Programmers Are Out of Danger: Why AI Won't Replace Developers Anytime Soon
Lately, there have been more and more forecasts about the imminent disappearance of the programmer profession — supposedly, artificial intelligence has already learned to write code and will soon completely displace people. However, this does not mean that AI is about to leave developers without work, because real programming is not just writing lines of code, experts say. The Moscow Polytechnic Zen (Dzen) channel analyzed in detail why AI is not yet able to replace live developers, although it will seriously change their work in the future.
Neural networks are not omnipotent
So many neural networks are being created and trained in the world, that it is already difficult to list them all. They can do a lot, and no worse than people: they write texts, generate pictures (very good and realistic ones), create videos, code, animate, animate photos and much, much more. And with all its capabilities, AI is only an assistant to a person. It can be trusted with routine tasks, used to save time, improve the quality of work with clients, but it will not be able to replace a good developer. And there are several reasons for this.
Firstly, routine work is not the most intellectual work. And while AI performs monotonous tasks, IT specialists can focus on more complex, interesting and even creative processes. Secondly, neural networks are good at template tasks (for example, code completion), but are helpless in unique projects that require ingenuity. For example, AI cannot cope with the creation of a data compression algorithm that will surpass existing ones (ZIP, Brotli) for highly specialized biological signals (say, DNA sequences). Thirdly, the IT world is changing rapidly: new languages, frameworks, paradigms appear. Neural networks are trained on existing data and lag behind current trends. Fourthly, AI can generate code, but its debugging, testing and long-term support require human participation. Errors in logic or non-obvious bugs often go unnoticed by AI.
And programmers don't just write code - they understand business requirements, communicate with customers, and analyze the problem. Neural networks work with templates and cannot consciously take into account the nuances of the real world. As you can see, there are many reasons and they are significant.
In addition, programmers not just write code – they understand the requirements of business, they communicate with clients and analyze problems: neural networks work with templates and cannot consciously consider nuances of the real world. As you see, there are enough reasons and they are weighty.
How does AI affect the IT market?
Based on the wording of the question — yes, AI is not replacing good programmers yet, but artificial intelligence is still affecting the IT market. New professions are emerging, for example, a PMT engineer or an AI developer. The former formulates PMT for neural networks, gives clear instructions to get the desired result. For example, instead of simply “Write code,” the specialist writes “Write a Python function for parsing CSV with the columns “Date” and “Price.” Note: the date is in the DD.MM.YYYY format, fill in the blanks with zeros, handle errors.” Then the AI produces not general phrases, but a complete and working code. And the AI developer creates and trains new neural networks.
The best defense is a good offense
AI will not replace programmers, but programmers who use AI will replace those who do not use it. For example, the first programmer spends an hour writing a standard REST API (Representational State Transfer Application Programming Interface, or, in simple terms, interaction with clients). The second programmer, who uses neural networks (for example, Copilot or GPT), writes the code in a minute - the rest of the time is spent on finalizing the logic and optimization. If you want to be like the second programmer, learn to work with neural networks. Or you can even start developing new neural networks. By the way, this is what they teach at Moscow Polytechnic University.
Artificial Intelligence at Moscow Polytech
Moscow Polytechnic University introduces neural networks and their functionality in the Artificial Intelligence and Machine Learning profile. Our students learn to design, train and optimize neural networks for various tasks (computer vision, NLP and predictive analytics). After training, the guys develop applications and websites that use AI - from chatbots to recommendation systems and voice assistants.
