FrontEnd developer Denys Tkachenko recently received unexpected feedback on a test assignment that was hotly discussed on LinkedIn. It turned out that a test that was done too well can make an employer doubt whether it was really done by a human, and not by an AI.
FrontEnd developer Denys Tkachenko recently received unexpected feedback on a test assignment that was hotly discussed on LinkedIn. It turned out that a test that was done too well can make an employer doubt whether it was really done by a human, and not by an AI.
"Recently I received feedback on a test task. The task was pretty standard — to implement a dashboard with a table, sorting, metrics, and a settings page. The feedback was unexpected," the developer wrote on LinkedIn.
Because everything was done well, the company assumed that everything was done with the help of AI.
"Everything is too correct, structured, unified. It looks like it was made by AI. There is no sense of an individual approach...", they wrote in the feedback.
The company refused Denis and announced cooperation with another candidate.
According to Denis, in this test, he used AI only to generate data arrays.
In addition, Denis added that he believes it is wrong to draw conclusions based only on suspicion, which calls into question the meaning of test tasks in modern realities.
"Wouldn't it be fair to offer a technical interview in such a case, it's just a suspicion and everything would fall into place during a technical interview. They noted that this is not a technical flaw, but they need a developer who is more personally involved in the implementation of the project, and my 7 hours spent and the desire to prove myself and show a decent level are not involvement, but the use of AI?"
So the developer turned to the LinkedIn IT community with a question: "What should I do then? If you write poorly, you won't be accepted. If you write well, you still won't be accepted, because it's "too academic."
“I am more than sure that this feedback was generated by AI after it was given a project folder to analyze with one question: “Are there any signs of AI being used here? This game can be played both ways ,” wrote Middle Flutter Developer Yevgeny Myhalsky.
"This is another case that proves that the era of test tasks is already passing. If they are so afraid, then there is no point in giving tests at all. A rhetorical question arises: what is better for business: a person who uses AI and proves in a technical way that he understands what is happening in the code and as a result writes the code several times faster, or a person who writes completely by hand and as a result takes longer? So they will send you, how much will 2+2 be, you write 4, they will say you do not have a personal approach, it looks like you used a calculator, we do not see division in a column , "- said Python Developer Denis Tereshchuk.
“As an employer and a representative of a company that is actively growing, I can at least support it, in our case we are already on the verge of making the use of AI mandatory in code generation and other routine tasks, and for us it is a negative signal if AI is not used and everything is done manually ,” shared Oleksiy Sitar, who works at the international company WWG.
“Score, the person who wrote the feedback is the person from that meme with the no longer fashionable shoes and rolled-up pants with the inscription “if your boss dresses like this, you definitely won’t be replaced by AI” — would you like to work in such a team? ”, — noted Yuriy Skrypnyk, Team Lead at Avtologistika.
“That was also the case. “The text was written by AI, so it was rejected.” I just really forgot some phrase and, using the old-fashioned method, asked Google Translate for one single phrase (from 4-5 pages). And this very phrase was indicated as an example that the text was generated by AI)) well, it’s kind of boring… But on the other hand, it’s even nice. So it turned out pretty well, if they didn’t believe it was me,” Graphic Designer Alina Zamaraeva shared her own story, she said.
"Of course, everyone has their own requirements. But from my own experience, I can say that a lot of work still needs to be done for AI to produce working code , not to mention "structured, unified". Without knowledge and understanding, even with all the latest versions of AI, you won't be able to generate good code," says Frontend Developer Yevgen Rudofilov.
“Circus. It’s just that not all companies have yet directed their efforts in the same direction as AI, and are still pushing against it. I think that in just a few more years and a little let go, companies will understand that waving with AI is like waving with windmills, it’s more labor-intensive than using AI itself. Only from this point of view is it absurd. The industry is changing fundamentally, this is an obvious fact. The faster companies adapt, stop wasting energy on contradictions, the faster growth will come, ” wrote Junior Frontend Developer Artem Borisenko.
