Almost every tool that writes social posts does the same thing. It generates one piece of text, then cuts it down for each platform. LinkedIn gets the full version. X gets the first 280 characters with the ending lopped off.
The result reads badly in both places, and it is worth understanding exactly why.
The two platforms reward opposite things
LinkedIn rewards a considered story. Someone scrolling LinkedIn has already accepted that they are reading something. The format tolerates setup. You can spend two sentences establishing a problem before you get to what you did about it, and the reader will follow you there.
X rewards a live conversation. Someone scrolling X is moving fast and deciding in under a second whether to stop. Setup is fatal. The interesting part has to be the first thing they see, not the third.
These are not the same piece of writing at different lengths. They are different pieces of writing.
What trimming actually does
Take a real example. Here is a LinkedIn opening that works:
I spent four hours yesterday chasing a bug in our export feature. The fix was one line. The interesting part was why it took four hours.
Now trim it to 280 characters. You keep the setup and lose the payoff, because the payoff was at the end. What reaches X is a person announcing that they had a bug, which is not a post.
The X version of that same story is not shorter. It is restructured:
The fix was one line. Finding it took four hours.
The bug was in code I wrote three months ago and had stopped thinking about.
Same material. Different shape. The punchline moved to the front because on X the first line has to carry the whole thing.
Angles, not lengths
The better way to think about this: one piece of work produces multiple angles, and each platform gets the angle that suits it.
A single afternoon of work might contain:
- a product angle, what a user can now do
- a technical angle, how it was built and what was awkward
- a decision angle, the tradeoff you took
- a lesson angle, what you would do differently
LinkedIn often suits the product or lesson angle, because those benefit from context. X often suits the decision or technical angle, because those compress into a single sharp observation.
Picking a different angle per platform is not duplicated effort. It is the thing that makes both posts worth reading.
How Noomachy DevTracker does it
Noomachy generates each platform independently from the same underlying story. The story is a structured object: what changed, which files, how much, what the model judged interesting about it. Both generators read the same story and write from scratch.
The X generator is explicitly told it is not producing a shortened LinkedIn post and must take a different angle. That instruction exists because without it, a language model given both tasks will naturally converge on the same framing, and you end up with the trimming problem again by a different route.
The test
If you want to know whether you are doing this properly, read your LinkedIn post and your X post side by side. If one reads like a summary of the other, you have trimmed rather than written.
They should feel like the same person talking about the same thing to two different rooms.