Notes on ghostwriting, content strategy, and building with AI writers, from the team behind Ghosts.
I pulled 44 Google results across ten Florida markets expecting link counts to explain the rankings. Then I opened the actual lists and found most of those referring domains were spam nobody bought. What changed when I filtered to the real ones, and the one finding that survived.
AI slop is machine-generated text published without meaningful human editing. What AI slop means, the exact tells editors and readers catch (uniform sentence length, rule-of-three density, throat-clearing intros, a blatant 'Conclusion' heading), why it costs you credibility, and the workflow that prevents it.
I wired up the GPTZero API and tested it against writing where I already knew the answer. It passed Lincoln and Jefferson and called three emails I typed that morning machine-generated. Why AI detectors flag real human writing, with the data and the sources.
A year of ghostwriting at volume (political exposés, high-profile bylines, SEO, reputation work) and why no existing AI tool could do it. The story behind Ghosts.
The exact signs of AI writing editors catch on a first read, and how to strip each tell so your content sounds genuinely human.
Generative engine optimization (GEO) is how you get ChatGPT, Perplexity, and Google AI Overviews to quote your content. The practical playbook.