ChatGPT is leading the revolution in natural language technology, opening up a new world of possibilities in AI-powered chatbots. The chatbot, developed by OpenAI, has already seen huge success, with 100 million monthly active users since its release in November 2020. With its ability to understand natural language, generate content, and even detect and create fake news, ChatGPT has the potential to revolutionize how we interact with technology.
It’s been a long time since there’s been such a stir in the tech industry. Everyone wonders if ChatGPT and the AI that drives it will change the face of the tech industry and take jobs away from creative professions. Are we dealing with a significant leap in the development of technology, or is it instead an impulse to get stupid and hatch a generation of bullshit? Or it could be another significant reinforcement in the Big Tech world, which, during the economic slowdown, desperately awaits innovation that will cause investments and stock prices to increase again.
Table of contents:
ChatGPT Chatbot: The Natural Language Revolution in the Internet
ChatGPT is a chatbot powered by artificial intelligence. We already know chatbots and have been using them for several years, for example, in customer service. But what about him has really warmed up the Internet in 2 months? The key is to talk to him in natural language on all the topics that come to mind. You do not need to use rigid phrases; you can address him while talking to others. So far, only in writing – this is a specific limitation – but you can ask him questions, banter, or even confide in him. ChatGPT will respond with some fact, essay, or less in-depth advice. What ChatGPT hasn’t mastered yet are jokes. A sense of humor is a superior driving school.
OpenAI – The Company Behind ChatGPT
Before I move on, I want to emphasize that artificial intelligence is not a new idea, and work on it has been going on for several decades. However, ten years ago, they seriously accelerated.
The American research company, OpenAI from San Francisco, established in 2015 as a non-profit organization, is behind Chat GPT. Elon Musk, a perpetrator of the last year’s Twitter takeover, is best known among its original investors but withdrew relatively quickly from the project. Before that, however, he managed to attract the best minds in artificial intelligence to the project. In addition to Musk, the founders of the company also included:
- Peter Thiel (co-founder of PayPal, investor, and adviser to Donald Trump),
- Sam Altman (current CEO of OpenAI and previously manager of the White Combinator accelerator),
- Jessica Livingstone, founding partner of the seed stage venture firm Y Combinator.
- Reid Hoffman, one of the founders and directors of LinkedIn.
So behind OpenAI is a group of powerful people with a lot to say in Silicon Valley, both from the technical and investor sides. And with this technology, money is very important because the creation of OAI was a signal to Big Tech that now it’s time to focus more on accelerating work on the possibilities of artificial intelligence. OpenAI’s intenders wanted to study how the world can use artificial intelligence for the general good. Let us remind you that Elon Musk’s Fix idea is that artificial intelligence should not destroy humanity.
Investments In AI Are Growing Exponentially
In 2016, one year after OpenAI was founded, McKinsey & Co. and Forbes calculated that the most prominent players in the AI market have invested in it from 18 to 27 billion euros. Seven years ago, it was already a lot of money. Now the apparent investments are even more prominent.
However, by then, the world had already received a clear signal that AI was no longer just about theoretical talk but was starting to enter into practice. Before ChatGPT, OpenAI had researched and worked on many other AI applications. They, however, didn’t decide to release them to the general public.
Artificial Intention To Detect (And Create) Fake News
For years, she has been working on solutions using AI for chatting, such as GPT 3.0 (ChatGPT) or generating images (Dali). Moreover, the GPT 2.0 algorithm created by OpenAI for detecting fake news turned out to be so good that it recognized fake news and started to create it efficiently. It scared the researchers behind the project who decided not to make this tool available for general use. We could say after Nietzsche: “When you look into the abyss, it also looks at you.”
The E-mail That Changed Everything
In November 2022, specialists from OpenAI were already working on the GPT 4.0 bot, a newer model than the ones we are currently playing with. Version 4.0 was supposed to debut at the beginning of 2023, but work on it was so extensive that the team needed at least a dozen months to refine it. But then one e-mail came that changed everything. From the e-mail, they learned that they were to release the product as soon as possible. Pay attention – the mail came in mid-November, and on November 30, ChatGPT 3.5 was released to the world (13 days later).
Why such acceleration? Everything indicates that the leading investor of OpenAI, i.e., Microsoft (MSFT.US), has sensed the market demand. So big that it was necessary to release it as is – even in an unfinished version. Today’s version differs from the polished version that was initially supposed to come out in early 2023. However, OpenAI dug up this slightly dusty model, refined it, and released it last November.
100 Million Users In 2 Months
Let’s summarize: ChatGPT was refined in two weeks, and two months after OpenAI was made public, it became the most downloaded application on the Internet. According to SimilarWeb, there were 100 million monthly active users in January. For comparison, let’s add that Instagram needed 2.5 years to gain this number of users, and TikTok – took nine months.
And thus, ChatGPT started the AI craze and gave OpenAI itself global recognition. The company eventually changed its status to a for-profit organization, and in 2019, Microsoft invested a billion dollars in it. The publisher of Windows continues to invest and has added a round of $10 billion this year. The company aims to integrate the technology behind ChatGPT with the Bing search engine, which I mentioned recently in my article about the chatbot race.
Why Should We Get Excited About ChatGPT?
Ok, but why is ChatGPT electrifying the world so much? Here are some arguments:
- Has excellent natural language skills in many languages, not only English.
- It is a generative chatbot – it can search databases and create. Can write essays, articles, song lyrics, and poems.
- Ease of use – until now, you had to have some technical knowledge to use artificial intelligence to create something.
- ChatGPT has enabled anyone to enter this tool, enter a query, and have the result immediately.
Free – no comment here.
Of course, AP-powered generators make mistakes but note that the ChatGPT currently available is an earlier version. What will happen when OpenAI releases another one? Can creative professions start to be afraid? What about the authorship of scientific papers?
The worst thing, however, is that the technology that underlies artificial intelligence systems is Large Language Models (LLM). It’s known for, well, generating crap. When they don’t have access to information, these models make things up. Regardless of whether or not chatGPT knows the answer to a question, “something” will answer anyway. They use only the language resources they know. Significant risks exist when using tools powered by artificial intelligence and based on the LLMs listed above. The risk of making up biographical data, making up scientific articles, making up articles, and making up fake news. Chatbots get lost even in basic tasks, such as determining which is heavier: 100 kg of feathers or 100 kg of lead.
Will the Newer Models Bring the Desired Improvements?
Current AI models give a statistical answer. If we process much relevant information that says X is wrong, the AI will consider these statics and “spit out” such an answer. Let’s remember that AI trainers are still people, and people can be racist, have their political preferences, etc. The search engine algorithm encountered a similar problem. The only salvation is the hope that chatbots will be better trained and have better-verified databases. We all count on that.
