Hey I'm Maciej from Eonea talking about data security. Today I'm going to talk about something that privately ... pulled out of my shoes seriously .... ChatGPT from OpenAI. You can chat with it, ask anything and the answers well surprise you. Rumor has it that Google has something to be afraid of. This could be a real, real game changer.

Google - How it is now

It's been said that Google has everything, and it probably almost does. We are used to the fact that we type a keyword into Google and get a set of results, pages, links that according to him best match our expectations. Not even necessarily correctly answer our question. After which we spend a shorter or longer, and sometimes much longer, time searching the suggested sites and most often we find what we need. This has been the case so far, but will it continue?

ChatGPT (Generative Pre-training Transformer)

This is where ChatGPT unexpectedly comes into play. Because, for example, you might think this process at Googling takes too long or is just not efficient, seriously. Want to know the recipe for apple cider? Want to write a short article on any topic? Want to generate an original wish? Want to know if god exists? Ask and you will get a coherent, factual and not uncommon answer.

What is ChatGPT?

It's a chatboot based on OpenAI - machine learning. Colloquially though to the point of being correct sometimes called "artificial intelligence."
ChatGPT is something like a large database of knowledge about the world additionally equipped with something so far unique, an interface specially designed and trained for conversations with people, to understand, analyze and answer questions and even have conversations with people in a way that is natural to us and surprisingly coherent. Often it can deceptively resemble a conversation with an actual person. Interestingly, ChatGPT, according to its creators, has no current access to the Internet and information created after 2021 may not be relevant to reality. Everything I'm talking about today is current as of December 2022.

 

Keep in mind that this particular solution is non-commercial, still under development and testing, the developers mention the possibility of daily conversation limits. ChatGPT is learning all the time. And how is it learning? Well, by talking to us. As it is with machine learning, simplifying the more data the better the learning results.

Is ChatGPT lying?

As the creators claim, it is designed to "speak" the truth according to its own knowledge and, of course, within the framework of how much it will be able to correctly understand our speech. In my experience, there should be no interference in communication to us because my impression is that it communicates all too clearly and unambiguously.

Fun with ChatGPT

I think it's really worth spending at least some time playing with ChatGPT. I played with it for a while and will tell a little about it. First, as it happens with IT news, I assumed that he did not know how to speak Polish and I communicated with him in English. After some time, I decided to ask him to translate what he wrote into Polish and ... It turned out that my previous strength in English was nothing more than knocking down an open door ... well it happens.

 

As a non-professional in baking, greeting my wife, I started with a question about apple cider. Well it was probably too trivial because I immediately got a list of ingredients and how to prepare it.

 

 

Then to be a little more timely I asked for Christmas wishes for IT professionals. I did not expect shots well and there were none so I asked to write more witty. Slightly better, but well the sense of humor is very subjective even among people, so I guess expecting a sense of humor and yet universal is probably a bit too early.

 

 

OK I know I know asking AI to be witty is not fair because it's not fair even to humans let alone a fledgling AI At least yet.

 

It's time for a slightly more practical test. Suppose we want to transport books from China and whatnot. So what question might we have? How many books will we fit in a shipping container. Well I got the answer in a nutshell that it depends a lot, but I got clear information on what it depends on. Well, and the "final" answer from a few to several thousand books. This is already for me a pretty good estimate as the fact of reserch at the level of 15 seconds to write a question, do not you think? Interestingly enough, I asked this several times about received different answers. However, they differed in content and not in substance. In one attempt, I even learned that around 17,000 books. In another attempt, I learned that when estimating, it should be remembered that books in containers must be packed stably and securely, which affects the quantity. You will admit. that quite factual and specific.

 

 

Ok since I already have an assistant I asked him to write a sample message for me to Luke for example, explaining that I am unable to complete the sentence. The answer was nice but, after all, it is currently about fun so, I asked for a more official version. The result was more than satisfactory. Still at the end I asked for a signature and it was done. Please see with what ease ChatGPT re-edits the content it throws out based on our instructions.

 

 

So at the end I somewhat philosophically asked about the future. So the obvious question seemed to me to be whether humanity would survive the next 1,000 years. And here, too, our ChatGPT did not disappoint me. It told me in a rather optimistic tone, first pointing out that the question is a difficult one. He told me that humanity has many problems like climate change wars terrorism etc. but in general as humanity we should be able to cope. At the end he wrote that he could not guarantee that humanity would survive for 1000 years ? Encouraging right?

 

 

Out of chronicling duty. Although I present this case of predicting the future somewhat jokingly, it is as if the impression of omniscience of such a ChatBot should be produced in us over time. However, please remember, it is only processing the existing information provided to it. Alternatively, it can draw conclusions from the information provided to it.

 

What can ChatGPT be?

Admittedly, the easiest way to prove someone incompetent is to ask about the future, but the most obvious solution is to use this technology as just assistants to start with in companies then probably individuals. This is nothing surprising, but imagine such an assistant connected to the boarding house. Able to order a cab, shop, play music, I know still nothing this is already happening. Google assistant, Alexa, etc. It's just that these days, yes I know I'm exaggerating, these assistants in practice nowadays their primary job is to listen to us to sell us the best advertising. Occasionally search for something on the Internet or add something to the calendar. Unfortunately or fortunately, it remains to be seen ChatGPT is already something that could become a universal, human-machine interface. It is easy to imagine an assistant which, in practice, based solely on simple voice commands, will make appointments for us, answer emails, write posts on social media, what's more, it will filter information for us according to our tastes, and not necessarily according to what we see on G bear, FB or YT. It will be able to carry out conditional commands, for example, if someone calls let me know or, on the contrary, ignore for two days. Such an assistant will search the web for us for opportunities, solve complex business and personal problems by searching the entire Internet. He can write articles for the blog, learning our style over time. The content and, in fact, the manner and formulation are, as it seems, unique.

ChatGPT - Google unnecessary?

Please pay attention to this may happen soon. Wanting to get information of value to us, we bypass ALL, google, ads, social media, search engines, web portals. Until now, Google has even been synonymous with finding information while collecting information about us and making a very good living from it. If Google doesn't do anything about it, it may be that the golden age of Google has just passed. Not surprisingly, according to various reports, we currently have all kinds of red alarm lights blinking at Google. While I'm obviously not betting on the demise of Google here ? I do think that a strong reshuffle could happen here.

Summary

I seriously believe that there is a breakthrough in the field of human-machine communication. And whether this is a good thing or a bad thing has no idea. Here's to probably seeing.