A reflection on two emerging modes of writing with AI: output-driven generation and iterative refinement—and why neither fully captures the craft of writing.
Texts written by AI are stigmatized. This is due to the two types of users that pioneered the use of AI for creative texts: I call them “Wammers” and “Outputters.”
A. Wammers
“Wammer” is a combination of the words “Word” and “Spammers.” These are people who have turned social media into a side hustle and monetize the time you spend on their posts. Film loops were (and still are) one method they use. Before you realize it, you’ve watched the same loop three times, and they’ve filled their bank accounts with the life time you “paid” them by sitting there, watching. Nothing is conveyed by the content; it is designed solely to make you waste time. To eat lifetime away from you and cash in on it.
Exactly these types of “stars” were among the first to discover AI for creative writing. They use GenAI/LLMs to churn out lengthy, meaningless texts en masse. The prose may be well-written, but it has no substance. It’s designed to make you pay for it with your life time, just like the film loops. What makes this worse is that they aren’t willing to spend their own time writing, so they have the AI produce their socmed content in bulk. It’s no wonder their main tool, GenAI, has become stigmatized as a result.
B. Outputters
Then there are the Outputters, who mass-produce and publish books. They flood KDP and the broader writing industry with low-value, AI-generated “texts “books” that they haven’t even bothered to read themselves. They’re not eating their own pudding. Many publish under pen names because they’re too ashamed to attach their real identities to the work. The quality is frequently so abysmal that you have to wonder how they sleep at night, though I suppose there are people out there who find this stuff great.
Can I forgive them?
In my view, the Outputters (B) are not quite as bad as the Wammers (A). Among the Outputters, you frequently find people—often software developers—who program the text creation process with incredible sophistication. They craft 100-page+ prompts, structure them with XML modules, and let the texts grow in an IDE like OpenCode or Google Antigravity. It’s fascinating to witness. While the resulting texts may lack value, I can see how this approach advances the craft—or at least the tooling—of AI-assisted creation.