Once worked at a 3rd party coding interview company as an interviewer, and we had a bunch of grad students from a college who were all clearly cheating. They solved the problem from the top of the page down (rather than organically, as someone would when doing actual programming) in a very specific way.
The company didn't want to disqualify those candidates, since we couldn't prove cheating, but it was pretty fascinating to witness.
Fortunately our company gives us the option to reject a candidate if we feel they are cheating and can provide an explanation. Once had a candidate solve a problem extremely quickly but they could not explain how they solved it all. I then made a small modification to the problem which wouldn't have changed the solution much and they were completely lost.
This is the way. I believe a simple 45-minute exercise one-on-one between a programmer and a candidate is all you need to gather the necessary information to hire. Simple followup questions like, "Before you start coding, talk me through your thought process and let's discuss a few implementations", "What's another way to solve this problem?" will weed out any bullshitters, and even find hidden strengths in candidates that would never be discovered in an automated Leetcode type interview.
Real discussion between two programmers, sounds actually like real interview. Unlike robotic leetcode type: “problem prompt-memorized solution from LC.com-memorized BigO answer” type interview
I feel that most avid readers of the green book would behave like that. If you consider that cheating its probably better to find another interview method.
In my experience that isn't accurate. There are a lot of folks who studied Hacking the Coding Interview and are able to use the skills they've gained, but who aren't willing to cheat in the process.
Probably, but small instances of cheating aren't going to push the curve all that much. We were mostly functioning as a first layer as well, so if the likely cheaters couldn't perform, they'd still get filtered out during the companies' onsites. At least that was the rationale.
The company didn't want to disqualify those candidates, since we couldn't prove cheating, but it was pretty fascinating to witness.