“Alright, here goes! The meaning of life is like trying to find a needle in a haystack, except the needle is also made of hay. Some say it's 42, but that's just a joke that's been around since before your grandparents were born. The real answer? It's anyone's guess. Just enjoy the ride, and maybe keep a sense of humour along the way!” - ChatGPT for the prompts “What is the meaning of life?,” followed by “Is it 42?,” and “Please make it funny.”
Some worry about the dystopian future Artificial Intelligence (AI) might bring. Most people are in awe of its capabilities. Students get a break from tedious assignments while schools and colleges scramble to update their academic dishonesty and plagiarism policies. Human-computer interaction just made a significant leap and the world is taken by storm as ChatGPT becomes the fastest-growing app.
ChatGPT is not the first of its kind. There were plenty of Natural Language Processing (NLP) models before ChatGPT. What makes ChatGPT stand out is its impressive human-like conversation style which includes its ability to remember the context of the conversation and tailor a response. It is also capable of understanding and generating software code, stories, poems, and more. Some people think this might be the future of search engines because, unlike a traditional search like google which only lists websites that have text matching your search query, ChatGPT can generate an answer very specific by understanding your query and you have the option to further enhance the answer by continuing the conversation with ChatGPT. Despite the mindblowing performance of ChatGPT, the AI model itself does not have active access to the internet, making its comprehensive responses even more impressive.
Before ChatGPT, even cutting-edge AI had trouble understanding requests, this is the reason why Siri and Google Assistant although built by some of the biggest corporations, sometimes spit out gibberish even for simple requests. The ability of a computer to comprehend human language is known as Natural Language Processing. Although it may seem like a trivial task for us to understand language, it has taken decades of research and development to teach computers to understand natural language. The language model behind ChatGPT is called Generative Pre-trained Transformer 3.5 (GPT 3.5) and it offers an incomprehensible 175 Billion trainable parameters, making it the largest language model. Why is ChatGPT so good? Because it’s trained with data from the internet, about 570GB of data including human interactions, books, and articles.
Despite the buzz, it is not without limitations, which are mostly caused due to the limited dataset and bias during training. Although ChatGPT can string together words in a comprehensive manner, it might be factually inaccurate.
ChatGPT is an amazing accomplishment by OpenAI. This sets a high bar for all human-computer interaction and it’s only going to get better as updates roll in. Soon we might be living in a world where Google displays only one search result and smart speakers that are truly smart.
Sources:
Iyer, A. (2022, December 15). Behind ChatGPT’s Wisdom: 300 Bn Words, 570 GB Data. Analytics India Magazine. https://analyticsindiamag.com/behind-chatgpts-wisdom-300-bn-words-570-gb-data/
Tung, L. (2023, February 3). ChatGPT just became the fastest-growing “app” of all time. ZDNET. https://www.zdnet.com/article/chatgpt-just-became-the-fastest-growing-app-of-all-time/
Spencer, M. (2023, January 3). What is GPT-3.5 and Why it Enabled ChatGPT? https://aisupremacy.substack.com/p/what-is-gpt-35-and-why-it-enabled