Act I: Every Company Is a Tech Company
The world is getting faster and more connected by the day, and it's practically impossible for a business to grow without technology. Cash flow management, supply chain tracking, internal and external communications, HR: all of it depends on software systems.
Think of any mid-to-large brand and picture the technological infrastructure holding it together. The website, the online store, the customer database, the live support -- and all of it scaling as the brand grows.
A brick-and-mortar's payment system going offline during Black Friday? Unacceptable. An architecture firm running entirely on paper? Impossible.
"Today, pretty much every company is a tech company."
So it's clear that, despite market fluctuations caused by a hundred different factors, the demand for tech professionals only tends to grow...
...and with it, the number of people trying to break in.
How do you evaluate that many candidates? How do you separate those who'll actually deliver results, even after a ramp-up period, from those who impressed HR with buzzwords and jargon?
Act II: The Parrot Is Smarter
Picture someone who keeps a parrot as a pet. Looking for interaction with its owner, the parrot mimics the sounds of words and phrases. Some words go ignored. Others get rewarded with treats, affection, or laughter, and there's no question which ones the bird will learn to say perfectly. To imitate, if we're being honest.
To imitate, the way a language model imitates human sentence structure based on input, then feedback from moderators and users, maybe?
No. The parrot is smarter. Parrots are surprisingly intelligent animals that can learn from humans and other animals alike. They're capable of social interaction, mischief, play, and affection. This parrot doesn't just respond to treats and head scratches: it can learn to act differently depending on its owner's mood, trying to cheer them up or calm them down. It can learn to communicate to get what it wants, like water, or the TV switched to a specific channel.
Parrots are organic beings with natural intelligence, and far smarter than any language model.
And you are far smarter than a parrot.
Act III: You Can Copy, Just Don't Copy It Verbatim
I'd bet you've said something like that at some point, or heard a classmate say it. No judgment here. We're being honest.
Part of your academic journey, if you think about it carefully, didn't get harder because the content got more complex. It got harder because the questions started demanding more from you. That's the difference between "2 + 2 = 4" and "calculate the difference in distance traveled between the two cars." Between "what's the main character's name?" and "can we trust a first-person narrator?"
That evolution, of reasoning and independence, is what matters, even if you still needed a passing grade to move on. It isn't easy, especially when it's expected from students who are young and haven't had enough life experience to draw parallels between content and reality. That demand is what makes each grade level or university semester feel harder than the last.
That's exactly why academic institutions have cared about plagiarism for so long, and why some have used copy detection software for years. There's a real difference between using sources to back up your research or argument, and "copying, just not copying it verbatim." One lets you grow and expand your horizons. The other keeps you dependent on everyone else.
Even the parrot doesn't limit itself to just imitating.
Conclusion: Time Is Money
Let's come back to the original problem: companies need to filter candidates, and it's costly and complicated.
Option 1: Hire based on resume review.
Sure, but we all know you can have four hundred certificates and know nothing, or know a lot with nothing on paper to prove it. The odds of a failed hire are high.
In the end, beyond all the costs of hiring and then letting someone go, you're also looking at slower development timelines, the need to spend even more time finding the next candidate, and the risk of the existing team feeling overwhelmed and quitting: a real snowball effect.
Easy conclusion: in tech, resume-based hiring alone doesn't work.
Option 2: The hiring process is led by the tech professionals on the team that'll work with the new hire.
Sounds ideal, being evaluated by the people who'll actually work alongside you.
Except: these people don't necessarily know how to run an interview or evaluate a candidate. They weren't hired to be interviewers. They may have a full plate and can't do everything and still interview dozens of candidates. They may get increasingly frustrated or burned out and more likely to quit...
It's one thing to spend 15 to 30 minutes with pre-screened candidates in a final round, to give the search that last comb-through. It's another thing entirely to be hired as a developer and spend more time interviewing than actually writing code.
Ah, yes, then let's hold on. Maybe option 2 isn't the best idea.
Option 3: A take-home challenge where you build a feature, or part of one, to specific requirements.
Now we're getting somewhere! Less pressure than doing it live and on camera, and you can look things up the way you would in a real job. Much better...
...unless the company asks you to build an entire project from scratch, taking up hours and hours of your time, and then doesn't even bother to send a rejection email. Hours and hours you could have spent studying, resting, or doing other interviews.
And who reviews your code? Here we're back to the path that takes excessive time from the company's tech team, who'll spend hours going through dozens of candidates' submissions. With any luck they'll have a rubric (which they still need to build), and manage to stay consistent across evaluations without getting tired or burned out... right? Yeah. Probably not ideal either.
Option 4: Hire a platform specialized in code challenges for initial candidate filtering.
There are several of them, HackerRank, Codility, and others, where companies can use pre-built questions or insert their own to create an online test.
Just don't get excited thinking it'll be enough to switch tabs and ask ChatGPT: your mouse movements are usually recorded and can be reviewed by interviewers, who can watch the full video to track your typing speed and development rhythm. The number of times you switch windows is also logged, and the platform often flags whether, and what, you pasted into the code editor. That's before you even consider whether you'll need to share your webcam feed to confirm it's you and that you're not using your phone or a second device.
You only move to the next round if the AI analyzes your code, your plagiarism score, and decides you passed.
A lot of pressure on the candidate. But a much better option for the company.
And time is money.
Conclusion, Part II: The Return
"Your code failed two of the hidden tests. Would you mind taking another look and seeing if you can figure out what's missing?"
That was one of the questions in an interview, led by developers at the company, after I'd passed the second stage of their hiring process: the online test, which had come after an initial interview with their internal recruiter.
The code review was done as a conversation. I got to walk through how I got to that result, discuss possibilities for optimization, and even ask for hints on how I could reorganize to cover the two tests I'd missed. I managed to cover one more, but I didn't solve the last one. We also ran out of time: a short conversation, around 20 minutes that felt like an eternity.
"We really liked how you walked us through your reasoning and organized your thinking. Some candidates pass all the challenge tests but can't explain what they did."
Look, it's obvious there's no shortage of answers online. Google, Stack Overflow, Reddit, LeetCode's own solution sections: all of that existed before ChatGPT was even a thing.
But what's the point of copying without understanding? How do you understand something and not be able to explain it? How do you build a whole solution path in your head and not be able to articulate it to an interviewer? How do you plan a solution using every tool at your mental disposal and not be able to translate it into code?
That's what's being evaluated. And that's what ChatGPT couldn't do for you even if it were known for writing perfect, error-free code, which is very far from the truth.
Don't fear the little robot, but don't depend on it either. Use it to help you prepare, as support along your journey if you want, and keep in mind that the hiring process will also use AI to evaluate you.
And that's okay.
Even with a buggy implementation, you are still smarter than a parrot who is smarter than a language model.
Done is better than perfect.