Yeah, hiring is scary. Hiring is insanely expensive on all fronts. Firing people is difficult, it's expensive and legally exposing. Hiring the wrong person, allowing them access your systems and potentially exfiltrate your IP to them is a hazardous but necessary venture.
The thing is, none of these things really changed with AI. People have been lying about their experience for literally centuries. IMO the advent of AI-laden candidates is going to nudge the hiring process back to how we did it 10 years ago, with a good old fashioned face-to-face interview and whiteboard questions. This means a lot of things that we've grown accustomed to in the past 5 years is going to have to melt.
- people are probably going to have to fly out for interviews, again.
- awkward and/or neurodivergent people are going to have to learn social skills again.
- And yeah, you guys, it's time to buy a suit.
Companies should consider reverting to forking the upfront $13-1500 dollars for a set of plane tickets for their hiring team and rented conference rooms for a week. It's a whole lot cheaper than spending 50k because you hired the wrong person for half a year.
Others have already commented on this, but do you work in tech? IME getting interviewed by directors and even VPs in t-shirts is the norm. I’ve worn jeans to work my whole career. If anything, I think people in tech have a strong prejudice against people in suits (ie “obviously this person isn’t a real software engineer, they’re wearing a suit.”)
Anyway, probably not good career advice to wear a suit unless dress codes at tech companies are suddenly subject to drastic changes.
I always get a positive response.
I got the job, but was then told "don't listen to your mother"!
> - people are probably going to have to fly out for interviews, again.
Fly where? Many companies don't have offices anymore.
> - awkward and/or neurodivergent people are going to have to learn social skills again.
Ahh... the age old, "just do better" position for neurodivergent people. Classic bigotry.
> - And yeah, you guys, it's time to buy a suit.
Suits were out in tech 30 years ago when I first interviewed. They have only gone more out of style. Fashion doesn't work the way you think it works.
My guess is that we'll see more contract-to-hire positions and "talking through code" style interviews. Though I think we'll see lots of things tried which will be a general improvement over what much of the industry was doing before.
Hiring is all about finding the best candidate. If you find you cannot function sitting in conference room with three other people for an hour, there is a 100% chance there is a better candidate suited for the role, even if his/her technical skills are less than yours.
Jobs have soft skill requirements, and there is nothing bigoted about that.
I think you're too eager to throw personal attacks on those who raise valid points that are you feel are uncomfortable to address.
You should be aware that engineering is a social activity that requires hard skills. In any project that employs more than one person, you need to be able to interact with others. This means being able to effectively address and interact with others around you.
If you give anyone a choice, anyone at all, on who they work with, they will of course favor those who they are able to effectively interact with.
This is not bigotry, is it?
While i dont agree with the idea we'll be flying anywhere for interviews, havent most companies gone back on remote work. "hybrid" is a benefit now and being in the office is the expectation.
Presumably to meet the boss. And maybe the key people on the team.
Or maybe it works exactly the way they think? Suits are so out, that wearing one is a strong signal of "different thinking" in a way that being casual once was. A colleague of mind would wear a three-piece on "casual Friday", and always showed up to the nines for interviews. Never harmed him, just reinforced his "think different" bona fides.
I agree that lying was possible before AI, but something about AI has emboldened a lot more people to try to lie.
Something about having the machine fabricate the lie for you seems to lessen the guilt of lying.
There's also a growing sentiment online that using AI to cheat/lie is "fair" because they think companies are using AI to screen candidates. It's not logically consistent, but it appeals to people who approach these problems as class warfare.
How is it not logically consistent?
At my previous employer, I had to convince several people in my team that wearing a suit was NOT a reason to reject a candidate out of hand. It's really difficult to gauge the expected dress code at a company beforehand, but it's not good advice to just blindly dress up.
Me personally, I like working at places where people can wear shorts and flip-flops. One level up is "pants and shoes, with socks", not even a collared shirt. Maybe a single-color new T-shirt, to be safe. A full suit would be an alarmingly bad read of the culture, and at that point we'd have made you come eat lunch at a burrito place to get a read on whether you're really a bad fit or just socially awkward.
