Field notes / growth / 17 November 2026
13 days, 6k to 10k followers: the most valuable part of that teardown is one comparison
In that teardown of 6k to 10k followers in 13 days, the most valuable part is one same-day comparison.
Speed is a publishing window, not diligence
- - DeepSeek released on 17 October; he ran model tests the same day.
- - GPT image 2.6 shipped; the sticker skill was published within three hours.
- - Over the next two days it was iterated five times; the repo went from 500 to 1.9k stars.
His rule used to be "publish when it is finished". It is now "publish when it can be caught".
That shift had a price: he spent two weeks building a memory-site product, published it, nobody used it for a day, and he cut it. The test changed from "am I done" to "is anyone catching this".
The two content types he stopped making
| Type | What the data showed | Why he stopped |
|---|---|---|
| Pure tool round-ups | High saves, low follows | They do not retain readers; useful as filler only |
| Very niche tracks | Plays but almost no saves | Cannot break out of the niche |
His correction: popular model plus an underserved use case. Going somewhere less crowded is not the same as walking into a desert.
What this means if you ship products
If you have a working project, it is raw material, not a landing point.
A landing point is the one thing a reader can do immediately after reading: a command, a checklist, a repo that runs. He iterated the skill to nine use cases and 1.9k stars in two weeks, not by writing better, but by making every version directly usable.
My own account has 39 projects, a pile of failure logs and real numbers, and almost every post says "here is what I built" instead of "here is what you can do with it".
The original post: https://x.com/Leviwoyun/status/2122583907581223172
His last line is: LIKE WIND, LIKE WATER, LIKE FIRE. I will not explain it.