Conversely, let& #39;s talk about one of my favorite questions.

"You find yourself on a box. Maybe it& #39;s Linux, maybe it& #39;s Unix; it& #39;s definitely Unix-like. How do you figure out what operating system you& #39;re running?" https://twitter.com/QuinnyPig/status/1265286980859908102">https://twitter.com/QuinnyPig...

The only single canonically correct answer to this is "uname." If the candidate says that, you modify it with "someone has removed the uname binary."

Now here& #39;s what I love about this question: there& #39;s no longer one correct answer. There are literally hundreds of them.

It shows how folks think about systems, which is really the point of the question. Networking people think network stack, ops people think package management, Serverless people don& #39;t think at all, they just condescendingly tell you it doesn& #39;t matter, etc.

I like to ask it early in the process, because it& #39;s an icebreaker question. The only "wrong" answers are giving up, or something that& #39;d amount to an HR violation.

It gives me a chance to be super encouraging; it impresses upon the candidate that this isn& #39;t a me vs. them moment.

People are already responding with a bunch of answers. I& #39;ve seen all of them and more so far. And that& #39;s great!

So far the most creative I& #39;ve seen was "I& #39;d use nmap to fingerprint the box." That one was awesome!

"I& #39;d poke around in /etc/." Great! What& #39;re you looking to see? What does that tell you about the box?

"I& #39;d check /etc/issue." What if that file doesn& #39;t exist?

"I& #39;d turn it off and see who shrieks about the X box. X will tell me what OS it is!" Points for creativity!

One of the best answers I& #39;ve ever gotten was "I& #39;d ask a coworker. Why reinvent the wheel when there& #39;s already someone right there who probably knows?"