2023’s buzz word in tech is surely going to be ChatGPT, and it’s only January. Indeed since it’s initial release in November 2022, OpenAI‘s ChatBot has been making all kinds of headlines from describing how it would take over the world to generating song lyrics in the style of Nick Cave. As you might know, we’ve talked about AI before in terms of video game development when discussing how to create a Follow AI using a NavMesh in Unity. Now while that particular AI methodology was movement based, it still appears very innocent in comparison to ChatGPT, which is a sophisticated chatbot that can respond very intelligently and almost appear to have a personality. Almost.
In ChatGPT’s own words, here is how it would describe itself:
“ChatGPT is a large language model developed by OpenAI. It is trained on a massive dataset of text data, which allows it to generate human-like text. It is based on the GPT (Generative Pre-training Transformer) architecture and can be fine-tuned for a wide range of natural language processing tasks such as language translation, text summarization, and question answering. Additionally, it can be used for other tasks such as text completion, text generation and conversation.
One of the main advantages of ChatGPT is its ability to generate text that is indistinguishable from text written by humans. This makes it well-suited for tasks such as chatbot development, where the goal is to create a conversational agent that can hold a natural and coherent conversation. The model can be fine-tuned on specific conversational data, which allows it to generate responses that are tailored to the given context.
ChatGPT is also highly adaptable and can be fine-tuned on a wide range of datasets and languages, making it suitable for a broad range of natural language processing tasks. It can also be used in combination with other models, such as a speech-to-text model, to create a complete natural language understanding system.
Overall, ChatGPT is a powerful and versatile language model that can be used for a wide range of natural language processing tasks. It’s ability to generate human-like text and adapt to different languages and tasks make it a valuable tool for developers and researchers working in the field of natural language processing.”
That quoted text, generated by asking ChatGPT to summarize itself, is a pretty concise and clear. That’s because ChatGPT excels in parsing its dataset to spit out straight up definitions, although it can and will make mistakes. I asked it some questions about some local geography and municipal boundaries and while it knew about the places in question, it confused exactly where certain towns were located (on the mainland vs on an island). So it really depends where ChatGPT is getting it’s information from and how it puts it together, contextually. Knowing that, beware when asking ChatGPT to write an essay for you and make sure you proofread the hell out of it if you do.
Interested? Sign up and try ChatGPT for yourself at OpenAi