Transcription
Not long ago, I published a video about ChatGPT, showcasing the capabilities of the bot that are rarely discussed. For example, it can search for vulnerabilities in code, write complex programs in various programming languages, and so on.
The main idea of this video was: can ChatGPT replace programmers, copywriters, designers, and other IT specialists? Many people in the comments were divided into two camps. The first group argued that it would take just a few years, at most 5-10 years, for this artificial intelligence to take over the universe. Programmers would become unnecessary, and all these specialists might be automated. This could lead to a situation where the only job left would be to maintain this bot.
The second group of people was skeptical about the bot and provided real examples of when it made significant mistakes. Sometimes, you can ask an obvious question, and the bot fails to answer. There are moments when it writes brilliant code for projects, but sometimes it uses outdated libraries or APIs that don't compile at all. At the very least, the code generated by the bot needs editing.
Practically no one mentioned that we should be happy because we have such a tool. By using ChatGPT to study material, you can outpace all your competitors and learn much faster than those who don't use ChatGPT and the method I will discuss in this video.
There have already been three videos on the channel where I talked about methods for quickly learning programming. If you combine all these methods with ChatGPT, you can develop at an incredible speed.
At the end of the video, I will also share some pitfalls that guarantee that the bot won't be able to replace even junior developers, let alone experienced specialists. So, let's look at the method and actively learn.
Let's imagine that I just entered the programming field. I want to create a content plan for learning a programming language. In our example, it will be Python, but you can use any language or even any field. You can also create a content plan for learning English or any other area, not just programming.
Okay, let's ask the bot to create a content plan for learning Python. I know that many people are looking for this material, and ChatGPT can generate it in just a few seconds.
It's also very important to base your plan on the specific area you intend to work in. This could be web development, desktop development, game development, or creating parsers or Telegram bots. You can specify this in your question. The bot will create a roadmap tailored to your needs.
Moreover, you can ask it to review all job vacancies for the same position in web development, such as backend developer, and request that it prioritize the requirements for the position. This means placing the technologies you need to learn in-depth at the top and those you need to learn more superficially at the bottom.
This bot has a vast amount of data and will create an ideal content plan for you, even for job applications. Then, for each technology, you can ask it to provide links to articles, videos, and so on, or even break down important aspects of that technology.
For example, if it's Python, and you need to create interfaces, you can ask what specific aspects you need to study in that technology. For instance, you might need to learn about creating windows, styling, and so on, but you don't need to delve into deeper topics. The bot can think through all of this as well.
Next, the most interesting part: we get a study plan that we can save in a separate file. Then, we can take each topic one by one and ask the bot to provide links to websites where we can read more about it. It gives us answers.
The most interesting aspect of this method is that I can ask it for more information and inquire about what to learn next and which technologies to study for a deeper understanding. I can repeat this question several times, and at each stage, it expands the content plan and provides new information.
I also want to tell you about the University of Artificial Intelligence, which is an innovative online educational institution, number one in the market in Russia and the CIS. Over 5,000 students from around the world have already passed through it. They have a huge content base on AI, and why this topic is so interesting is that it is an area where creativity intersects with science.
AI specialists need to come up with new architectures for neural networks, test them, and find new interesting applications. Right now, there is an opportunity to participate in a free three-day intensive course where you can develop nine neural networks even without programming skills and try yourself as a specialist in a full-fledged course.
There are also a lot of cool things, such as writing your thesis and creating a portfolio that will include 32 completed projects in the form of neural networks. There is also a so-called cost buyout option, where you are hired after training, and the cost of the course is returned to you through your salary.
Additionally, there is a job placement guarantee written into the contract, and the university has already placed students in companies like Huawei, Dell, Amazon, VTB, Yandex, and many others.
But let's not forget about the high salaries in the IT field, especially for in-demand AI specialists. The minimum salary for a junior is from 70,000 to 100,000 rubles. This course is suitable for anyone looking to increase their salary and expand their skill set. More detailed information about the free three-day intensive is available via the link in the description.
Next, you start studying each topic in turn and ask the bot about that technology. For example, if we are studying loops, we can ask the bot for examples of loops in Python. If we don't like them, we can ask for five more examples. If we need to combine loops with another technology, we can ask the bot how to implement that.
