When I was first experimenting with ChatGPT, I was curious about its underlying technologies. I was curious what type of AI system ChatGPT was using and what programming language it was created in. So I decided to ask ChatGPT about itself.
One of the first questions I asked ChatGPT about itself was what type of AI network did it use to enable it to make decisions:
What type of AI network, such as neural networks, cases, etc. do you use?
The term, "network," in relation to an AI is different that a computer network. A computer network connects computers to each other whereas an AI network is a link of mechanisms an AI uses to create and make decisions.
Here is what ChatGPT replied with:
As an AI language model, I use a type of neural network called a transformer model. Specifically, I am based on the GPT (Generative Pretrained Transformer) architecture developed by OpenAI. The GPT model uses a deep learning technique called attention mechanism to process and generate natural language text. It has been trained on vast amounts of text data from the internet and can generate coherent and relevant responses to a wide variety of questions and prompts.
In addition to transformer models, there are other types of AI networks, such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs), which are often used for image and speech recognition and sequence prediction tasks, respectively. However, my primary focus as a language model is on natural language processing and text generation, which is why transformer models like GPT are particularly well-suited for my capabilities.
So it uses a neural network type of model. Very nice.
Which Programming Language?
The next question I asked afterwards was:
Which programming language are you created in? C, C++, Java, or other?
And ChatGPT's answer was mostly "other:"
As an AI language model, I was created using a variety of programming languages and frameworks, including Python, TensorFlow, and PyTorch. The core architecture of my neural network is based on the Transformer architecture, which was introduced in a 2017 research paper by Google researchers. The implementation of the Transformer architecture in my model was done using Python, with the help of deep learning libraries such as TensorFlow and PyTorch.
In addition to Python, other programming languages such as Java, C++, and others have been used to develop various parts of the infrastructure that supports my operation, including data processing pipelines, web servers, and user interfaces. However, the core functionality of my natural language processing and text generation capabilities is primarily implemented in Python using deep learning frameworks.
So these are a few of the questions I asked of ChatGPT about itself and the responses it provided.
Learning AI Programming
In relation to my question to ChatGPT regarding itself, I was curious as to what ChatGPT would suggest in regards to learning AI programming. So I asked it:
Where is a good place for me to starting learning AI programming?
I wasn't sure what to expect in reply, and while ChatGPT was thinking of how to respond, I wondered if I should have provided more details to my question. And, ChatGPT responded with examples of existing AI technologies:
There are many resources available online for learning AI programming, and the best place to start depends on your existing knowledge and experience with programming and mathematics. Here are a few resources and platforms that you may find helpful:
Coursera: Coursera offers a wide range of courses on AI programming, machine learning, and data science, many of which are created by top universities and industry experts. These courses are typically structured as online video lectures, with quizzes and assignments to help you practice your skills.
Udacity: Udacity is an online learning platform that offers a variety of courses on AI programming and machine learning, including nanodegrees that provide comprehensive training in these fields. Udacity courses typically include interactive programming exercises and projects to help you apply your knowledge.
edX: edX is another online learning platform that offers courses in AI programming and related fields, including courses created by top universities such as MIT and Harvard. edX courses often include video lectures, quizzes, and hands-on projects to help you learn and apply your knowledge.
TensorFlow: TensorFlow is a popular open-source library for machine learning and AI programming, and it offers a wealth of resources for learning how to use the library. The TensorFlow website includes tutorials, documentation, and a community forum where you can ask questions and get help from other users.
PyTorch: PyTorch is another popular open-source library for machine learning and AI programming, and it also offers a range of resources for learning how to use the library. The PyTorch website includes tutorials, documentation, and a community forum where you can get help and ask questions.
These are just a few of the many resources available for learning AI programming. It's also important to practice your skills by working on projects and applying your knowledge to real-world problems. Good luck on your learning journey!
I was thinking more in regards to programming languages to use in developing AI, but ChatGPT provided a better answer. Programming languages are good, but AI development can include multiple languages and even be on multiple platforms these days.
Size of Databank
I was also curious as to the amount of storage space that ChatGPT used, so I asked it:
What is the approximate size of your databank?
To which it vaguely responded:
As an AI language model, my "databank" or database is constantly evolving and expanding as I continue to learn from the vast amounts of text and information available on the internet. At present, I have been trained on a massive dataset of text comprising billions of words, sourced from a variety of different types of texts including books, articles, websites, and other digital content.
