7 ms·

I've also done a lot of interviews for over a decade. I figure I'm an average or above average interviewer. Talked to people for an hour, discussed projects, th

I've also done a lot of interviews for over a decade. I figure I'm an average or above average interviewer. Talked to people for an hour, discussed projects, they looked good. Failed basic coding screens. Over and over. Maybe you've got an amazing sense for interviewing people. I doubt many others do and processes need to accommodate the average interviewer and not the amazing outliers.

pantulis 4y ago
I am not saying that all the people I hired were x10 performers, but they were certainly better than average, imho, with a leaner process. For example I did not hire anyone that lacked basic coding skills. Perhaps I am a great interviewer or, more probably, the fact that the specific profile I was targeting (RoR developers) included a high-pass filter.
- KajMagnus 4y ago
  No idea if this could be related, but @pantulis you're in Madrid (says your profile. Well, "from"), whilst @marcinzm you are (I'm guessing) statistically in the US, maybe SF? Is it possible that it works better to "just talk to people" in some parts of the world, than other parts? In SF, software engineer salaries can be $X00k/year, but a lot less in Madrid. I wonder if this can influence people's behavior. (Nonetheless personally I'd never do only such unstructured interviews.)
yamtaddle 4y ago
Raises hand Dude here who designed and built an early multitenant mobile app server-side support system from scratch, serving millions of daily users (the spikes in traffic when our users sent out pushes to some of those apps... yikes), among a bunch of other stuff. Actually, in-fact did that. Code-language-polyglot, often considered the one to go to for low-level stuff (nothing I seek out, I just keep getting that reputation), getting paid to write code for ~22 years now, et c., et c. I've nonetheless bombed "basic coding screens" a couple times, including after having done years of nontrivial and highly-independent coding for pay. And I don't mean leetcode shit, I mean really basic stuff. Some percentage of those folks who seemed great in the interview but failed your basic screens are false negatives, no matter how much you believe no one who had a tenth of a clue could possibly fail them. Maybe that rate's negligible, but maybe it's not. I do think unjustified confidence that very simple tests can't have a significant false-negative rate is part of why so many people are convinced our industry is drowning in expert-level bullshit artists who can't even write a for loop.

pantulis 4y ago

I am not saying that all the people I hired were x10 performers, but they were certainly better than average, imho, with a leaner process. For example I did not hire anyone that lacked basic coding skills. Perhaps I am a great interviewer or, more probably, the fact that the specific profile I was targeting (RoR developers) included a high-pass filter.

KajMagnus 4y ago
No idea if this could be related, but @pantulis you're in Madrid (says your profile. Well, "from"), whilst @marcinzm you are (I'm guessing) statistically in the US, maybe SF? Is it possible that it works better to "just talk to people" in some parts of the world, than other parts? In SF, software engineer salaries can be $X00k/year, but a lot less in Madrid. I wonder if this can influence people's behavior. (Nonetheless personally I'd never do only such unstructured interviews.)

yamtaddle 4y ago

Raises hand Dude here who designed and built an early multitenant mobile app server-side support system from scratch, serving millions of daily users (the spikes in traffic when our users sent out pushes to some of those apps... yikes), among a bunch of other stuff. Actually, in-fact did that. Code-language-polyglot, often considered the one to go to for low-level stuff (nothing I seek out, I just keep getting that reputation), getting paid to write code for ~22 years now, et c., et c. I've nonetheless bombed "basic coding screens" a couple times, including after having done years of nontrivial and highly-independent coding for pay. And I don't mean leetcode shit, I mean really basic stuff. Some percentage of those folks who seemed great in the interview but failed your basic screens are false negatives, no matter how much you believe no one who had a tenth of a clue could possibly fail them. Maybe that rate's negligible, but maybe it's not. I do think unjustified confidence that very simple tests can't have a significant false-negative rate is part of why so many people are convinced our industry is drowning in expert-level bullshit artists who can't even write a for loop.