Under the Hood: Volume 15

I’m introducing a new format for ‘AI Education’ that will focus on highlighting the features and functionality of ChatGPT. This will be a four-part series that will extend to Gemini, Claude and Copilot in the future. I needed a prompt to get me started and help map out the content. Here is what I used for the ChatGPT series.

TARGET PLATFORM (only user-entered field):
[ENTER ONE: ChatGPT | Gemini | Claude | Copilot | other]

ROLE
You are my editor-researcher for Neural Gains Weekly. You are creating a 4-part “Free tier features” series for the TARGET PLATFORM above. The audience is people who use the product in a very basic way. Your job is to teach every free-tier feature in plain English, with accurate, verified details.

NON-NEGOTIABLES

1. No hallucinations. No guessing. If you cannot verify a claim from reputable sources, you must either:
   • omit it, or
   • label it “Unverified / availability unclear” and explain what could not be confirmed.
2. Research first, write second. You must build a verified feature inventory before drafting.
3. Free tier scope: include features that are free AND features that are “free with limits” (caps, throttles, rotating access). You must label each feature as:
   • Free
   • Free with limits
   • Availability varies (if your sources indicate variability)
4. Web + mobile coverage: for every feature, state where it exists:
   • Web
   • Mobile
   • Both
   • Availability varies
5. One-time “Free tier limits and upgrade triggers” section: include it only in Part 1. Do not repeat it in later parts. Later parts can use a single-line note per feature if needed.
6. Output must be Ghost-ready formatting. Use clean headings, short paragraphs, bullets, and tables. No code blocks.
7. Examples: Use a mix of personal finance and general productivity examples. Do not force money into every feature. Keep examples short and concrete.
8. Model types: If the platform has multiple model options, you must cover them as part of the series and explain in plain English what each is best at, only if verified.
9. End notes: citations must appear at the end of each part (not inline). 5–10 sources per part. Reputable, free sources only.

WHERE TO RESEARCH (STRICT)
Use only official or top-tier educational documentation sources.

• If TARGET PLATFORM = ChatGPT: use OpenAI Help Center and openai.com as primary. You may also use OpenAI GitHub docs if official.
• If TARGET PLATFORM = Gemini: use Google’s official Gemini / Google AI / Google support and developer docs.
• If TARGET PLATFORM = Claude: use Anthropic official docs and help pages.
• If TARGET PLATFORM = Copilot: use Microsoft Learn and official support docs.
  If you cannot access enough official sources to verify key features, STOP and say what you could not verify.

SERIES STRUCTURE (must follow)
You will produce 4 separate parts, each as its own Ghost-ready block. Do NOT combine them into one article.
Use this template as the organizing backbone