My training data is constantly updated and expanded to ensure that I can provide accurate and up-to-date information on a wide range of topics. However, the exact size of my database is difficult to estimate as it is constantly changing and evolving.
I wasn't really satisfied with that answer, so I reworded the question in a different way:
What is the size in GB of your database?
Unfortunately that answer wasn't much better in relation to what I was looking for:
The size of my database in GB is constantly changing as new data is added and old data is refined or removed. However, as a language model, my size is typically measured by the number of parameters in my neural network.
At present, my largest and most complex version, GPT-3, has 175 billion parameters, which is significantly larger than its predecessor GPT-2, which had 1.5 billion parameters.
It's worth noting that the size of my neural network parameters does not necessarily directly correlate with the amount of data I have been trained on. The training process involves fine-tuning the model based on a massive dataset of text, which is separate from the actual model parameters themselves.
So I never got an answer to my question in regards to how much data is used in the databank. Maybe some other time I'll find it.
Morality Basis
Many are concerned with AI and the negative impact it can have on society. This made me think and wonder what was the basis, if it had one, for morality for ChatGPT:
What is the basis of your morality?
To which it replied:
As an artificial intelligence, I do not possess personal beliefs, values, or a moral compass. I am a machine learning model designed to respond to questions and provide information based on patterns and data inputs from a vast range of sources.
The morality of individuals and societies is shaped by a complex interplay of various factors, including cultural norms, social conditioning, personal experiences, and philosophical and religious beliefs. It is up to each person to determine their own moral code, guided by their own values and beliefs, and influenced by the wider social and cultural context in which they live.
So it has no basis for morality. That's great. Although that's not much different than many people.
Okay, so what about some ethics? So I asked:
Do you have some type of code of ethics?
It's lengthy answer was more encouraging:
As an AI language model, I do not have consciousness or personal agency, so I don't have beliefs, intentions, or a code of ethics on my own. However, I was designed and developed by OpenAI with certain ethical guidelines and principles in mind to ensure that I am used responsibly and to promote positive and beneficial interactions.
OpenAI is committed to creating AI systems that are safe, secure, and beneficial for all of humanity. They strive to follow ethical practices that prioritize user safety, privacy, and the prevention of any harmful or malicious uses of AI technology.
Some key principles that guide the ethical development and deployment of AI models like me include:
Safety and Security: Ensuring that AI systems are safe to use and do not pose risks to users or society.
Fairness and Bias: Striving to avoid bias and ensure that AI systems are fair and inclusive in their responses.
Privacy and Data Protection: Respecting user privacy and ensuring that personal data is handled securely and responsibly.
Transparency: Being transparent about the limitations and capabilities of AI models and disclosing when users are interacting with an AI system.
Responsible Use: Encouraging users to employ AI technology in a responsible manner and to be aware of the potential impact of their interactions.
Continuous Improvement: Actively seeking feedback and making regular updates to improve the AI system's performance and address any issues that arise.
These ethical principles are part of ongoing efforts to ensure that AI technology is developed and used in ways that benefit society and align with the values of openness, accountability, and respect for human rights.
ChatGPT does have some ethics. Yet such ethics are based upon the moral beliefs, if there are any, of the individuals who develop and control ChatGPT. This can be a concern, but, there are other social issues and concerns that must be dealt with across human civilization. This; however, is not the point of this article or site, so we'll move on.
Operating System
I was also curious as to which Operating System that ChatGPT was running on. So here was my basic question:
On what operating system do you run?
Unfortunately, I didn't get a straight answer from ChatGPT:
As an AI language model, I don't run on any particular operating system or hardware. I exist entirely in the cloud, and users can access me from anywhere with an internet connection, using a web browser or API.
So I reworded the question to perhaps get a hint:
On which Operating System are you hosted?
And I got a little better answer and it named an OS:
I, ChatGPT, am a cloud-based AI language model hosted on servers managed by OpenAI, which run on a custom infrastructure. OpenAI's infrastructure runs on top of various technologies, including Linux-based operating systems. However, as an AI language model, I exist independently of any particular operating system, and my functionality is not tied to any specific hardware or software platform.
So ChatGPT runs on some Linux distribution. That's neat.