7 ms·

> Hire someone right away and let them quit if they want to and hire someone else. It's just business for crying out loud. This is missing the point of interv

> Hire someone right away and let them quit if they want to and hire someone else. It's just business for crying out loud. This is missing the point of interviewing. The goal isn't to find any warm body to fill the chair, the goal is to find someone qualified to do the work who also has a history of doing good work at previous employers. You also don't have unlimited headcount and hiring budget, so it's worth making the investment to find the top 10% of your applicants rather than picking first-come first-serve. One of the things you don't realize about the hiring market until you've been reviewing resumes for a while is that problem employees are over-represented in the candidate pool. The most qualified employees spend the least time job searching because they're given offers right away. The most problematic and underqualified employees are frequently searching for jobs after being fired or let go. If you sample applicants at random, they're far more likely to be in the underqualified and/or problematic group than in the great employee group, statistically. The other thing that isn't obvious is just how damaging a single bad hire can be to a team. Hire someone who clashes with their peers and fails to deliver any good work and you'll find yourself losing the good team members very shortly. Nobody likes working with painful coworkers. That said: The analogy of a "walk-on" job isn't dead in tech. If you pick a company you want to work for, find someone on LinkedIn, and send them your resume with a short pitch about why you want to work there, there's a good chance they'll at least strongly consider your resume. Nobody is guaranteed a job this way, but it's one route to getting your foot in the door even when you don't see the exact job posting you want on the website.

primeblue 4y ago
mixmastamyk 4y ago
The GP was using the proxy of "shows up at 7:30AM, ready to work" as signal for motivation and a lesser extent, competence. Not a morning person, but would prefer this to leetcode hazing.
vgel 4y ago
The first thing you mention is a common problem in marketplaces: used cars, online dating, etc. Term is "market for lemons"
ChrisMarshallNY 4y ago
I actually very much agree with this comment. I was a manager for 25 years. A bad "team fit" was not good. I never had a technical failure in my hires, but I did have a couple of "bad cultural fits." These usually weren't toxic people, but people that couldn't handle the responsibilities and pressures (we were a small, high-functioning team, and everyone's visibility was fairly high). But this: > who also has a history of doing good work at previous employers. makes me wonder how LeetCode tests can tell you that, as they seem to be the single most important component of all software engineering hires, these days. In my experience, they just drive out the qualified people that can see projects through, and leave you with ... the ones that are really well-practiced in short, academic exercises.
- lordnacho 4y ago
  I'll second that. I never hired anyone who couldn't do the work. The only times things went badly were times when the person basically didn't want to do the work, due to some personal hangup. No amount of interviewing is going to weed out that guy who can code perfectly fine but deep down is yearning to be a psychologist instead. Like any other job that pays bills, you are vulnerable to paying his bills until they find what they really want. > In my experience, they just drive out the qualified people that can see projects through, and leave you with ... the ones that are really well-practiced in short, academic exercises. Yeah beats me how anyone thinks LC is useful, other than for weeding out the most unqualified people, like people who genuinely have never coded. I suppose what it really does is finds you people who are willing to put in the time to study all the hundreds of questions.
  - ianlevesque 4y ago
    >I suppose what it really does is finds you people who are willing to put in the time to study all the hundreds of questions. Yes, this is it.
    - ChrisMarshallNY 4y ago
      As opposed to putting in the time to learn how to write and release ship software. I won't study LC, because I'm waaaaaayyyy too busy, learning Swift, UIKit, AppKit, WatchKit, SwiftUI, DocC, MapKit, SiriKit, device SDKs, networking, USB, etc. I literally work every single day (like seven days a week), and learn something new every single day, yet I am barely keeping up. I would be nuts to sacrifice any of this time, studying schoolboy questions that have little to no relevance in the software that I write. These technologies result in actual applications that you can sell and market. Just another way to look at it.

primeblue 4y ago

mixmastamyk 4y ago

The GP was using the proxy of "shows up at 7:30AM, ready to work" as signal for motivation and a lesser extent, competence. Not a morning person, but would prefer this to leetcode hazing.

vgel 4y ago

The first thing you mention is a common problem in marketplaces: used cars, online dating, etc. Term is "market for lemons"

ChrisMarshallNY 4y ago

I actually very much agree with this comment. I was a manager for 25 years. A bad "team fit" was not good. I never had a technical failure in my hires, but I did have a couple of "bad cultural fits." These usually weren't toxic people, but people that couldn't handle the responsibilities and pressures (we were a small, high-functioning team, and everyone's visibility was fairly high). But this: > who also has a history of doing good work at previous employers. makes me wonder how LeetCode tests can tell you that, as they seem to be the single most important component of all software engineering hires, these days. In my experience, they just drive out the qualified people that can see projects through, and leave you with ... the ones that are really well-practiced in short, academic exercises.

lordnacho 4y ago
I'll second that. I never hired anyone who couldn't do the work. The only times things went badly were times when the person basically didn't want to do the work, due to some personal hangup. No amount of interviewing is going to weed out that guy who can code perfectly fine but deep down is yearning to be a psychologist instead. Like any other job that pays bills, you are vulnerable to paying his bills until they find what they really want. > In my experience, they just drive out the qualified people that can see projects through, and leave you with ... the ones that are really well-practiced in short, academic exercises. Yeah beats me how anyone thinks LC is useful, other than for weeding out the most unqualified people, like people who genuinely have never coded. I suppose what it really does is finds you people who are willing to put in the time to study all the hundreds of questions.
- ianlevesque 4y ago
  >I suppose what it really does is finds you people who are willing to put in the time to study all the hundreds of questions. Yes, this is it.
  - ChrisMarshallNY 4y ago
    As opposed to putting in the time to learn how to write and release ship software. I won't study LC, because I'm waaaaaayyyy too busy, learning Swift, UIKit, AppKit, WatchKit, SwiftUI, DocC, MapKit, SiriKit, device SDKs, networking, USB, etc. I literally work every single day (like seven days a week), and learn something new every single day, yet I am barely keeping up. I would be nuts to sacrifice any of this time, studying schoolboy questions that have little to no relevance in the software that I write. These technologies result in actual applications that you can sell and market. Just another way to look at it.