Netflix took three and a half years to gain one million subscribers, Twitter two years, TikTok one year, Facebook 10 months, and Instagram only 75 days to reach its first million users. ChatGPT-3 had one million active users five days after its official public launch.
“It’s frighteningly good,” said the controversial Elon Musk, one of the founders of Open AI, alongside current CEO Sam Altman, who developed this language tool. Of course, Musk had his fingers in the business that will turn entire industries upside down.
ChatGPT (Generative Pre-trained Transformer, the one that creates new, original content) is a language tool designed to answer questions, provide information, and create content on a wide range of topics, trained on a vast amount of textual data from the internet such as various websites, books, news articles, and scientific papers, essentially everything published up to the end of 2020. It may not know that there is a war in Ukraine or that central banks have raised interest rates, nor that Prince Harry has severed ties with his family, but it can effortlessly generate texts in various styles and formats in a manner very similar to human writing.
Recent IPO?
For example, it knows everything about the financial crisis of 2008 and can write an analytical news article on that topic. With the right prompts, it can play the role of a financial consultant and create a presentation on trends in, for example, ecological food production in Eastern Europe, prepare a legal contract for a startup that will regulate the mutual rights and obligations of the owners, and besides that, it can write programming code. With a little skill, it can also develop a business idea based on a simple query, outline basic, long-term, and short-term goals, describe who it is intended for, what the specifics of the business will be, and how it will ultimately be financed.
It even goes so far as to suggest clients in the domestic market. Not only can it write excellent social media statuses, but it can also summarize any unread literature, plan a family trip to Madrid, and even design a weekly menu for a family of five based on fruits and vegetables available in Croatia during winter. It can generate video content and create visuals. And if it can do all that, it’s no surprise that it can flawlessly write a student paper, which is why some faculties, so far only in Australia, and it’s only a matter of time before others implement it, have introduced the requirement that papers be written by hand, while exams are conducted orally again. Even this moment has created a business opportunity: GPTZero is a tool that detects whether a text was written by a human or a bot (chatbot, or robot), and it already has millions of users.
A new version of ChatGPT-4 will soon be released, and the founders announce that it will be smarter and more current, having around 100 trillion parameters instead of the current 175 billion, which simply means that if you find the current ‘playground’ (as the field where questions are entered and where the bot outputs answers is called) impressive, just imagine what it will be like when it is even ‘smarter’ and equipped with knowledge from the current time.
ChatGPT is not the only tool based on artificial intelligence. There are already dozens of applications and bots that more or less successfully do just that – communicate with users on various topics. The vast majority of them are paid, and it is very likely that new versions of ChatGPT will also be expensive because the company that designed it certainly did not invest a huge capital to share it out of pure altruism. Their investors expect profits, especially Microsoft, which has already had two rounds of funding, and everyone is cheering for the upcoming IPO of the company, which could be a hit. Indeed, a recent presentation for investors revealed that the company expects $200 million in revenue in 2023 and $1 billion annually by 2024. By the way, OpenAI was recently valued at $29 billion.
– Training such systems is expensive. For example, training GPT-3 as the main component of ChatGPT cost around four million dollars, but the cost of using the model is currently just a few cents. This is actually very cheap and cannot be compared to the cost of human labor. However, this is just the first such model, and we expect that the systems coming will be even more powerful, and the cost of use even lower. Otherwise, we haven’t had such a large potential market in IT since the late 90s and the popularization of the internet. Predictions are that new mega-companies like today’s Facebook, Apple, Amazon, Netflix, and Google could emerge from this wave. One of the candidates is OpenAI, whose ambition is to become worth a trillion dollars, and I am sure it will not be the only one in that range,” emphasized Davor Runje, an IT expert and co-founder of the startup airt, which is based precisely on artificial intelligence.
