Content will make you rich, but NOT SLOP!
As you can see my account has 115K karma and I love writing content. I recently helped a friend get to 50K on twitter in 2 months.
Why content matters now
Building products is "easy". Distribution is the bottleneck. And the cheapest distribution channel when you have no audience is content: tweets, Reddit posts, blog posts, LinkedIn. Content is the first skill you need to master to sell anything you ship.
Even outbound is content, it's how you will write the right DM to the right person at the right time.
What notslop is
A lot of friends asked me to share my content systems so I made a little CLI called notslop. Please do not expect magic, i built it in one day but i believe it's a good representation of the systems I hae in place.
With it you can:
- write a tweet, X article, LinkedIn post, Reddit post (tuned to the sub), Reddit reply, blog post, cold DM grounded in the recipient's recent posts
- REPURPOSE ANY OF THE ABOVE INTO ANY OF THE OTHERS (this one is important and it's how the best content is made)
- get a deduped feed of what is being said about a topic across the subs, blogs, and X accounts you care about, in 2 seconds. The same news from Reddit, HN, and X collapses to one entry, you don't have to scan 5 versions
- see what is trending in your niche right now, or track mentions of a topic over 7 days clustered into themes
- see the top voices on a topic, your favorite accounts, the experts you respect, anyone, and write your version grounded in what they actually said today
What is good content
Good content is timely. Most of what we say is repurposing. The thing that matters is having the right info at the right time, with the right tone and style.
For that, you need two things:
- A way to get timely and relevant data. I use Orthogonal for this. One API key, $10 free credits at signup, gives access to X scraping (and via ScrapeCreators, also Reddit, LinkedIn, TikTok, etc. for future skills).
- A way to rerank and embed that data. I use ZeroEntropy. zerank-2 takes the ~200 raw posts you scraped and picks the top 10 most relevant to your topic. zembed-1 dedups the same news appearing across Reddit, HN, blogs, and X (you only see it once) and clusters noisy mentions into themes. Raw scraped data is NOT usable as-is, Claude would drown in noise !! Rerank + embed are what turn it into context that fits in a prompt window and actually matters.
One warning
Please take notslop as an amplifier, not as something magical. AI is a big lever, it does not turn a bad take into a good post. If you have NOTHING to say it won't help you !!
Please contribute back
The 14 content skills are starting points. The quality ceiling on each surface comes from people who write that kind of content every day. If that is you, your taste is what is missing. Sharpen the relevant SKILL.md, send a PR.
Repo
I am not affiliated with ZeroEntropy or Orthogonal. You can use other models for reranking and embeddings, and other scrapers for the data. The code is MIT.