Rewriting a Quote on Monday, I Realized My Copy Had Nothing Left to Say

At 10 p.m. last Monday, I was revising a quote email for a freelance project in Shenzhen. The client asked why I recommended content operations, so I wrote three paragraphs: “Unlock brand growth,” “build a closed loop,” and “unlock potential.” Before sending, I deleted all of it and started again. That was when I realized: AI isn’t useless, but I had let it pull me into writing like a template nobody believes.

That is what the No AI Slop Manifesto is about. It isn’t another writing tool; it’s a set of judgment criteria: fewer sweeping conclusions, fewer trend words, no pretending everything can be solved, and more visibility of the actual person, the actual problem, and the actual next action. The manifesto traces the idea to a discussion at the AIE CODE conference. Host Dex proposed it in his closing keynote and invited several speakers in attendance to co-sign it.

I also tested it in my own client group. Replacing “improve efficiency” with “move the two hours I spend each week organizing reports from Monday to Friday” got a much faster client reply.

What I Actually Spent Trying to Replicate It

The replication cost was almost zero: ¥0 and about 30 minutes for my first check. The technical barrier is low: I only needed a browser, the checklist, and a piece of copy to review—no coding required. The first step is to open slopmanifesto.com, read the manifesto from beginning to end, then choose a recent client email and review it against the criteria.

I wouldn’t start by rewriting everything. First, I’d find the sentence that sounds most like filler. For example, replace “provide comprehensive support” with “complete the two missing pages of information by Wednesday.”

If I were still unsure, I’d test it on a real client email rather than my personal bio. I’d keep the old version and compare reply length, follow-up questions, and whether the conversation continued. This may not work for everyone. If my work depends on formal language for industry reports, there’s no rush to try it.

How I’d Use It at Each Stage

Just Getting Started

I’d only review my landing page and direct messages, replacing one piece of filler each day with a concrete number or action. The goal wouldn’t be to sound impressive. It would be for prospective clients to know what happens next.

Working With 1–2 Clients

I’d turn the manifesto’s criteria into a reusable checklist and run through it before sending every proposal. I’d also keep both versions, before and after. Once I had more clients, my team could use the same language standards.

Scaling Up

I’d include “does this sound like AI wrote it?” in content reviews, but I wouldn’t rely on detector scores alone. I’d spot-check sales emails, client updates, and personal-brand posts so greater efficiency didn’t quietly reduce trust.

I make mistakes too, and AI has pulled me off course. I don’t think we need to reject AI; we need to stop its polish from hiding the useful information. Start by changing one sentence in the email you’re about to send today. That’s enough.