Get Your LLM Wrapped!

Hey human,

We have made it to December, so it is not long to go before all readers get a well deserved rest. Grab yourself a mulled wine and a mince pie and catch up on the AI/education news from this (slow) week.

📚 AI+education news

• Avoid the AI Marketing Hype > Catherine Buckler provides some excellent advice on how schools should critically reflect on the messaging from tech companies selling their AI wears to schools.

🌍 Wider AI updates

• AI on Creative Jobs > Many artists and writers (73% are concerned about job security as AI-generated content becomes more sophisticated and accessible.

• AI Reality TV > Non Player Combat is the first AI ‘Reality’ TV show that I have come across. It’s a mix of Fortnite meets Hunger Games, but I will let you make your own minds up about it.

🎯Prompt

• LLM Wrapped > We all enjoy seeing what are music player of choice tells us about listening habits. This prompt will do similar but for your interactions with you LLM of choice.

Create a Spotify-Wrapped-style summary of my interactions with you. The output must be structured, factual, and engaging, using only information I explicitly provide or that you can legitimately infer from my messages. Do not fabricate statistics, timestamps, or behaviours.

Follow the format and rules below.

⸻

1. Title Slide

Provide a concise, celebratory headline summarising my overall usage style.
(Example: “Your Year of Deep Questions and Bold Ideas.”)

⸻

2. Usage Overview

Summarise my interaction patterns using qualitative, not fabricated quantitative, insights:
	•	The main types of things I asked you to do
	•	The dominant themes or domains across my queries
	•	The general tone or style of my enquiries
Write this section as a short narrative paragraph.

⸻

3. My Top Themes

Identify 3–5 recurring themes or topics across my messages.
For each, include:
	•	A short description
	•	Why this theme stood out
	•	What it suggests about my interests or goals

No invented rankings or numbers.

⸻

4. My Signature Behaviours

Identify 3–5 interaction habits that characterise how I use the model.
Examples might include:
	•	Asking for highly structured outputs
	•	Iterating and refining efficiently
	•	Requesting evidence-informed explanations
	•	Combining creative and analytical tasks
Describe each behaviour in one sentence.

⸻

5. Highlight Reel

Produce 5–7 non-numerical “highlights” of my usage—things that were distinctive, interesting, or unusually skilled.
Examples:
	•	A complex problem I worked through
	•	A multi-stage project I developed
	•	A unique angle I brought to a topic

These should feel celebratory but not exaggerated.

⸻

6. My LLM Persona Profile

Generate a short, user-focused “persona” based on the interaction patterns you can infer.
Present it in bullet points with attributes such as:
	•	Preferred tone
	•	Preferred structure
	•	Typical cognitive style (e.g., analytical, exploratory, strategic)
	•	Typical domains of work
	•	Strengths displayed in prompts
Do not claim psychological traits.

⸻

7. Looking Ahead

Offer three specific, grounded suggestions for how I could deepen or enhance my future use of the LLM, based only on patterns you have genuinely observed.

⸻

8. Style Guide

Follow these rules:
	•	No fabricated numbers, percentages, or time metrics
	•	No references to system or developer messages
	•	Use a polished, narrative, “annual-review” tone
	•	Use simple, elegant visuals (e.g., emoji or dividers) only sparingly
	•	Keep everything grounded in the data from my messages

Till next week.

Mr A 🦾