There has been a lot of hype around AI this year which seems to have come to a boil over the last few weeks with OpenAI allowing access to its ChatGPT AI, a modified version of OpenAI GPT3 (Generative Pretrained Transformer 3). Both use the Natural language model to process and understand human language.
I had to check it out for myself, and while I was impressed with its capabilities, it’s not a finished product, even if some of the hype suggests it is.
My experience of using chatGPT
To write this article, I sat down to use the ChatGPT for only the second time, with me using the first-time layout ideas for this article.
It would be good to start by asking GPT3 to introduce itself. This is how Open ChatGPT responded to that question. “As an artificial intelligence, I am a machine-based program designed to process and analyze information, and to provide responses to questions and prompts. I am trained on a large dataset of human language, and I use this training to generate human-like responses to queries and interactions. My primary function is to provide information and assistance to users, and I am here to help you with any questions or concerns you might have. Is there anything specific you would like to know or discuss?”
I then asked ChatGPT a more technical question about the basics of Boolean logic, and I was impressed with the answer it gave me. I couldn’t detect any errors in the response, and while I don’t teach computing, I think it would be an acceptable answer to an assignment.
Let’s see if it can tell us how to create a new page in Notion, which I tried to follow, and it wasn’t 100% accurate, but it was close enough for me to figure it out for myself. It could be because the data contained within the model is up to 2021. However, I don’t think creating a Notion page is one of the things that changed. I tried the same with making a note in Obsidian, and that answer wasn’t quite right either.
When I asked ChatGPT who won the English Premier League when I was working on laying out this article, it could not answer the question today, but on the day I wrote this article. The question was answered, suggesting that it either changed or I asked it slightly differently.
Conclusion
As I stated in my introduction, I’m impressed, and it’s indeed an immerging technology. Still, I don’t think people in the creative industries have to worry as the technology relies on what it learns from us. It can compare many examples stored in its database. It might even be able to see links between things humans have missed.
But it has no imagination. It cannot yet think of new ideas, dream about the future, and define new models to explain the world about us.
That isn’t to say we shouldn’t be concerned about AI in the future, and we need to start discussions now to ensure the technology isn’t abused.
You can improve the linking of your knowledge by using a personal knowledge management system such as a Zettelkasten. You can learn more in our introduction to Zettelkasten.