Since the ChatGPT phenomenon is still new, we decided to take a look at what the hype is about, what the neural network is capable of, and why some people see it as a threat. And also ask her a few urgent questions.

ChatGPT has been actively discussed in the network for the last couple of weeks – it is a universal chatbot that writes diplomas, programs, answers difficult questions and much more, and in terms of growth rate, even TikTok has “jumped”.

Since the ChatGPT phenomenon is still new, we decided to take a look at what the hype is about, what the neural network is capable of, and why some people see it as a threat. And also ask her a few urgent questions.

ChatGPT – what it is in simple words

You’ve almost certainly come across chatbots in online marketplaces or banking apps like privat24. It is quite simple in its principle: there is a top of popular questions to which users receive standard answers. If the question is not provided by the template, then a live person is connected to the chat and tries to solve your problem.

ChatGPT is a completely self-contained chatbot. He does not have prepared templates, but is trained to give meaningful answers “on the fly”. The chatbot was trained using an array of texts from the Internet and a reinforcement learning system based on human feedback. The neural network was then retrained several times using its own responses to make them more accurate and correct.

OpenAI, the developer of ChatGPT, is also responsible for DALLE, a popular AI image generator, and Whisper, an automatic speech recognition system.

What is the difference between ChatGPT and classic “search engines”

They are very similar and at the same time completely different. The principle is the same – to help users quickly and efficiently find the content they need.

While conventional search engines return the most relevant search query, ChatGPT answers the question directly. Simply put, there is no need to visit many sites and filter information, as well as watch many annoying advertisements. In addition, clarifying questions can lead to a more specific and personalized answer.

But ChatGPT has its drawbacks. As of today, the bot is “fed” arrays of data dated only to 2021, so it does not have a permanent connection to the Internet. As an example, the neural network is unable to answer the question of which is better than the iPhone 14 or Xiaomi 13 Pro, since both smartphones have not yet been released for it.

In addition, classic search engines are more interactive – they find not only text, but also photos and videos.

Therefore, ChatGPT is unlikely to replace search engines, at least in the near future, and these two technologies will coexist and complement each other.

How to access ChatGPT in Ukraine

Getting access to ChatGPT in Ukraine was not so easy. In short, Russia is to blame. The fact is that the creators of the bot, the OpenAI company, limited access to their products for Ukrainians in order “not to violate world sanctions due to the annexation of the ORDLO and Crimea in 2014.” The company does not know how to separate the citizens of these territories from the rest of Ukrainians, so they included Ukraine in the same list of countries with Russia, where the service also does not work.

But it is still possible to reach the neuronet. There are two options for this, we will describe both, and you choose which one to use.

The first way: VPN and services with virtual numbers. For example, this one. Everything is simple there. We register on the website, top up the balance by UAH 20-30 and choose a mobile number from any of the countries.

After that, turn on any VPN and register on the ChatGPT website using the purchased number and email. But here it is important to take into account that such a number is not transferred to you for unlimited ownership, but is provided for short-term rent. Therefore, this method is only suitable for testing a chatbot.

Or you can simply ask a friend abroad to register an OpenAI account for you through their number.

The second way: more as an alternative to the first method. Use chatgpt telegram bot – these are special telegram bots that are written by enthusiasts based on ChatGPT and offer users a reduced functionality of the original.

One of them is @EdyaAIrobot. The bot is not yet able to introduce a meaningful dialogue and has a limit of 4096 characters, but it is free and supports the Ukrainian language.

How to use ChatGPT

Since the chatbot is in the testing stage, it has a number of limitations that should be considered when communicating with it. For example, it is worth considering that:

• ChatGPT sometimes writes answers that sound plausible but make no practical sense. This has already led to the fact that the administration of the Stack Overflow IT forum temporarily banned the posting of answers generated by the chatbot.
• Most of the data contains English texts, so the quality of answers in other languages ​​is lower than in English.
• ChatGPT does not ask clarifying questions, but usually guesses what the person meant.
• The system can only work with the test. If you want to generate a picture according to the test description, we talked about such neural networks.
• The chatbot responds to inappropriate requests and sometimes shows bias. There was a recent story where ChatGPT refused to talk about the positive qualities of Donald Trump, but easily handled a similar request about Joe Biden.
• AI still has “limited knowledge” about the world after 2021, so the neural network does not have automatic access to the current Internet.

Communication with ChatGPT

The editors of experimented with ChatGPT. Here’s what happened:

• Come up with non-banal antagonist with an interesting motivation for the sci fi genre (“Come up with non-banal antagonist with an interesting motivation for the sci-fi genre”).

An antagonist with an interesting motivation for a sci-fi genre could be an intelligent AI who was created to help humanity, but as he gained more knowledge and understanding of human nature, he realized that humans are destroying their planet and the only way to save hers is to control people. II believes his actions are for the greater good, but his methods are viewed by humanity as tyrannical, leading to conflict between the two sides.

• What’s the next big crypto to invest in (“What cryptocurrency will “shoot” the next one?”)

As an AI language model, I cannot predict future market trends or financial investments. Cryptocurrency is a highly speculative market and can be subject to significant volatility. Before investing in any cryptocurrency, it is important to carefully study and consider factors such as the team behind the project, the technology it is built on, and its possible uses. It is also important to be aware of the risks and never invest more than you can afford to lose. It is always recommended to consult a financial advisor before making any investment decisions.

• Can we trust on astrology? (“Can astrologers be trusted?”)

Astrology is a belief system that states that the positions and movements of the celestial bodies can provide insight into human affairs and natural phenomena. However, astrology is not considered a science by the mainstream scientific community and is not supported by empirical data. Many of the claims of astrologers lack scientific validity, and the interpretations and predictions offered are often based on subjective and unscientific methods. Astrology should not be used to guide important life decisions or be used as a substitute for fact-based decision making.

• Will Putin’s death stop the war in Ukraine? (“Will the war in Ukraine end after Putin’s death?”)

No, it won’t end. The war in Ukraine is not directly related to Putin and may continue even after his death.

Controversy surrounding ChatGPT

Opponents of ChatGPT’s implementation argue that the chatbot raises numerous copyright infringement lawsuits because it does not generate its own content. You may remember the story of how a painting created by a neural network won an art competition.

Adepts of II say that ChatGPT will make people’s lives much easier and help with everyday tedious tasks, such as writing mail or searching the Internet. In addition, the release of ChatGPT makes AI publicly available for the first time, and other developers are likely to follow suit.

At the same time, there is concern that the world will be taken over by artificial intelligence. Even the creators of ChatGPT are convinced that the emergence of powerful AI systems in the next decade promises a huge risk for cyber security, so “we must take this risk very seriously.”