Why all the chat about ChatGPT?

By now if you don't go one day with at least 5 posts on LinkedIn where someone has asked ChatGPT, a service from OpenAI, to write some song lyrics or some other parlour trick you're either not connected to anyone on LinkedIn or you're in a very interesting corner of the platform.

Why so much interest? this a fundamentally different way of training and using data for the purposes of AI. Traditionally you would take a load of your data, mark it all up, and use that data set to train your machine learning model or train a specific set of models for AI to use. OpenAI has unlocked the use of massive and general data sets to help make these sames models smarter, faster, and critically fill in gaps that arrive from data missing in unavailable.

As a simple test, and to get away from the "tell me a joke", "write a poem" approach I asked ChaptGPT to explain itself....

David: "write an interesting LinkedIn article about ChatGPT that highlights the use case for businesses and mentions the differences between traditional approaches to training AI models and how this will help banking and healthcare users"

ChatGPT: "Artificial Intelligence (AI) has become a vital tool for businesses in today's digital age. With the rise of natural language processing (NLP) and machine learning (ML), businesses are able to automate tasks, improve customer service, and make more informed decisions. One of the latest developments in AI is ChatGPT, a large language model developed by OpenAI. In this article, we will explore the use cases for ChatGPT in businesses and how it differs from traditional approaches to training AI models.

ChatGPT is a transformer-based model that has been trained on a massive dataset of text. This allows it to generate human-like text, making it ideal for tasks such as chatbots, language translation, and text generation. With ChatGPT, businesses can create chatbots that can understand and respond to customer inquiries with high accuracy. This can help to improve customer service and reduce the workload on human customer service representatives.

One of the key advantages of ChatGPT is that it is pre-trained on a large dataset of text, which means that it can be fine-tuned for specific tasks with relatively little data. This is a significant difference from traditional approaches to training AI models, which require large amounts of labeled data for each task. This makes ChatGPT ideal for businesses with limited data resources, as well as for businesses in industries that have strict regulations on data usage, such as banking and healthcare.

In the banking industry, ChatGPT can be used to improve customer service by creating chatbots that can assist customers with account information, transactions, and other banking-related tasks. It can also be used to automate compliance-related tasks, such as monitoring for suspicious activity and flagging potential fraud.

In the healthcare industry, ChatGPT can be used to assist doctors and nurses with tasks such as medical record keeping, appointment scheduling, and prescription ordering. It can also be used to improve patient communication, such as providing symptom checkers and answering common health-related questions.

In conclusion, ChatGPT is a powerful tool for businesses looking to improve customer service, automate tasks, and make more informed decisions. With its pre-training on a massive dataset of text and its ability to be fine-tuned with relatively little data, it has the potential to revolutionize the way businesses use AI, especially in industries such as banking and healthcare."

I think you'll agree that is an interesting result given the nature of the dataset in the background. Remember, general data that this system has drawn inferences and conclusions from.

Satya Nadella as always explains it well in his sit down with WSJ's Matt Murray at WEF in Davos:

Satya: "My data is the only thing that matters. All that's now questioned, you now have a large model which can be used as a retrieval engine on your data that's going to perform better than the sort thing that you trained on your data. I mean, just the intuition, just so that people have what this is, is suppose you train a model on, let's say, all the math formulas.

Yeah, it'll get pretty efficient in math. But let's say you train a model on all the math formulas

and all the human text there is on the web, it gets better at math.

And you say, why? How is that possible? Of course it's possible because when we went to school, it's not like we just learned math, we learned math, we learned history, we learned language. And so that emergent capability is a pretty different thing than like why we were even a couple of years ago as state of the art in terms of AI."