What if companies interviewed translators the way they interview coders?
Jose Aguinaga

Okay, finally commenting here after about a dozen people have shared this with me.

To start off with: Yes, I get the humor here and that it’s satire except that it’s also intending to make a point.

But it’s also… intending to make a point, and that point is pretty significantly off base.

I’ve never heard of a company asking about the history of computer architecture or algorithms or anything like that. I’m sure there is some company out there doing that, and that would of course be ridiculous. That is not the “classic software interview” is like.

Of course, that’s not actually the analogy this article was trying to make. This article was assuming that the classic whiteboard interview is totally unrelated to the job, so it substituted it with another totally unrelated topic and a totally different job, and in order to show that, well, the interviews are unrelated to the job. This is the begging the question fallacy. You’re presuming the very conclusion you’re trying to prove.

Here’s an updated analogy: The candidate is given a piece of text to translate that has multiple interpretations (the trade-offs of which will be discussed with the interviewer), is significantly more complex than the typical work, and is about a different side of medicine than this particular translator would be dealing with. The candidate is also not given a dictionary, largely because it’s not necessary for the task and the interviewer will overlook any little issues.

Does that seem quite so ridiculous?

Sure, they could have just asked the candidate about his/her skills. But many people oversell themselves and many other people undersell themselves. It’s really better to actually assess people’s skills.

Sure, they could have given the translator actual medical labels to translate and just left him in the room for an hour. But the issue there is that you might be too biased by someone’s current skillset (the candidate with a background in medical labels has an advantage, even though a good candidate could learn that quickly), and easy tests clump “okay” and “good” scores together (and minor, random mistakes can make people look worse than they are).

If you want to identify people who are great, asking challenging questions and actually assessing their questions is a pretty good way to go.

Whiteboard interviews have their flaws (and they have lots of companies who screw them up), but this article hasn’t hit on them.