The rise ofAIhas been rapid, moving from tools used mainly for research to systems that can now assist with complex tasks such as coding. AI agents have become an integral part of many developers’ workflows, helping generate functions, fix bugs, and speed up routine programming work. With this, developers have started relying more on AI tools to produce working code and handle repetitive programming tasks quickly.
Recently, a developer reportedly struggled to explain a function he had included in a PR (pull request) during a review.
The incident was described in a post on X by user @sickdotdev, who said the incident happened within her own team.
Also Read: Techie admits struggling to code without AI tools: ‘I don’t think I’ve learned much’
Developer struggles to explain code:
According to the post, the developer had submitted a function as part of a PR but could not walk through the logic when asked during the review.
“A developer on my team recently couldn’t explain a function he submitted in a PR,” the post adds.
The post explained that the issue was not about coding skills. Rather, the developer had copied the function from Claude, an AI tool capable of generating code and text.
In the post, the user said the developer had simply copy-pasted the function after the AI tool generated it. Since the code worked as expected, it was included in the pull request and pushed forward.
However, things became awkward when the reviewer asked the developer to explain how the function worked.
“Not because he didn’t understand coding. Because he didn’t write it. He copied-pasted it from Claude; it worked, he shipped it,” the post added.
According to the post, there was a long pause as the developer struggled to describe the logic behind the code.
“When I asked him to walk me through the logic, there was a long pause,” the post adds.
HT.com has reached out to the user for more details. This report will be updated when she responds.
Also Read: Senior tech worker claims AI reduced 75% work but raised workload: ‘Everyone is expected to output 5x’
Check out the post here:
Here’s how people reacted to the post:
X users reacted by sharing their own experiences with AI tools in programming. While some said such tools help save time, others warned that relying on them without understanding the code could create problems later.
One of the users commented, “I would bet you even at large companies this is happening in 50%+ of PRs.”
A second user commented, “If the tests pass and the code works, the ‘how’ is secondary.”
A third user commented, “Developers used to copy-paste stuff from Stack Overflow not that long ago. I remember quite a few similar situations from back in the day. At least back then, you had to check if your code worked as expected.”
“He’s a good guy; if it was someone who’s good at lying, he would have manufactured words for you,” another user commented.
