The Deuce

The Deuce
[ The Deuce ]

This is my first round of hiring and interviewing with this gen of AI and boy is it a problem

Posted by ISeeDeadHawkeyes User icon on October 5, 2026, 09:20:04

We're getting resumes that were clearly generated, including required certifications that the candidate didn't actually possess.

During interviews we've also had candidates clearly using some sort of AI tool to answer questions. When you ask for specific work experiences from their resume, there's a pause and then they just give a high level overview of how the software works as opposed to a specific experience.

Probably an unfortunate reality of conducting interviews remotely.

Are any of you seeing, or using AI to winnow applications?

Posted by memyselfandillini User icon on October 5, 2026, 15:33:51, in reply to "This is my first round of hiring and interviewing with this gen of AI and boy is it a problem"

Previous Message

"Did somebody say 'boy'?" -- Dennis Hastert's Ghost

Posted by doubledown User icon on October 5, 2026, 11:31:58, in reply to "This is my first round of hiring and interviewing with this gen of AI and boy is it a problem"

Previous Message

timely topic. i was on an interview this morning and the candidate withdrew 10 min into it

Posted by lcubed User icon on October 5, 2026, 10:11:15, in reply to "This is my first round of hiring and interviewing with this gen of AI and boy is it a problem"

could tell they were nervous at the beginning and they acknowledged it, so i had hope, but it didn't get better and they apologized and withdrew from consideration. felt bad for them

i also think the first person we interviewed in the search was wearing meta or some sort of smart glasses

Previous Message

I dream of being able to do that

Posted by Sounder User icon on October 5, 2026, 10:39:18, in reply to "timely topic. i was on an interview this morning and the candidate withdrew 10 min into it"

Not the getting nervous part, but we've all been in interviews where you know pretty early on that it's a waste of everyone's time to even be there, so it would be awesome to be able to pull the plug.

Previous Message

AI is one thing, but straight up lying about having certifications is another.

Posted by Warrin'_Illini $<iframe src='https://mb.boardhost.com/TheDeuce/info/user=Warrin\'_Illini' width='330' height='300' border=0 style='border:none;' tabindex=0>$ on October 5, 2026, 09:30:40, in reply to "This is my first round of hiring and interviewing with this gen of AI and boy is it a problem"

I would start making a (naughty) list of people misrepresenting themselves and spread it far and wide in your industry.

Previous Message

I said forget about it cuh

one really bad habit with AI is people not checking its work. see: all those stories about lawyers*

Posted by torinese User icon on October 5, 2026, 09:34:30, in reply to "AI is one thing, but straight up lying about having certifications is another. "

Previous Message

I would start making a (naughty) list of people misrepresenting themselves and spread it far and wide in your industry.

Previous Message

Heh...speaking about that, I asked Perplexity a question about the copyright on Winnie the Pooh

Posted by ATL User icon on October 5, 2026, 10:09:55, in reply to "one really bad habit with AI is people not checking its work. see: all those stories about lawyers*"

yesterday and as it turns out....it's first answer to me was wrong. Specifically, it said that the character Tigger was still under copyright while Winnie was not. I asked it a follow up and it corrected itself.

Not sure why it got it wrong at first and then corrected itself within a couple of minutes but it got it right.

Previous Message

I would start making a (naughty) list of people misrepresenting themselves and spread it far and wide in your industry.

Previous Message

“We are old and tired and just want to be taken care of.” - Republican voter, Dec. 2024.

Did your follow-up question the correctness of the first answer?

Posted by doubledown User icon on October 5, 2026, 10:57:15, in reply to "Heh...speaking about that, I asked Perplexity a question about the copyright on Winnie the Pooh "

Also, what model were you using?

A lot of models will respond without researching so you get an answer fast, though Perplexity's value is supposed to be that you always get back grounded answers. A follow-up that says the answer seems wrong typically calls a tool that does a more thorough search.

Tigger entered PD at a different time than the other characters, so it's likely that the model was trained on content that mentioned Tigger not being PD yet.

Previous Message

I would start making a (naughty) list of people misrepresenting themselves and spread it far and wide in your industry.

Previous Message

The follow up question was: “why is tigger’s copyright not expired” so it was a request for a reason

Posted by ATL User icon on October 5, 2026, 11:29:19, in reply to "Did your follow-up question the correctness of the first answer?"

not questioning the accuracy of the first response. I’m using the “best” model - the perplexity default.

Previous Message

I would start making a (naughty) list of people misrepresenting themselves and spread it far and wide in your industry.

Previous Message

“We are old and tired and just want to be taken care of.” - Republican voter, Dec. 2024.

Yeah, it likely called a tool that engaged a more capable model, or did a search. It corrected

Posted by doubledown User icon on October 5, 2026, 11:34:50, in reply to "The follow up question was: “why is tigger’s copyright not expired” so it was a request for a reason"

itself based on the result.

Previous Message

not questioning the accuracy of the first response. I’m using the “best” model - the perplexity default.