The best thing I heard from an interviewee that was wearing a suit was that they interviewed elsewhere nearby that morning, and those people needed to be impressed with clothing.
I could even go further and say that NOT hiring anyone who shows up in a suit will give you better results than the other way around. You filter out a lot of career guys who are really poor programmers and will try to end up as mediocre middle management that way.
That said, I very much agree with your last paragraph. In the late 1990s and early 2000s, a lot of hiring was done this way in the US.
In an American context, this is generally true in 49 out of 50 states, except that the probation period covers the entire duration of your employment. The people who say firing is expensive are thinking about something else.
And I’m a hiring manager. I’m trying to slot new hires with the training they will need and give them realistic tasks I know they can accomplish. And it’s not easy. I’m already 30 days in on a new hire that I’ve been able to peer with for 2 days. And I’m constantly apologizing for the lack of time.
It's not like these are skills that they haven't learned, these are things that they have a hard time with. Expecting them to be 'normal' is like asking a person of medium stature to be taller. They could mask them but ultimately it's not who they are and expecting everyone to be the same is a fools errand.
I do agree that there’s no reason face to face interviews shouldn’t be the norm again after an initial screen.
If some of those things don’t appeal to some candidates? <shrug> I don’t totally mean that. But some practices should be the default even if some candidates don’t really like them (and even if they’re less convenient or more costly for hiring managers.)
Not sure about the suit at a lot of tech companies but dressing neatly and even throwing on a sports jacket probably doesn’t hurt.
Employers didn't have a whole lot of choice in that matter for a long time. Candidates wouldn't show up if you tried to impose that upon them.
Granted, nowadays it does appear that the tide has turned back to employers getting to call the shots, especially for lower-level positions. It is less clear how desperate the top talent is.
Millenia. Just ask Nanni what happened when he trusted Ea-nāṣir.
https://en.wikipedia.org/wiki/Complaint_tablet_to_Ea-n%C4%81...
There are very few companies I'd fly out for TBH.
IMO Make firing easier, pay people a massive severance if you're firing them for a mistake you made in hiring, and initially start them out as remote so you're not forcing a lifestyle change for them if you realize you made a mistake.
They didn't fire many people quickly, but it had a deeply chilling effect when someone was only at the company for a month or two before disappearing.
One of the unspoken difficulties of firing fast is that the person does a lot of relationship building with people who don't work with their output. It was often the case that someone would become well-liked by people who never saw their code, who would then become distraught when the likable person vanished one day.
No, business casual is just fine. Who wants to try to do a grueling technical interview in a suit? No thanks. I sweat enough as it is in interviews.
I worry about the retreat to networks. I think it's an inevitable response to the rise of machine-generated fakes, that people are going to start strongly preferring to be physically next to someone talking to them simply in order to verify that they're real, not one of the billions of apparitions knocking on their virtual door. But it also pushes back to networks of preferred universities and preferred drinking societies within those universities. All of which have the opportunity to be little discriminatory clubs.
We just need some sort of qualification which tests practical knowledge.
The real answer here is: know the dress expectations of the place you’re applying to, and consider whether you feel comfortable working for a place that won’t pay you the courtesy of letting you dress yourself.
Whatever future interviews look like, I sure as hell hope we don't maintain this ^ attitude.
No company worth working for would refuse to hire someone just because they didn't wear a suit to their interview
Dress nicely, sure. Wear a collar? Yeah probably. A tie? Meh.
Let's get rid of this old fashioned boomer nonsense from hiring please
I agree no one will explicitly decide one way or another based on how you dress. But making everyone in the room feel comfortable with each other will help the whole process.
No. Those are costumes that benefit no one but the seller of the costume. They wear the costume precisely once and never put it on again. It's an old classist ritual that forces people to spend money on clothes they dont want or need.
The problem with hiring in IT is that it's a bit of a broken system. On one hand you have companies that are overly picky and are alienating good candidates with their convoluted/mildy unfair hiring processes. Being overly picky when it's hard to get good candidates to show any interest because they are in demand is of course counter productive. This has been the default for the last decades. Lots of demand, not a lot of supply of great candidates.
And then on the other hand you have the recruiter / bodies for hire market with candidates that are maybe not that great being lined up with opportunities that are a bit ambitious for their skills. There you need good filters.
