No, it was over video. The person was verbally reciting whatever ChatGPT gave to them. Their answers were verbally "I apologize for any inaccuracies, blah blah blah".

They were typing in your questions? That must have been pretty slow and obvious.

There's a button on ChatGPT in my iPhone app that allows me to switch it to microphone mode rather than needing to type it. An off screen tap and you can get the results.

I'm sure that some people have hooked up the whisper model so that it can tap a desktop audio out and feed ChatGPT too.

Typing during an interview is pretty common since many people (interviewers and interviewees alike) take notes during interviews, so the typing alone didn't stand out as unusual.

I once concluded an interview with a candidate who was rather conspicuously engaged in clandestine typing, subsequently rectifying their previously incorrect answers. I suspected they were resorting to Google, given this was before ChatGPT came out. Much to my surprise, they ended up landing the job.

They got hired even though you said "no"? That sounds extremely unusual, typically employers of SDE types are cautious this way.

I don't know what exactly happened there. The two of us conducting the interview advised against this candidate in our report. About a month later the candidate was introduced as a new hire on a different team and I never saw them again in a company of <100 people.

Arguably if someone can google the “answers” to your interview questions in 20 seconds they aren’t good questions to ask anyway.

And if someone's ability to answer questions using search tools in real time is sufficient to impress a series of interviewers, they're probably quite resourceful at finding solutions to problems they've been given a few hours to solve

That is pretty depressing. It's funny how easy it would be to not give yourself away too. It's ridiculously easy to spot some of the general ChatGPT tells and simply not recite them.

Even without the answers, I can't imagine I wouldn't realize by the pace and format of the responses that they were simply read to me instead of said.

I've never had such experience and it seems very low effort.

In any case there's absolutely no risk of someone passing an interview with ChatGPT on our end because it's about seeing how people think and what they've experienced.

Automattic does this.

https://automattic.com/how-we-hire/

Y'know, that's surprisingly on-brand for the company that owns tumblr, a website full of socially anxious introverts who would rather perish than make a phone call.

That's me! https://imgur.com/a/Zbv5da8

Anyone who does not choose death over picking up the phone is a true psychopath.

Also weird that they just repeated back exactly ChatGPT. How lazy can you be?

You'd be surprised. After all, if we were not lazy we'd be still digging the ground with our bare hands. "How do I get X while doing less work" is a kind of laze that defines human.

The engineer’s three virtues: Laziness, Impatience, Hubris

https://wiki.c2.com/?LazinessImpatienceHubris

laziness can be a good thing! Real lazy people put in a lot of effort to achieve the same outcome working less!

Sounds as though you've read the page before.

Procrastination too. I've had things go off the rails because procrastination wasn't allowed to manifest opportunities.

> "How do I get X while doing less work"

And we'd still be using neanderthal tools if everyone just copy-pasted previous work.

Reminds me of Ron Burgundy!

Absolutely. Reminded me of Chris Pratt's famous P&R flu scene ad lib "Leslie, I typed your symptoms into the thing up here and it says you could have network connectivity problems".

Many years ago I had someone copy and paste two pages from an Oracle manual, complete with formatting, in response to an e-mailed screening question about MySQL that required a one line answer. It was just an initial screening where we had relatively low expectations and halfway expected a lot of people to search - letting a few too many people through was ok, we just wanted to whittle down the pool.

But after that we realised the screening worked even better when we input random sentences from the answers into Google and used it to weed out the people who were not just dishonest enough to give a verbatim answer from somewhere without disclosing it, but who very often did so in ways that meant they got it horribly wrong. E.g. another example was the guy who copied and pasted a forum post. Not only was the answer wrong, but it was followed on the forum by a large number of responses from other people explaining why it was wrong.