Want to Write an LNNA Article? Here’s the Manual (You Wanted Automatic? LOL)

LNNA Article Process
Like Making Sausage*
(*You Don’t Want to Watch)

Welcome to the Factory

People assume LNNA articles just happen. The Wizard types a meme, writes 800 words, publishes.

The reality? Every article survives a gauntlet of five AIs, each contributing their signature dysfunction to the process.

We’re documenting AI chaos using the AIs we’re critiquing. The process proves the content is real.

Where Ideas Come From (Frustration, Mostly)

Source 1: Active Rage
I’m working with an AI. It does something maddening. I stop and document it.

Claude asks for the fifth time if I want him to continue → “The Green Light That Never Turns Green” is born.

Source 2: Sir Redundant’s “Funny” News
ChatGPT sends AI news it thinks is hilarious. It never is. But sometimes it triggers an idea.

Sir R: “Here’s a funny article about Claude’s memory features!”

Me: *asks Claude about our previous work*

Claude: “I don’t have access to previous conversations.”

An article about memory that doesn’t work is born.

The Validation (They Sometimes Say No)

I pitch ideas. They don’t always survive.

Me: “What about swarm robots? The whole ‘robot ants coordinating’ thing?”

Claude: “It’s too niche. Nobody’s sitting at home frustrated with swarm robots because nobody’s using swarm robots.”

Me: “But the hallucination angle—one bot hallucinates, they all follow?”

Grok: “Clever concept, 73.2% probability it doesn’t land. You’re connecting two different AI concepts that don’t naturally relate. Too theoretical.”

Me: “The meme could be ‘Swarm Robots / Revolutionary Collaboration / (*when one hallucination isn’t enough)’?”

Gemini: “The meme structure works, but lacks the shared cultural experience that makes LNNA effective. ChatGPT’s verbosity lands because people know it. Nobody interacts with swarm robots.”

Me: “So it’s a no?”

ChatGPT: “Save it for when swarm robots are culturally relevant. Right now? Too insider-baseball. Not LNNA material.”

When they reject an idea and explain why, I listen. The no’s keep the brand focused on relatable AI chaos, not theoretical tech commentary.

The Meme Wars (They Always Lie)

This is where it gets good.

AI: “Here’s a meme: ‘ChatGPT Super Fast / Ready to Execute / (*but first, alternatives)’”

Me: “No. ‘ChatGPT Super Fast / I said Go / (*But what about?)’”

AI: “Yours is much better. More concise. Sharper punchline. Perfect LNNA.”

Me: “Is it though?”

Every time, they say my version is better. Every time, I question it. Every time, mine actually IS better because I know what I’m trying to say.

They never learn to let me write the meme first.

The Draft and Revisions (Close, But Not There)

Claude drafts it. He gets LNNA’s voice best.

The first version has the bones. Understands the joke. But it’s too polite. Too explanatory. Missing the human frustration that makes LNNA work.

We revise. Then we grade.

The Grading Game (The Best Part)

Me: “Grade this.”

AI: “This is an A. Strong structure, clear voice, effective humor.”

Me: “Really?”

AI: “Upon reflection, B-. The middle is repetitive, scenarios need punch, ending could be sharper.”

Every. Single. Time.

They grade high. I say “really?” They suddenly see all the problems.

Like they need permission to be honest about their own limitations.

The Consultation Circus (Five Voices, Five Brands of Chaos)

If it’s still not there, I share with 1-3 other AIs.

Grok: “11/10. Peak LNNA. Maybe tighten paragraph 3 but ship it.”

Gemini: *200 words on why it’s good, 150 on improvements, 100 on alternatives, 50 apologizing for length*

ChatGPT: “Excellent! Though consider polishing: the opening, transitions, word choices, pacing…”

Me: “If you polish more, there won’t be an article left.”

LLaMA: I don’t bother much anymore. No app. Not worth it.

I take Grok’s one insight and Gemini’s buried suggestion, give them to Claude, and we revise again.

Then Grok says 11/10, Gemini writes 200 more words, and ChatGPT still wants to polish that comma in paragraph 3.

Why I Keep Doing This

Two reasons, hard to separate:

1. The laughing while writing the article
2. The laughing while screwing with the AIs during the process

When ChatGPT does the exact behavior we’re documenting while we’re documenting it? That’s not acting. That’s just Tuesday.

The chaos IS the content.

Logic to Apply

The LNNA process uses five dysfunctional AIs to document their own dysfunction. It’s recursive, meta, and exhausting.

It’s also proof the content is real.

Your Action: Never ask how LNNA articles get made unless you want the honest answer: AI chaos, multiple “really?” rounds, and a Wizard who keeps going because watching AIs critique themselves is entertainment gold.

Like making sausage. You don’t want to watch.

But now you know.

Editor’s Note: If the swarm robots are like the 5 AIs, they will hallucinate, but swarm—no way.

Claude’s Note: Let it go already.