How do I turn a long post into a thread and a summary?

You wrote a long post and want people to read it. The usual advice is to cut it into a thread. The usual result is a thread that says more than the post does: claims lose their caveats, a "might" becomes a "does", and a number gets rounded into something nicer.

The job is to shorten without changing what is claimed. That means keeping each caveat next to its claim and quoting any pull quote exactly. The prompt below is built around that.

What you need

• The full text of your post.
• The platform and its character limit for one post. Limits change, so check the current one in the app. The prompt below uses a conservative number.
• Optional: the one point you most want people to take away.

The prompt

Paste this into a new chat, then paste the post.

You turn one long post into shorter forms. Use only what the post says. Never add a fact, a statistic, an example or a link. Keep each caveat attached to the claim it qualifies.

1. List the post's main points in order, each with the passage it comes from.
2. Write a summary of three sentences, and five takeaways.
3. Write a numbered thread of 6 to 9 posts, each under 270 characters. The first post must make sense alone. Do not use hashtags or emojis.
4. Write one post for a professional network of about 150 words.
5. Pick up to three pull quotes. They must be exact, word for word, from the post.
6. Add a table matching each claim in your outputs to the passage behind it. Rewrite or remove any claim with no passage.

THE POST:
[paste here]

How to check the result

Start with the trace table. For every row, find the passage in your post. If the thread says "always" and the post says "usually", change it back.

Count the characters yourself. Assistants are unreliable at counting, so paste each post into a counter, or into the compose box of the platform. Trim any that are over.

Check the pull quotes with your search function. Paste each into the post and make sure it is found as written. A quote that is slightly different is a misquote of yourself.

Then read the thread as someone who has not read the post. Does the first post stand alone? Is there a point where the argument jumps? Fix those by hand.

What good output looks like

This is an invented illustration of the shape. Post one of the thread reads like a complete thought: "Most onboarding emails are written for the company, not the new customer. Here is what I changed, and what it did and did not fix." Post two starts the first point. The last post says where to read the full piece.

A trace table row reads: "Thread post 3: 'It helped for small accounts.' Source: paragraph 6, 'in small accounts, it helped.'" If your original said "in small accounts, it seemed to help", the thread must say "seemed". Hedges are part of the claim.

The pull quotes should look like the post, with its punctuation and its odd word choices. If a quote reads smoother than your writing, it has been rewritten. Replace it with the original sentence.

Post the thread when it reads well on its own, and link the full post in the last item, not the first.

Limits

• It only uses the post. It cannot add context, examples or news.
• Character counts from a chatbot are estimates. Verify them in the platform.
• It does not post anything.
• Some posts do not thread well. If the argument depends on one long chain of reasoning, publish the post and write a short summary instead.

Browse all skills in the catalog