Sometimes it will repeat itself or provide obvious information, but if you ask the questions correctly and repeat them, it will explain almost any topic to you without any problems. Just think about the scale of this bot. You can spend the whole day asking it to show different examples of loops, inquiring about their features, asking which is faster, a for loop or a while loop, and why one is faster than the other.
You can ask it to explain in more detail. You can think of this bot as a personal consultant that will answer any of your questions and is available 24/7 to find information for you.
Now, compare these advantages you get from this bot to another person who wants to start programming but either doesn't know about GPT or has the stereotype that it's just another dumb tool that can't do anything. This person will be forced, in case of errors, to visit multiple websites, gather information bit by bit, ask questions on forums or in chats, and spend hours waiting for answers to their questions, while also bothering other developers because they ask the same question a million times, like "Why does string plus give an error?"
On the other hand, you can learn much faster than these people. For example, I had a problem with parsing websites where content is rendered only through JS. In other words, it was impossible to scrape these sites without using a browser. I asked the bot to find tools that would allow me to do this purely through requests, and it provided me with five libraries.
Previously, when I searched for such tools, I visited all the websites one by one and couldn't find a single decent tool. But the bot gave me five libraries that I had never heard of, even though I have been developing for about four years.
You can ask the bot for the names of different libraries for your goals and request code examples for implementing the necessary functionality. Imagine you are studying loops and don't know how to make the bot traverse nested structures. You can ask ChatGPT to provide an example of such code, analyze it, and understand how it all works.
If you encounter an error or can't grasp the principle, you can clarify it with the bot, and it will respond just like a real person. So, while it's free, make the most of it—ask questions, take screenshots, and save its answers. Because when this bot becomes paid, it will be too late.
Right now, you can learn almost any technology just by asking this artificial intelligence.
At the end of the video, I promised to answer the question of why this bot won't be able to replace programmers, designers, and any IT specialists in general. The fact is that artificial intelligence can only provide dry answers based on its pre-defined logic. A human, on the other hand, has a way of thinking that machines simply do not possess.
Most of the time, clients themselves cannot clearly formulate a technical task because they don't understand how it should work at a logical level. Only a programmer can understand this and clearly describe the technical specifications. If you give some task to an artificial intelligence, it simply won't be able to write software because there is no clear assignment.
This bot needs everything clarified down to the smallest detail, and even then, it won't write you a complete project right away. It will most likely give you small pieces of code that may not even work. Therefore, in any case, you need a programmer at the very least to create a technical specification, break it down into components, and then these small components can be created by the bot.
Even then, many parts will need to be rewritten and adapted. So, it can serve purely as a convenient replacement for a stack of airflow and nothing more.
Just imagine a situation: you write a Telegram bot, and you need to embed a Python interpreter so that the user can send code to the bot and get a response back. A human can immediately understand the pitfalls involved. They know that access to the web needs to be restricted to prevent malicious requests from your server, and many system commands need to be prohibited to avoid compromising the system.
Additionally, this module needs to be placed in a separate sandbox so that in case of a security breach, it cannot harm the system. There may also be a need to protect the bot from DDoS attacks or spam.
Now, can you imagine how GPT would handle this? I honestly can't even guess how it would implement all of this because it's an unrealistic task for it.
Now, imagine you work at Yandex, and your project is 45 gigabytes. You have one class that is connected to many other implementations. Let's say some coder created such bad code that one module is tied to 200 other modules, and minimal editing breaks the entire project.
You give this AI the task and say to rewrite it so that this class is no longer dependent on anything else. This means it needs to rewrite 200 modules with other classes. If it is inherited in many places, it will need to rewrite all of that as well.
This logic also needs to be adjusted so that all 45 gigabytes of the project interact correctly with this class. Just think about the scale of such a project and understand that this bot won't be able to solve this problem in 5 or even 10 years.
Even if it is trained on quantum computers, it won't make a difference. Just think about how many resources would need to be spent to achieve this and how much it would cost to maintain such a bot when the same programmer can be paid much less.
At the end of the video, I have very important information that is not an advertisement. Recently, I found a really cool project that also helps you develop much faster. This definitely deserves your attention, and I haven't seen such videos on YouTube at all.
Let's try to get 3,000 likes. I made a really cool video reviewing this project and will show how you can set yourself up for learning new material and how to do it much more effectively. These are not some banal tips; this is a real working method.
Also, write in the comments what you think about whether ChatGPT will replace programmers in your opinion or not, and explain why you think that way.