Hacker News new | ask | show | jobs
by mattlondon 406 days ago
Yep. It started with COVID where understandably 100% of interviews were remote.

But now with COVID a thing of the past, for "fairness" reasons (DEI?) we still do 100% remote interviews, but now have the ludicrous situation where we're asking interviewers to do absurd things like look for the reflections in the candidates' eyes/glasses to see if they're using ChatGPT, ask the candidate to swing the webcam around to make sure there are not other people in the room, ask them to hold their hands up to the camera to show they're not typing a prompt (which is even more stupid than it sounds because voice recognition is amazing these days), or ask them not to look away from the camera when answering questions (so not reading answers from another monitor) and other stupid things. How ridiculous.

The sooner we get back to in-person interviews the better. Get them to come to the office (which they'll need to do one day if they get the job) and sit next to them while they code on a work laptop).

Sorry to all those folks who want 100% remote, but this is why we can't have nice things.

5 comments

And similarly forbid them from using AIs while they code on that work laptop in person? Are employees forbidden from using AIs for work? If not, why require that during evaluation? If it's not required during evaluation in person, why require it remotely?

(I don't know the answers to how to interview in this brave new world, but I'm increasingly skeptical of forbidding tools that people will be using for the job.)

Because job interviews don't test real-world programming skills, which is a whole other issue.
The closer you can get to doing so, the better.
I think the best interview question, and really the only one you need to determine technical ability is ask someone to describe a http request in as much detail as possible.

To write code (even with the benefit of AI) effectively you need a mental model of the systems you work with, reading the chatGPT response doesn't prove you have that.

That's a stupid interview question for the vast majority of software jobs. Many people don't work with HTTP or web software at all.
So replace it with something from the relevant field.
Yes, technical interview questions should be relevant to the job field. What's your point?

The hard part is selecting good questions that act as reliable predictors of actual job performance. Very few hiring managers can do that reliably, although many fool themselves into believing that they can.

The point is that someone gave a specific example of the much more general concept of probing for mental model by way of detailed explanation of a process he ought to be familiar with. You objected to the specific details - knowledge of HTTP. That's not an indictment of the general approach.

That said ML models have gotten to the point where I'd have to disagree with OP that this approach will necessarily filter their use. However there are plenty of available mitigations, from latency of response to requiring a video feed that fully covers the candidate, his screen, and his keyboard.

The "what happens when I enter facebook.com in the browser and hit enter" is/was a well-known FAANG question a few years back so I would expect that all the LLMs are well versed in it, as will be NK infiltrators
If you want to work as a clerk at Target, the video is not even an interview, it’s a one-way audition you record to be judged anonymously.
My suspicion is that it's purely monetary and driven by the finance people.

a) Don't have to pay to fly candidates out, pay for their hotel, etc.

b) Don't have to pay relocation

c) Get access to a larger pool of candidates, so can price the wages lower than local wages would require

My last company there was a top down directive that in-person interviews were straight up not allowed, everything had to be over Zoom. Even for local candidates, for a job that was supposed to be in-person! Completely crazy IMO.

The advantage of a larger pool of candidates is not mostly a financial benefit, IMO. The benefit is mostly the ability to hire from a larger pool of people especially with a specialized skillset, and also to have less of an echo chamber.

But yes, that directive to interview local candidates over zoom does seem very silly.

My experience is that yes it opens up the wider pool, but it makes the filtering process much more difficult in trade.

Opening up the wider pool without the in person interview is where things hit the wall since the filtering criteria everyone learned over their careers went out the door thanks to the online interview process. And the online interview process is much more subject to cheating--not exactly a huge concern in-person.

What is the local pool like?

If you want a software engineer silicone valley you can stay all local. There are companies in remote small towns who need a software engineer - they have to open up to non-local candidates as there are zero people in town who could do the job that don't work for them. There is always someone from elsewhere excited to move to a small town, but finding those people is hard. (and for those people finding a company that wants them is hard)

I'm in a tech hub so the local pool is wide and deep. But I was flown up for an interview, and I've known other people who were flown out, were hired, and have become locals.

This didn't used to be a huge problem.

I agree that there's a trade-off in filtering, but I really just don't resonate with this "cheating" issue.

I haven't run into this thing where I'm talking to a video AI, but maybe I'll sing a different tune if that ever happens and is high fidelity enough to trick me.

If "cheating" just means using AI assistants to answer my interview questions, honestly I think I've done a poor job structuring the question and interview.

I do recognize this as a giant challenge right now, to structure interviews in a way that provides real signal, while allowing candidates to use the tools they'll actually be using for the job. But I don't think the challenge is significantly different between remote and in-person.

Only a) is valid, as you can fly candidates for interviews and have them go back to their home city to work remotely.
Yeah after a disastrous remote hire I started requiring in-person 2nd round interviews. Company policy is that all future hires are hybrid only (not that we or anyone else is hiring these days...) so it just makes sense.

For developers I share my screen on MS Teams so everyone can watch, then hand them my laptop with Visual Studio. They've got 90 minutes to complete a small assignment while we look at them code - Google is allowed, so is copying and pasting from Stack Overflow, and we'll probably allow Copilot as well. The code needs to run and return the expected results. One candidate said, "this was great, it felt like real work".

For cloud admins, our Devops lead creates a new resource group, hands over his laptop, and we ask them to create a few resources and do the network and authentication to make them talk to each other. Most candidates can't do that anymore - we're finding they've become Terraform operators that don't know how the underlying technology works.

COVID isn't in the past, just no one doing anything about it. :)
The 1918 Pandemic is still around, too... A/H1N1
Different species, you can't generalize like that. It's pretty unclear what actually happened with H1N1. Scientists were able to resurrect the more virulent strain in the lab two decades ago and it was just as potent in lab animals...

Two possibilities are that it did in fact mutate to become "milder" or those strains were already circulating. Either way, H1N1 killed so quickly it ran out of victims and the highly lethal strain went extinct. Another notable aspect of H1N1 is that is mostly didn't kill directly, it made victims weaker to opportunistic lung infections and that's what killed them. Antibiotics have made this kind of attack vector much more difficult for viruses.

Omicron is only loosely analogous to the "flu fairy tale" as the major threat is Long COVID now and it is circulating at high levels. Other viruses have had vastly different natural histories, 1918 is only a single reference point, and a muddy one at that.

I was under the impression that it mostly killed by causing cytokine storms.
I won't claim to be an expert on 1918 influenza, but here's a reference for the claim I was making: https://pmc.ncbi.nlm.nih.gov/articles/PMC5481322/

After reading this, I am less sure of the claim "most", but it seems that opportunistic infection was an important factor. There were a few unpleasant ways to die...