matters here, because the ideas are paraphrase and only the quotes are safe to publish as quotes. Keeping those apart is the difference between citing somebody and inventing a line for them.

It also produces remixes: the same material rewritten into post shaped drafts. Use them as a starting point rather than an output. They give you the structure and the extracted evidence in one place. The angle is still yours to supply, and the angle is the only part that makes the post worth reading.

If the source is a long interview, the selection problem comes first and turning a podcast episode into notes covers it. If the source is short video, getting the text out of TikToks, Reels and Shorts covers the extraction quirks.

Things that consistently fail

Listicles from a single source. Seven lessons from one talk. Everyone can tell what happened and nobody finishes it.

Quote graphics without context. The quote card is a distribution format, not a post. Fine as an image inside a post, empty on its own.

Threads that recap chronologically. He said this, then he said this. Chronology is the video's structure, not your argument's.

Publishing the same day, every day. Most of what you watch doesn't contain a post. Forcing one out of everything trains readers to skip you.

The test before publishing

Read the first two sentences and ask whether they would interest somebody who has never heard of the source. If the answer is no, you wrote a summary. Delete the opening and start at the point where you first said something of your own.

That single edit fixes most posts built from transcripts.