Don’t Cheat on Interviews. It’s Obvious When You Do

The interviewer is incentivizing this behavior by playing gotcha with computer science knowledge. The only reason I know about radix sort is from college and I have never used it. If they are hiring a college professor, then I understand but otherwise this is ineffectual and biased towards young college grads.

If your interview is breaking when someone googles information, then your interview is flawed.

I want to point out the insanity of filtering on an interview question that has a low success rate.

Presumably your candidate has a bachelor's degree. By this point in their life they've answered thousands and thousands of questions. They taken countless exams, and there is solid aggregate data which summarizes their performance (GPA, standardized tests, certificates, etc).

By filtering on a few challenging questions, the interviewer is throwing these statistics in the trash and rolling the dice to disambiguate the best and worst candidates. What do you think the variance is on a five question exam? Not great.

Rather than try to assess decades of expertise with five questions, I think you are better served by asking questions which pull in orthogonal information.

Do the candidates career goals align with the companies? Do they seem pleasant to be around (also high variance, but orthogonal). Are they enthusiastic about the position?

Hiring someone based on whether or not they know a particular algorithm that may or may not have any business relevance to their work, speaks to the process. Obviously the candidate was in the wrong here, but I question what this sort of teaser question actually buys you.

Experience and depth are varied. If you happen to ask a niche question where someone has a lot of depth, then you've identified perhaps deep knowledge on that narrow topic, but what about general problem solving and programming skills?

Funny that the Author says counting sort is O(n) when it's really O(n+r). It's almost like making interviews around the knowledge of some case specific sorting algorithm is a bad idea...

> At this point, it was very obvious to me that he googled the solution. I was amazed that he had the guts to do so in an interview. Just WOW.

Frankly, I'm amazed at the guts to say you shouldn't just search for the answer. It's what you want an employee to do. Not just sit around and "think about it" for an hour, only to write something terrible. Search it and find someone who's already done it and use their working solution.

I've also had the "guts" to search for things on google in interviews. I've done it doing remote interviews where I share my screen with them. Best time was when I was asked "how do you iterate through a json dict in javascript?" Not being a javascript developer, I'm not sure why they were asking me. But I searched it, and found that one of results to stackoverflow was purple. I told the interviewer "this must be good, since all the other ones are blue." We all laughed (it was a group kind of interview) and I was told "Your Google-Fu is strong."

I very quickly received an offer.

This is a low quality cheat. There’s a website where people share how they got through interviews and it includes stuff like “Wait for the interviewer to look at the code you’ve written for this specific problem and spot the bug on Line k. They will then ask you if you can think of any edge cases. Try the following edge cases that appear to pass until you hit this edge case that detects the bug. Fix it. Wait for a hint on making it more efficient, then depending on the hints follow this action”. Oddly enough k the person is probably well trained on the problem by that time. Don’t know why they memorize the tree.

For the people who talk about work experience in the past, there are also extensive guides for this sort of thing. They assume a simulation of what they did. Pretty cool. And incredibly exasperating on the other side.

I feel like the moral of this is to be better at cheating.

I mean, if the interviewee hadn’t cheated, he still would have gotten the esoteric sorting question wrong, and probably failed the interview because another person had happened to remember the esoteric sorting question.

So it was really a situation where there was (aside from ethical considerations) no incentive to not cheat.

If it were a situation where we were evaluating the candidate based on actual work related skills, this wouldn’t be an issue. Interviewer could develop a conversation with the candidate and have an interesting discussion about the problem that will give insight into how smart they are and if they can hack it.

That could have been done in this case as well, but instead of giving useful guidance, (“what if you thought about them as strings”) interviewer gave a meaningless clue. That is really just restating the problem parameters and has little to do with the implementation.

This is selection bias. Of course the cheaters who get caught are obvious. The question is whether the cheaters you don't catch are obvious. Academics say the same thing about students who cheat in university, apparently unaware of the vacuousness of the argument.

Man... I do about 2-3 screening interviews a week. Always fun when you hear them furiously typing on a mechanical keyboard when they are trying to answer how they did something on their last project. I've had people cheat by printing off common answers on the wall behind the camera. Last week had someone try having someone else respond to the questions while they were on video. Good times. Interviewing is a skill (for both sides) For something as important as hiring, I'll include my new dev leads/potential s as part of the hiring process to see how you suss out if a person might be a solid inclusion to your team. Tell me your greatest weakness or O notation gotcha isn't it...

> not exactly what I was looking for

Here lies the problem, as long as the interviews are about what you want and/or what I don’t know - I don’t care.

Anyone ask an algorithm question, I am just going to suggest using quantum bogo sort for everything.. Let the damn program run till the heat death of the universe. Then interrupt them during the next question, and try to find out how bad there backend really is..

I don't understand why an interviewer would create a sterile exam situation which is nothing like the real world or the job of interest. I'm supposed to be appalled by the interviewee but I'm more appalled by the interviewer.

This is very common in my phone screens

Anyone else think this was a half-ass attempt at cheating ?