Previous Message

I would start making a (naughty) list of people misrepresenting themselves and spread it far and wide in your industry.

Previous Message

Yeah, I had this problem when I was applying to jobs last year

Posted by DT User icon on October 5, 2026, 10:02:35, in reply to "one really bad habit with AI is people not checking its work. see: all those stories about lawyers*"

I tried using AI to customize my resume for job postings but it would regularly just make up experience to match the posting, even though I was prompting it explicitly to not include anything that wasn’t in my base resume.

It ended up taking more time to use AI than just to do it myself, given the hallucinations.

Previous Message

I would start making a (naughty) list of people misrepresenting themselves and spread it far and wide in your industry.

Previous Message

Yes indeed. You know when you are reading it too.*

Posted by Warrin'_Illini $<iframe src='https://mb.boardhost.com/TheDeuce/info/user=Warrin\'_Illini' width='330' height='300' border=0 style='border:none;' tabindex=0>$ on October 5, 2026, 09:37:18, in reply to "one really bad habit with AI is people not checking its work. see: all those stories about lawyers*"

Previous Message

I would start making a (naughty) list of people misrepresenting themselves and spread it far and wide in your industry.

Previous Message

I said forget about it cuh

what i don't understand is how these people are getting to speak with you in the first place.

Posted by torinese User icon on October 5, 2026, 09:28:56, in reply to "This is my first round of hiring and interviewing with this gen of AI and boy is it a problem"

it might require some more work upstream, but those should be filtered out relatively easily?

Previous Message

Is ISDH selecting the candidates or is HR using AI to give him a list?*

Posted by RVA User icon on October 5, 2026, 09:35:17, in reply to "what i don't understand is how these people are getting to speak with you in the first place."

Previous Message

it might require some more work upstream, but those should be filtered out relatively easily?

Previous Message

Yeah I am not selecting these candidates myself

Posted by ISeeDeadHawkeyes User icon on October 5, 2026, 10:03:34, in reply to "Is ISDH selecting the candidates or is HR using AI to give him a list?*"

I'm a pretty narrow corner of the data world, and there's a lot of different roles and ways to do the work in this corner.

Previous Message

it might require some more work upstream, but those should be filtered out relatively easily?

Previous Message

I think torinese is saying that HR isn't doing its job

Posted by ATL User icon on October 5, 2026, 10:17:12, in reply to "Yeah I am not selecting these candidates myself"

(autofill)

Previous Message

I'm a pretty narrow corner of the data world, and there's a lot of different roles and ways to do the work in this corner.

Previous Message

it might require some more work upstream, but those should be filtered out relatively easily?

Previous Message

“We are old and tired and just want to be taken care of.” - Republican voter, Dec. 2024.

We're getting 500+ applicants for each role

Posted by ISeeDeadHawkeyes User icon on October 5, 2026, 10:25:26, in reply to "I think torinese is saying that HR isn't doing its job"

It's really hard to explain to people that dont do this work, but I can see why these candidates were selected for interview.

We're a team of developers - these candidates have been analysts so they're one step downstream from the work we do, but have relevant certifications

Previous Message

(autofill)

Previous Message

I'm a pretty narrow corner of the data world, and there's a lot of different roles and ways to do the work in this corner.

Previous Message

it might require some more work upstream, but those should be filtered out relatively easily?

Previous Message

Write your own filtering agent. It won't be perfect, but you can make a better bloom filter

Posted by illinibone User icon on October 5, 2026, 12:14:51, in reply to "We're getting 500+ applicants for each role"

than accept everything

Previous Message

(autofill)

Previous Message

I'm a pretty narrow corner of the data world, and there's a lot of different roles and ways to do the work in this corner.

Previous Message

it might require some more work upstream, but those should be filtered out relatively easily?

Previous Message

Step 1 - install tampermonkey: https://www.tampermonkey.net/index.php
Step 2 - install this script: https://greasyfork.org/en/scripts/577609-boardhost-idiot-hider
Step 3 - Add/modify user list

You're right to push back on that. *

Posted by doubledown User icon on October 5, 2026, 09:30:05, in reply to "what i don't understand is how these people are getting to speak with you in the first place."

Previous Message

it might require some more work upstream, but those should be filtered out relatively easily?

Previous Message

We are doomed as a society*

Posted by Sounder User icon on October 5, 2026, 09:26:06, in reply to "This is my first round of hiring and interviewing with this gen of AI and boy is it a problem"

Previous Message

Like, they're reading answers off their screen? *

Posted by doubledown User icon on October 5, 2026, 09:23:05, in reply to "This is my first round of hiring and interviewing with this gen of AI and boy is it a problem"

Previous Message

Correct*

Posted by ISeeDeadHawkeyes User icon on October 5, 2026, 10:01:33, in reply to "Like, they're reading answers off their screen? *"

Previous Message

Why should this surprise anyone, this is what we are teaching people to do*

Posted by Sounder User icon on October 5, 2026, 09:28:14, in reply to "Like, they're reading answers off their screen? *"

Previous Message

« Index

323

[ The Deuce ]