Вікторія ГорбікAI Eng
1 February 2025, 11:00
2025-02-01
IT residents are debating whether AI will replace developers by the end of 2025. Which answer option won?
Product Owner Oleksandr Sosnovchyk recalled Mark Zuckerberg’s words when in January during a podcast with Joe Rogan he stated that AI has already reached the level of mid-level developers in programming and companies can start «automating» their work by the end of 2025. The specialist asked programmers, engineers, and developers how they feel about this statement and what, in their opinion, the scale of automation of code writing will be. Here are some of the specialists who are already preparing to switch to another profession.
Product Owner Oleksandr Sosnovchyk recalled Mark Zuckerberg’s words when in January during a podcast with Joe Rogan he stated that AI has already reached the level of mid-level developers in programming and companies can start «automating» their work by the end of 2025. The specialist asked programmers, engineers, and developers how they feel about this statement and what, in their opinion, the scale of automation of code writing will be. Here are some of the specialists who are already preparing to switch to another profession.
«Given the speed of progress we have made in AI since the first announcement of ChatCPT (fall 2022…just think about it), I tend to believe Mark and think in 2-3 years, code will be written almost exclusively by AI,» says Oleksandr Sosnovchyk.
However, he adds that the developer profession will not go away (because who will support old products), but instead will transform. «How it will transform is an open question,» the specialist emphasizes.
And how did the developers react?
I believe
Some programmers and engineers are confident that many companies will have AI that will be capable of writing code at a mid-level. As Software Engineer Stanislav notes, there were already «alarm signals» for this in 2013. «I regret not investing a penny in myself or my business,» he adds. However, the deadlines are being pushed back a bit, for example,
Some of the developers who support Zuckerberg’s claim are already preparing to change careers or at least to train and expand their skills.
«I am already actively trying to improve my business thinking and product thinking so that in the future, when I will definitely have to change my profession again (I have already changed from an engineer to a programmer), there will be no shock from the fact that „Oh my God, I have been replaced by AI, how can I live now, how can I make money?!“, but on the contrary, there will be understanding and even a certain joy that „Cool, now with the help of AI you can create cool projects, implement your own ideas, without hiring a bunch of employees. Which means, I can try myself as an entrepreneur/startup and implement my own ideas, and not work on other people’s ideas“,» writes Senior Python Developer at SoftServe Oleksandr Viter .
He advises everyone who is currently afraid of losing their job due to AI to see this not only as a negative, but also as a positive side. According to Oleksandr, it is worth accepting:
The negative side is that you are no longer a unique specialist, your work has been automated.
On the positive side, you yourself can use AI to meet other people’s needs and make money from it.
As an argument that AI is already partially replacing programmers in companies, Petro, Tech Lead Full-Stack Engineer at ChampGroup, mentions ChatGPT version o1, which already demonstrates the level of a confident junior developer. The chatbot, according to the engineer, is able to make decisions independently and create working code. In addition, Petro also adds that for a year and a half there has been a concept where an AI supervisor, based on specified quality criteria using tests, manages a team of several AI models (QA, development engineers, etc.).
«AI is also acting as a consultant for many DevOps. Many hosting companies are already implementing AI-based automation of simple infrastructure solutions,» Petro notes.
However, according to the engineer, open questions remain, in particular:
Energy consumption — modern AI models require significant resources. Insufficient computing power — after solving this problem, models will be able to create higher-quality code, and their «work» will become cheaper.
Petro gave an example of a team he knows, which he does not name. The company implemented a solution based on an AI team to create a new project. Their estimate for a team of 10 people was 2 years. With the help of the AI team, they created an MVP in just 3 months. The cost of tokens for each «AI employee» was about $4,000 per month, plus 3 months of payment for the human team.
«Work will remain for everyone, but the threshold for entry for a person will increase, as more knowledge will be required. This can be compared to the evolution of media: television, newspapers, the Internet — all these platforms exist in parallel. «We need to learn,» the engineer adds.
I don’t believe it.
Some experts approach such statements with a certain amount of skepticism, particularly with regard to the term used.
For example, Artem, Senior .NET Developer, believes that this is unlikely in the next 5-10 years. «Today ChatGPT (4o and o1) convinced me for 30 minutes that Biden is still the president of the United States,» he adds.
Senior Software Engineer Javascript Pavlo Bozhok suggests that one day AI will definitely write code with the help of a person who knows how to write it. This means that fewer engineers will be needed than now, but they will still be needed.
«But the ones who really need to worry are the managers, they don’t fit into the concept of the future IT sector at all,» the engineer writes.
In this case, according to Pavel, there is hope for managers with a technical background.
And Python developer Igor Rys also reminds us that artificial intelligence cannot be fully trusted yet. «In most cases, AI produces an approximate result that requires editing, reality checks (because it hallucinates, be healthy), and often a banal rejection, because somewhere the prediction did not work and the AI produced something completely different. I certainly did not spend $500 to play Devin, but for some reason I am sure that the same problems exist there,» — supports this opinion Yevgeny Nikitchenko, Senior software engineer at DXC Technology .
Other
Elena Arshinnikova, Software Engineer and CEO at EVA Solutions, touched on the cost of energy resources that will be required for AI tools. «We can assume that the mass use of AI in solving work tasks will be cheaper than paying salaries, bonuses, and shares to developers,» she believes.
Yevhen Nikitchenko does not rule out the possibility that AI agents will actually «dust the code.» «Well, there’s nothing left to do but develop a creative approach and learn not to be mediocre, then you can’t be afraid,» the engineer advises.
And according to him, if AI is no less creative than humans, then I don’t think there will be much room for any of us. «But it looks futuristic enough to me not to worry about it yet. I think it’s more realistic to see the overheated AI market meet the reality of inflated expectations,» adds Yevhen Nikitchenko.
Mark Zuckerberg said that AI has already reached the level of middle-level developers in programming and companies can start «automating» their work by the end of 2025
Мануал для джуна. Python Developer: запитання на співбесіді, практичні завдання та поради для початківців
Джунів, які шукають роботу, проходять купу співбесід та отримують відмови замість оферів, під час війни побільшало. А все тому, що до співбесід необхідно ретельно готуватися.
dev.ua продовжує серію матеріалів про те, що треба знати джунам для проходження співбесіди й отримання омріяного оферу від першого роботодавця.
Олександр Ковтунов, Python Software engineer компанії Ajax Systems, який також проводить співбесіди для junior- і middle-фахівців, розповів про професію Python Developer і надав поради, як початківцям долучитися до професії.
Як нейромережі бачать вільну та незалежну Україну? Тест dev.ua
Нейронні мережі для генерації зображень бачать світ по-своєму, їхню логіку зрозуміти часом зовсім неможливо. Але таки хочеться. На честь Дня Незалежності України редакція dev.ua вирішила провести невеликий експеримент.
Ми задали чотирьом різним нейронним мережам п’ять однакових запитів: «прапор України», «День Незалежності України», «український Крим», «перемога України» та «українці». Отриманими результатами ми ділимося з вами нижче.
У TikTok тепер можна генерувати фон за допомогою нейромережі. Ми протестували її та ділимося результатами
У TikTok з’явилася нова функція «Розумний фон». З її допомогою як фон для тіктоків можна підставляти згенеровані нейромережею зображення. Редакція dev.ua протестувала цю технологію і ділиться своїми враженнями.
