I asked ChatGPT to rewrite some of my glossary entries in the style of Enid Blyton. I said, “Please provide a rewrite of the following glossary entries, but in the upbeat style of Enid Blyton’s Adventure Series featuring Jack, Philip, Dinah and Lucy-Ann, together with Jack’s intelligent and talkative parrot Kiki.” Always ready for a new adventure, and without hesitation it replied, “Certainly! Here’s a cheerful and adventurous rewrite of the new glossary entries in the style of Enid Blyton’s Adventure Series, featuring Jack, Philip, Dinah, Lucy-Ann, and the ever-clever Kiki the parrot.”
Large Language Model (LLM):
Original: A type of AI computer program that is trained on vast amounts of text data to process and generate human language, capable of performing a wide range of language tasks. These models are typically trained on datasets containing hundreds of gigabytes to terabytes of text (i.e., up to one million books) sourced from diverse materials such as books, websites, articles, and social media. The scale of the training data enables LLMs to train on a wide array of linguistic patterns and contexts.
Enid Blyton Version: “I say, this Large Language Model is like having a library full of books in one place!” exclaimed Jack, his eyes wide with excitement. “It’s an AI that’s read millions of pages from all sorts of books, articles, and even things people say online!”
Philip nodded, always up for a new discovery. “Yes, and because it’s trained on so much, it can talk about almost anything! Imagine, it’s like having a clever friend who’s read every book in the world, and who can tell you about anything you ask!”
“And the best part,” added Lucy-Ann, “is that it understands all kinds of different ways people talk, so it can help us with all sorts of things—from writing stories to solving puzzles!”
Kiki flapped her wings. “Clever friend! Clever friend! Books and puzzles, books and puzzles!” The children laughed, delighted by the idea of an AI with the knowledge of a thousand libraries.
LLM:
Original: Short for Large Language Model, a specific type of AI model used for processing and generating human language.
Enid Blyton Version: “LLM—what a funny little name!” said Dinah with a grin. “It stands for Large Language Model, like the one we were just talking about.”
Jack nodded. “It’s a bit like a nickname, isn’t it? Just a shorter way to talk about that really clever AI that knows so much about language.”
Philip added, “And it’s the LLM that helps us whenever we need to understand something or need a bit of help with our writing. Handy, isn’t it?”
“Handy, handy! LLM, LLM!” squawked Kiki, hopping from foot to foot. The children giggled, happy to have such a clever new friend on their adventures.
Machine Learning (ML):
Original: Some people reserve the word “learning” for the processes by which living beings develop knowledge, understanding, and skills in complex environments. But the term “machine learning” has currency as a subset of AI that involves training algorithms to make predictions or decisions based on data, enabling systems to improve performance over time without explicit programming. Like “artificial intelligence,” the term “machine intelligence” serves to define a field of research and development, irrespective of the extent to which we think the anthropomorphic term “learning” applies.
Enid Blyton Version: “Machine Learning sounds like something out of one of our adventures!” exclaimed Jack. “It’s when a computer or an AI gets better at doing something all on its own, just by looking at lots of data.”
Lucy-Ann tilted her head thoughtfully. “So, it’s like when we learn to do something by practicing, like riding a bike or finding our way through a maze—except it’s a machine that’s doing the learning!”
“Exactly!” said Philip, always ready to explain. “It’s like the AI is teaching itself, getting smarter and quicker as it goes, without anyone needing to tell it exactly what to do.”
Kiki, ever the mimic, squawked, “Learning, learning! Clever machine, clever machine!” The children couldn’t help but smile at the thought of an AI learning new tricks, just like they did.
Matrix:
Original: A mathematical structure composed of rows and columns, used in AI to represent data, weights, and other parameters for computations.
Enid Blyton Version: “Matrix!” cried Dinah, looking at a grid on a piece of paper. “It sounds like a secret code or a treasure map!”
