Hacker News Reader

Comment by foobaw | Hacker News Reader

Most of the responsibility, minus the technical interview, seems like it could be done by an AI sometime in the future anyway. No more complaints afterwards!

View on HN

I found Google's dev interview to be especially difficult, I personally hate doing hours of whiteboard coding (sorry, I tend to code on computers instead, silly me) ... So I wasn't hired, but those who worked as devs there told me that's a blessing because after all the rigorous interviewing bluster they'll assign devs to remedial code refactoring anyway. I'm not a rockstar coder above doing the housekeeping work but doesn't sound fun either. Why hire top coders then assign them intern busy work? I wonder if the strategy was to just drain the local talent pool of top coders so competitors had less of a candidate pool. Just odd to set the interview bar so high, and for what purpose?

You leave your interns to do code quality refactors? Seems like something more experienced engineers should do.

I don't work at Google, but code refactoring is critically important in the real world. It's not intern work.

Re-factoring is probably 80% of the job.

New projects are rare, so very few are writing a lot of code from scratch.

Why not, like, test refactoring skills? Mind blown!

If that's the attitude you brought to the interview, then perhaps that's why you were not hired?

What percent of software engineers' time do you think is spent doing green-field development with no refactoring, writing tests, documentation, etc.?

I actually find refactoring legacy code to be one of the most challenging things I do at work.

While I definitely wouldn't want to do that for all 40 hours of work I wouldn't classify refactoring as "remedial".

> Just odd to set the interview bar so high, and for what purpose?

Check Google's stock price.

FWIW, interns on the team I interned on had the chance to work on experimental, new features that required little to no refactoring. Maintaining existing code was done by developers who knew the code well and were capable of making measured engineering decisions.

Just yesterday I had an interview with a Facebook recruiter and afterwards all I could think about was that the recruiter's entire job could have been replaced by an AI. And not one of the smart, modern, cutting edge Siri types of AIs, either. The entire process could have been just as easily accomplished by a robocaller bot that's been around for decades.

The interview process was a few generic, form-filled emails to schedule the interview time. Then during the interview call the interviewer read off a few pre-determined questions that were read directly from a questionnaire, and the interviewer recorded my answers. They didn't ask any follow-ups, only asked what was on the questionnaire. When we were done, I tried to ask them questions about the position and they told me "I can't answer those questions for you because I'm just a recruiter and I don't know details about the specific position. Hold onto those until you talk to the next person."

He might as well have just put on a robotic tone and said "My responses are limited. Please wait to talk to a real person if you have other questions."

It could have been automated, but having the human make the call leads to a slightly decreased level of people madly trying to google up answers to 'what is the expected exit code for a program that had no errors?' so that the next phase of the process is not a complete waste of time for all parties involved. The recruiters are just trying to see if there is any chance that you might actually have the skill set claimed on your CV; if you can't answer their simple questions then your resume is complete BS and they will move on, but after asking the same questions over and over and over the recruiting team also knows the response of someone for whom the questions are insultingly easy and can move those candidates to the front of the queue for the next stage.

Yea, that's all well and good and I used to do recruiting for my old company where that was the case, except that wasn't even what was happening here. The Facebook recruiter himself told me that he was simply recording my answers down to be reviewed by someone else. He didn't seem to know what a "good" answer was versus a "bad" answer, and he didn't care. The questions he asked weren't about nor related to my resume, they were just generic "tell me about a time when you worked on a team with other people" type questions, and the recruiter wasn't really more than a notetaker for my responses.

I get that having a "human element" makes the process seem slightly more personal, and I would hate to be a part of a recruiting process that literally is just robots, but in the case where your job can literally be completely replaced by nothing more than an online HTML form or a robocall, what are you even doing?

Actually, I'm wondering if an AI could be better at predicting the success of a given candidate given enough data (resume, online presence, portfolio, DNA sample, etc.) than "traditional" whiteboard interviews (which have a huge false negative problem).

I know someone in just such a startup: https://ideal.com/

They've been at it for a few years, they started with a focus on candidates for sales positions. I don't know if it's generalised or not.

Reply on news.ycombinator.com