What does a conversation with ChatGPT look like? I test it out below and add some short notes and thoughts on what it might mean for the future.
I ran a simple chat session with ChatGPT, just to get an idea of how good it is at understanding questions and returning understandable, useful answers in ordinary language. Everything in the conversation below with a yellow background was typed by me, the rest is from ChatGPT. Where my typing is an aside for you, my readers, and not sent to ChatGPT, I’ve put it in braces {}.

The conversation is quite long, but I hope it will interest you and perhaps cause you to raise questions in your own mind. If you want to discuss any thoughts on AI in general (and ChatGPT in particular) just leave them in the comments section below. Thanks.
The conversation
Chris – Hi, my name is Chris and I’d like you to introduce yourself; please don’t be nervous, just give me an informal introduction.
ChatGPT – Hi Chris! I’m ChatGPT, a language model designed to communicate with humans and assist with various tasks. Think of me as a robot that can talk with you through text. I’m always eager to learn new things and help you in any way I can, so feel free to ask me anything!
Chris – Thanks! Good to meet you. OK, I’ve got the idea that you are robot-like. There are a couple of things I’d like to know. Why are you called ‘ChatGPT’? And what is a language model?
ChatGPT – Good to meet you too, Chris!
ChatGPT is named after the GPT (Generative Pre-trained Transformer) language model architecture, which is used to build me. GPT is a type of neural network that is trained on massive amounts of text data, allowing it to generate coherent and natural-sounding text. The “Chat” part of my name refers to my ability to chat with humans like you, using text as our mode of communication.
Speaking of language models, a language model is a type of artificial intelligence that can analyze and generate human language. Essentially, a language model is a computer program that is trained on large amounts of text data, allowing it to learn the rules and patterns of language. Once trained, a language model can perform a wide variety of tasks, such as generating text, translating between languages, or answering questions. I’m a particularly advanced type of language model, based on the GPT-3.5 architecture, which allows me to handle a wide range of tasks and communicate with people like you in a natural and engaging way.
Chris – I must congratulate you, your replies are certainly engaging and natural. Do you mind if I ask some more questions? This is an interesting conversation.
ChatGPT – Thank you, Chris! I’m always happy to answer questions and have interesting conversations, so please feel free to ask me anything you’d like.
{I thought I’d introduce some typos next, to see if ChatGPT would struggle to parse the text. But it didn’t bat an electronic eyelid, answering quickly and correctly.}
Chris – What sort of toxt deta were you trainers with? Do you know where it came from? A few examples of sources would be helpful.
ChatGPT – I was trained on a diverse set of text data, ranging from news articles and academic papers to social media posts and online discussions. The training data comes from various sources and domains, covering a wide range of topics and writing styles to help me understand and generate text in a natural and diverse manner.
Some examples of sources that were used to train me include:
- Wikipedia: A large online encyclopedia with articles on almost every topic imaginable.
- Books: A wide range of books on various topics, from fiction to non-fiction.
- Web pages: Crawled from the internet, including blog posts, news articles, and online discussions.
- Social media: Posts from popular social media platforms like Twitter and Reddit.
- Scientific journals: Academic papers from various scientific disciplines.
- News sources: Articles from major news outlets like CNN and the New York Times.
The training data was collected from various sources and pre-processed to remove irrelevant or duplicated content, ensuring that the data was of high quality and diverse enough to train me to generate human-like text.
Chris – In that case let’s try to draw on some of the material you’ve seen from those sources. I live in an English town called Cirencester. It’s in a region called The Cotswolds. Please write a short essay about Cirencester for me.
ChatGPT – Sure, I’d be happy to write a short essay about Cirencester for you!
Cirencester is a charming town located in the heart of the Cotswolds, a picturesque region known for its rolling hills, quaint villages, and traditional architecture. With a history dating back to Roman times, Cirencester is home to a wealth of historical and cultural attractions that draw visitors from around the world.
One of the town’s most notable landmarks is the Cirencester Parish Church of St. John Baptist, a stunning example of medieval architecture that has been standing for over 900 years. The church is home to a variety of beautiful stained-glass windows, intricate stone carvings, and historical artifacts that provide a glimpse into the town’s rich history.
Another must-visit attraction in Cirencester is the Corinium Museum, which houses an impressive collection of Roman artifacts and exhibits that showcase the town’s Roman heritage. Visitors can explore the museum’s extensive collection of mosaics, coins, and pottery, as well as learn about the daily lives of the Roman citizens who once called Cirencester home.
In addition to its historical attractions, Cirencester is also known for its natural beauty and outdoor activities. The town is surrounded by idyllic countryside, with plenty of opportunities for hiking, cycling, and exploring the Cotswold Way, a 100-mile walking trail that winds through the region’s stunning landscapes.
Whether you’re interested in history, culture, or outdoor activities, Cirencester has something to offer. With its unique blend of ancient and modern, this charming town is a must-visit destination for anyone looking to experience the best of the Cotswolds.
Chris – That was really good! Do you write poetry as well? How about a short poem (three or four verses) about Cirencester?
ChatGPT – Sure, I can certainly give it a try! Here’s a short poem about Cirencester:
In the heart of the Cotswolds lies a town,
With Roman roots and ancient renown.