Jack peered over her shoulder. “It’s actually a grid of numbers, like rows and columns, that the AI uses to figure things out. It’s how the computer keeps track of all the bits of information it needs to do its job.”
Lucy-Ann’s eyes lit up. “So, it’s like the grid we use to find the buried treasure in a game of battleships, only this one helps the AI to solve problems!”
“Grid! Grid! Find the treasure!” chirped Kiki, flapping her wings. The children laughed, imagining the AI navigating its own grid to uncover hidden secrets.
Memory:
Original: Some people prefer to reserve the word “memory” for the processes by which living beings recall experiences. In AI, “memory” is the ability of a system to retain and utilize information from past interactions or data processing to improve performance on tasks.
Enid Blyton Version: “Memory in an AI? How fascinating!” said Jack, tapping his head. “It’s a bit like how we remember things that happened to us, only in a machine, it’s remembering all the information it’s been given.”
Philip grinned. “So, it’s like the AI has its own diary, where it keeps track of everything it’s learned and can look back at it whenever it needs to do something better!”
Lucy-Ann giggled. “Just like how we remember the best places to hide or the best way to climb a tree! The AI remembers too, and that’s how it keeps getting smarter.”
“Memory, memory! Diary of a machine, diary of a machine!” squawked Kiki, making the children laugh even more.
Model:
Original: A structured representation of a process or phenomenon, often implemented as algorithms and data structures, used in AI to simulate tasks and make predictions. In the context of LLM development, the term “model” can refer to different aspects: it may describe the entire trained system capable of processing and generating language; it can also refer to the underlying architecture (like a neural network) that defines how data is processed; and sometimes, the term is used to describe the specific configurations or versions of the system that are being trained or fine-tuned to improve performance on certain tasks. Thus, “training the model” can mean refining the algorithms, adjusting parameters, or enhancing specific capabilities within the larger system.
Enid Blyton Version: “A model, in AI? How jolly interesting!” said Jack. “It’s like a plan or a map that helps the AI figure out what it needs to do.”
Philip nodded. “Yes, it’s the way the AI is put together—like a blueprint for a building or a design for a new gadget. The model shows the AI how to take in information and make sense of it.”
“And when they talk about ‘training the model,’” added Lucy-Ann, “it’s like teaching it to be even better—just like how we train for a big adventure, getting ready for anything that comes our way!”
Kiki flapped her wings. “Training! Training! Ready for adventure! Ready for adventure!” The children chuckled, feeling as if they were right alongside their AI friend, preparing for the next big mystery.
Multi-agent system:
Original: Multiple autonomous agents interact or work together to achieve individual or shared goals. These agents can be software entities, avatars, robots, or even humans, and they operate in a common environment that coordinates their actions, shares information, and manages competition between them.
Enid Blyton Version: “I say, a multi-agent system sounds like one of our adventure teams!” exclaimed Jack. “It’s when different agents—like robots, avatars, or even people—work together to get something done.”
Philip grinned. “It’s just like when we all team up on one of our adventures, each of us playing our part, sharing information, and making sure we get the job done together!”
“And just like us,” added Dinah, “these agents can be working on their own tasks, but in the end, it all comes together in the same place, like a big, exciting puzzle!”
“Teamwork! Teamwork! Polly wants a cracker!” squawked Kiki, making everyone laugh. The children felt a surge of excitement, ready to face any challenge, just like a team of clever agents on their next big adventure.
Note
On request, ChatGPT provided the feature image with the caption: “Here is the updated adventure scene with a close-up of the mouth of a cave in the foreground, adding a sense of mystery and exploration. The illustration maintains the vibrant and nostalgic feel of a 1950’s children’s book, with the distant island still in view and the promise of discovery all around.”
Discover more from Reflections on Technology, Media & Culture
Subscribe to get the latest posts sent to your email.
Really enjoyed these, how do you intend to use them? Slip a few into your glossary?
Yes, as a short alternative glossary in the appendix. It’s too easy! You could fill a whole book with this material … and with pictures.