ChatGPT - now what?
The hype is down...what are the results?
Hi sharkies,
how often did you hear the following 👇
“ChatGPT is going to change everything, you have no idea”
I know that I have…and it got me thinking…
Now that the initial hype is down, now what?
In this edition we will do a quick ChatGPT introduction, how to use ChatGPT to your benefit and how some companies already built ecosystem of apps around ChatGPT.
Before we dive into another deep dive 👇
aaannndd 👇
Ok, now we can continue!
What is ChatGPT anyway?
First things first, ChatGPT in its essence is not revolutionary.
Surprised? 😮😮😮
How often did you see on a website a chatbot that is there to help you with your inquiry?
I’m not saying that ChatGPT is the same as those chatbots, but they do fall into the NLP category.
Natural language processing (NLP) refers to the branch of computer science—and more specifically, the branch of artificial intelligence or AI—concerned with giving computers the ability to understand text and spoken words in much the same way human beings can.
So, you might be wondering now…why ChatGPT is so special?
Because of its training and the amount of data it has been trained on.
Training process
ChatGPT runs on a language model architecture created by OpenAI called the Generative Pre-trained Transformer (GPT).
Generative AI models of this type are trained on vast amounts of information from the internet, including websites, books, news articles, and more.
Just a small refference 🔦
GPT-4 version is currently active, but OpenAI didn’t disclose numbers, so we can just guess on the number of parameters, the size of training dataset etc.
Greater numbers mean better results in terms of human-like conversation.
AIs pre-train using two principle approaches: supervised and non-supervised.
For most AI projects up until the current crop of generative AI systems like ChatGPT, the supervised approach was used.
Supervised training approach
In a supervised training approach, the overall model is trained to learn a mapping function that can map inputs to outputs accurately. This process is often used in supervised learning tasks, such as classification, regression, and sequence labeling.
For example, an AI could be trained on a dataset of customer service conversations, where the user's questions and complaints are labeled with the appropriate responses from the customer service representative. To train the AI, questions like "How can I reset my password?" would be provided as user input, and answers like "You can reset your password by visiting the account settings page on our website and following the prompts." would be provided as output.
Obviously, this type of training approach would work for a limited scope application. The majority of customer service chatbots you can see on websites today are trained using this method, and their limitations are obvious.
Non-supervised training approach
Non-supervised pre-training is the process by which a model is trained on data where no specific output is associated with each input. Instead, the model is trained to learn the underlying structure and patterns in the input data without any specific task in mind.
In the context of language modeling, non-supervised pre-training can be used to train a model to understand the syntax and semantics of natural language, so that it can generate coherent and meaningful text in a conversational context.
It's here where ChatGPT's apparently limitless knowledge becomes possible. Because the developers don't need to know the outputs that come from the inputs, all they have to do is dump more and more information into the ChatGPT pre-training mechanism, which is called transformer-base language modeling.
Now, I will stop here with the technical explanation because we would go into too much detail.
In case you are super interested in this topic, you can visit the following websites that can give you more technical context 👇
(this is more light version)
How ChatGPT Works: The Model Behind The Bot
(this is heavy technical version)
Short conclusion
ChatGPT is special because they managed to train the model on the largest dataset ever. The model training itself was different than before, which enabled users to use ChatGPT now just for a specific purpose, but for many, many different purposes.
But…ChatGPT is (yet) nothing more than a large generative AI language model. This is still not AI in terms of creating new ideas.
How to use ChatGPT to your benefit?
Now, this is the real question!
Long story short, you will have to learn several new skills as ChatGPT evolves, but the most important skill you have to learn is prompt engineering.
If you are not into IT, most likely you never heard for this.
Prompt Engineering
In simplest terms, prompt engineering (PE) is the process of communicating effectively with an AI to achieve desired results.
ChatGPT is a great tool, but if you don’t know how to communicate with it…what’s the point?
Tweeting in January 2023, Andrej Karpathy, Tesla's former director of AI, stated that the "hottest new programming language is English."
You know that I never leave you without a guide, and this is the best beginner prompt engineering guide I could find 👇
In this guide you will find all the things you need to start using tools such as ChatGPT in the most effective way possible.
ChatGPT plugins
Companies recognized that instead of trying to beat ChatGPT, they could build up their services around it.
It makes sense because it took the OpenAI team years of model training and a lot of 💵 to get where they are, and it’s not reasonable to expect anyone except big players (Google, Meta) to compete with OpenAI and Microsoft.
At the moment, there are over 900 ChatGPT plugins available, but I guess that there are still a lot of them on the waiting list.
Full list you can find here 👇
For you more visual folks, here are few screenshots 👇
As it seems now, this is slowly turning into a plugin store, similar to the web plugins we have at the moment.
However, one plugin could “rule them all”.
For paid ChatGPT users, Microsoft enabled Bing (Microsoft search engine) and ChatGPT integration. This has the potential to change the way we search the internet. Instead of getting countless website links for your search, ChatGPT can summarize the information you seek.
This could be a game-changer.
Let us know in the comments how you use Chat GPT, and what kind of plugins you use!