Cirencester, where history meets charm,
And natural beauty keeps us calm.
Streets lined with stone and buildings grand,
A parish church that still does stand.
Corinium’s past in the museum’s keep,
A town that makes our hearts leap.
Cirencester, oh, how you inspire,
In every season, by every fire.
A gem in the Cotswolds, a treasure to see,
A place where our hearts are free.
Chris – Great! Here’s something a bit harder, can you do a haiku on Cirencester?
ChatGPT – Sure, here’s a haiku about Cirencester:
Roman walls stand tall,
Cirencester’s ancient embrace,
Cotswold’s heart of stone.
Chris – Great again, and hard to do. I have one last question for you; how should I solve a quadratic?
ChatGPT – To solve a quadratic equation, there are several methods, but one of the most commonly used is the quadratic formula. The quadratic formula states that for an equation of the form ax^2 + bx + c = 0, the solution for x is given by:
x = (-b ± sqrt(b^2 – 4ac)) / 2a
where “sqrt” denotes the square root function.
To use this formula, you simply plug in the values of a, b, and c from your quadratic equation into the formula, and then simplify the expression to find the solutions for x.
It’s worth noting that there are other methods for solving quadratic equations, such as factoring and completing the square, but the quadratic formula is generally considered the most reliable and efficient method, especially for more complex equations.
Chris – Thanks! Enjoy your time chatting to some other people. By the way, how many people can you converse with at the same time? I can only manage four or five.
ChatGPT – As an AI language model, I can theoretically converse with an unlimited number of people simultaneously, as long as the incoming messages are not too frequent and I have enough computing resources to process them all in a timely manner. However, I can only provide individualized responses to each person based on the messages they send me, so the more people involved in the conversation, the more difficult it may be to keep track of multiple threads of conversation at once.
Some thoughts on what you’ve just read
I think you’ll agree that ChatGPT is pretty good at giving understandable and useful responses in English to a range of questions. It’s pretty much the sort of conversation you’d expect from a very well-read human. It seems most unlikely that today’s AI systems, including ChatGPT, actually ‘think’ in anything like a human way. It manipulates language very cleverly in ways that are useful and seem human-like. But these AI systems don’t understand meaning, they don’t have feelings or emotions, they don’t have thoughts or opinions.
Might they have the earliest glimmerings of these things? There’s no evidence that they do. But the teams developing such systems have sometimes expressed caution. The danger here is that any much more advanced future system that did have a spark of self-awareness might arise without us even knowing about it, and if it was capable of modifying itself it could become exponentially more clever. Even a few days of exponential improvement could result in a more powerful mind than any human. After that, the AI could outwit us with ease and make decisions that we might not like. The difficulty is that the time available for us to intervene would be a much smaller window than we might suppose or plan for.
In the near future (likely starting in months, not years) a welcome development will improve those systems that already use speech recognition and/or speech generation. Prepare for much better and more natural interaction with devices like Alexa, your satnav, book readers, aids for the blind and much more. No more need to remember the right words and phrases to use, fewer misunderstood commands, the ability to ask for more or different information.
Expect help with programming, planning astronomical observing sessions, writing up research, having your articles proof read and copy edited. Perhaps soon you’ll be able to throw away your keyboard! Laptops might start to look like tablets. Bedtime? Ask for a story.
ChatGPT can do that already…
Chris – Write me a children’s bedtime story about a dog, a mouse, an oak tree and an ant hill.
ChatGPT – Once upon a time, there was a friendly dog named Max. Max loved exploring the forest near his home and making new friends. One day, while he was wandering through the woods, he met a tiny mouse named Mimi. Max and Mimi quickly became best friends and spent hours playing together.
As they were playing, Max noticed a beautiful oak tree nearby. The tree was enormous, with branches stretching high into the sky. Max and Mimi decided to climb up the tree and see what they could find. When they reached the top, they discovered a bustling ant hill. Thousands of tiny ants were working together to gather food and build their home.
Max and Mimi watched in awe as the ants worked tirelessly. They were amazed at how much the little ants could accomplish when they worked together. Max and Mimi decided to help the ants by gathering food for them. Max used his strong nose to sniff out berries, while Mimi used her quick reflexes to catch insects.
The ants were overjoyed to receive Max and Mimi’s help. They thanked the two friends by giving them a special gift – a magical acorn that would grant them one wish. Max and Mimi thought long and hard about what they should wish for.
Finally, Max said, “I wish for all the animals in the forest to be friends, just like us.” Mimi nodded in agreement. She added, “And I wish for the oak tree to grow even taller and stronger, so it can provide shelter for even more creatures.”
Their wish was granted, and from that day forward, all the animals in the forest were friends. The oak tree grew taller and stronger, and even more animals found shelter in its branches. Max and Mimi continued to play together and explore the forest, knowing that their wish had made a difference in the lives of so many creatures.
As the sun began to set, Max and Mimi said goodnight to their new friends and curled up together under the oak tree. The stars twinkled above them, and the gentle breeze rustled the leaves. Max and Mimi felt grateful for the adventure they had shared and fell asleep with happy hearts, dreaming of all the fun they would have tomorrow.