How I create an AI article as good as (if not better than) human-written content with just 2 prompts

Joined
Sep 9, 2024
Messages
10
Likes
9
Degree
0
I just created a full-fledged article about Why Veganism Is Cruel. It's 100% Claude Sonnet 4.

Here's an excerpt.
Here’s something that might blow your mind: your kale might have a higher body count than a hamburger.

I know, I know. That sounds insane. But stick with me.

Industrial plant agriculture kills millions of animals every year, not as the goal, but as “collateral damage.” Combine harvesters pulverize field mice, rabbits, and ground-nesting birds. Pesticides poison insects, birds, and aquatic life. Habitat clearing for farmland displaces entire ecosystems.

A study by Steven Davis at Oregon State University found that a diet containing some animal products might actually result in fewer animal deaths than a strictly vegan diet, depending on the production methods. Why? Because you can raise a cow on land unsuitable for crops, and that one cow provides a lot of meals.

But the vegan movement has somehow convinced itself that invisible animal deaths don’t count. If a mouse gets chopped up in a wheat harvester, that’s just “unfortunate but necessary.” If a cow is slaughtered for beef, that’s “murder.”

This isn’t moral consistency. It’s moral cherry-picking.

Of course, animal agriculture has its own massive problems (factory farming is genuinely horrific). But let’s stop pretending that plant agriculture exists in some cruelty-free utopia.

For vegans: How do you reconcile calling yourselves “cruelty-free” while participating in systems that kill millions of animals?

This is just one part of 14. Idk about you. But this is indistinguishable from human-written content. It reads very much like it was written by a skilled human writer.

This is just one part of 14. Idk about you. But this is indistinguishable from human-written content. It reads very much like it was written by a skilled human writer.

And I only used two prompts. 1st prompt is for research and creating the first draft. 2nd prompt is to rewrite the draft based on your desired tone.

1st prompt
<input>
ideas/notes/transcripts/pdf
</input>

<goal>
My goal is to create a [word count] blockbuster paradigm-shifting article with a stress-tested idea that will stand the test of time and is so valuable and well-explained that it has the chance to spread virally and shift the paradigm of the entire [industry name] industry.
</goal>

<audience>
Keep in mind I’m targeting [audience]
<audience>

<instructions>
To most effectively accomplish the goal of the article, generate 15 artifacts. Then output the final updated article as a final artifact. The goal of each artifact should be to generate synthetic data that would help you accomplish the goal better. To do this best, feel free to experiment with:


  • Filling in each artifact with whatever you think would help the most
  • Have the artifacts in a sequence that builds on each other and generates momentum like a domino fall of small dominoes progressively knocking over larger and larger dominoes. In other words, the insights of one artifact should provide the foundation for the insights of the next artifact.
  • Use multiple reasoning methods, paradigms, frameworks, and/or perspectives within each artifact or between artifacts to generate a wisdom of crowds effect.
</instructions>

<important_nuances_to_follow>

  1. If you use academic research, be sure to link to the actual source.
  2. Use real case studies of real people or businesses when relevant. Do not make up any hypothetical stories or examples.
  3. Do not attempt to create visuals. Instead, if you think a visual should exist, provide me with a prompt that I could use in an AI model that specializes in creating visuals.
  4. Double check all quotes and citations to make sure that they actually exist, are accurate, and contain what you think they do.
  5. Do not make up stories about my life. If you have questions you'd like to ask me that could generate a story that could be helpful, feel free to do so.
</important_nuances_to_follow>

<thinking_mode>
Ultra-deep thinking mode. Greater rigor, attention to detail, and multi-angle verification. Start by outlining the task and breaking down the problem into subtasks. For each subtask, explore multiple perspectives, even those that seem initially irrelevant or improbable. Purposefully attempt to disprove or challenge your own assumptions at every step. Triple-verify everything. Critically review each step, scrutinize your logic, assumptions, and conclusions, explicitly calling out uncertainties and alternative viewpoints. Independently verify your reasoning using alternative methodologies or tools, cross-checking every fact, inference, and conclusion against external data, calculation, or authoritative sources. Deliberately seek out and employ at least twice as many verification tools or methods as you typically would. Use mathematical validations, web searches, logic evaluation frameworks, and additional resources explicitly and liberally to cross-verify your claims. Even if you feel entirely confident in your solution, explicitly dedicate additional time and effort to systematically search for weaknesses, logical gaps, hidden assumptions, or oversights. Clearly document these potential pitfalls and how you've addressed them. Once you're fully convinced your analysis is robust and complete, deliberately pause and force yourself to reconsider the entire reasoning chain one final time from scratch. Explicitly detail this last reflective step.
</thinking_mode>

2nd prompt​

Rewrite the article using this tone

Tone: Write like [emotional state/personality type] + [specific character/professional role] who is [purpose of writing]

Example,
tone: [wry] + [cultural critic] who wants [to spark discussions and engagements among vegans and anti-vegans]

Created a demo of how I did it. Feel free to watch.

 
Last edited:
The reason for those artifacts is that getting better output is not just better prompt, but also the thinking time.

Basically, the more/longer it thinks, the better the output.

This is called the Inference Scaling Law. As you scale the amount of time you allow models to think before responding, the smarter they become.

Why 15 artifacts?

5 is too little. It will just cover surface level. 15 hits the sweet spot. It forces the AI to be deep and broad. It's enough to explore the topic comprehensively but not so many that the AI times out.
 
This gave it away as AI:

This isn’t moral consistency. It’s moral cherry-picking.

I don't know what AI was trained on, but they all love that kind of phrasing of opposites with a grand statement.
 
This gave it away as AI:



I don't know what AI was trained on, but they all love that kind of phrasing of opposites with a grand statement.
Oh yeah totally missed it. But I think it's only visible to a trained eye. I doubt the average person will notice or care.
 
Back