I've been on both sides of the table.
My process for hiring is:
- Pre-screen CVs and look for smells (job hopping, a string of meh employers, poor technical skill match, lack of seniority, etc.). You can read a lot from a CV. I tend to give people the benefit of the doubt here. But given 20 CVs, I'm not talking to 20 people.
- Quick phone interview either with myself or somebody I trust to have good judgment. This is a critical call. Mostly this is to confirm the basics line up (availability, expectations, skills).
- Interviews in person (ideally). At this point I either like the person or I really don't. Yes this is subjective. But initial impressions seem to have a strong correlation to long term outcomes. Again benefit of the doubt here. But I'm not going to pretend it's not influencing the outcome.
- Decision to proceed with negotiations or not.
Note I don't do any coding interviews whatsoever. I hate those with a passion. They don't tell me anything. I prefer portfolios (e.g. Github) or having candiates talk about something they did. I'm not going to probe them for encylopedic knowledge of algorithms, doing some shitty IQ test, or whatever.
Most of your points I agree with, but this? Cmon grandpa
No man, it's not and never was - unless you are aiming for a "career" at JP Morgan and the likes.
You sound like a typical classist MBA; don't you have Linkedin posts to make and employees to micromanage?
In the past we took a chance hiring people with non traditional backgrounds but now that everyone thinks they can do complex engineering with the help of AI, we need to know people have truly studied the fundamentals over a period of 4 years at a university.
The only way to be sure that I know of is to ask questions in-person. They don’t have to be absurd, just things that you should be able to answer if you understand fundamentals, like “describe the differences between a binary tree and a B-tree,” or “describe the fetch-execute cycle.”
Yeah, no. I'd deduct points if the candidate wears a suit. What a huge red flag, missing all sorts of context and appropriateness cues.
This shouldn't be surreal at all. A candidate just wasn't able to make up relevant experiences on the spot.
Side note, as far as a job requirements goes the bigger issue is asking for impossibly diverse experience and asking for things that can be easily learnt. This promotes lying because the liars are the ones that are rewarded with an initial interview. I was talking to a fresh graduate with some volunteer experience who was having difficulty getting a job, and all I could hesitatingly recommenced was to tell him lie on his resume so that his resume could get past the screening.
That said, my personal ethics don't let me lie to an actual person.
This are not school exams, company wants to hire the best candidate. If all fail then best failing is still the best candidate, and this can be measured and/or perceived by skilled interviewers.
Really? That's the bigger issue?
Company wants to pay money to someone in exchange for services. They have unreasonable expectations. So that makes it OK for people to deceive them in order to have them believe that their unreasonable expectations have been met?
I don't think that unreasonable expectations should be rewarded. But an unreasonable expectation is just "being stupid and harming yourself."
Deceiving others in order to take their money under false pretences (which is fraud) is immoral and harms others.
The two are not remotely comparable.
> This promotes lying
No it doesn't. If someone feels "encouraged" to lie and defraud others because they want something from them (even if the "someone else" is objectively stupid), that is no one's fault but their own. And their wishes and desires are just as unreasonable as the company's. [The wish/desire on the part of the applicant is wishing that the company had reasonable expectations]
Indian SDE market is an extreme case of Goodheart's law, but that's a topic for another day!
If recruiters only pick up your resumes based on keyword matching themselves, what is one to do, if not adapt their resumes to said keywords so they can at least try to get to a human interview?
Not talking about India specifically, but in general. Hiring is broken, so everyone tries to fix it in their own ways to maximize their chances.
I think the point is that LLMs makes it easier and cheaper to produce a large volume of convincing lies. The candidate likely would not have been able to produce convincing-enough lies to get through the resume screen without LLMs.
For instance if you want to prepare for a C dev interview and would like to review what 'static' means and does (one of the super usual interview questions) you can just ask and immediately get a pretty much perfect explanation without noise. It's not cheating, it's just a better tool.
It would be better if we just stopped asking l33tc0d3 questions, since it's been shown over and over again it's a pointless waste of time on both side of the aisle.
If a candidate is taking the time to practice and master leetcode it does show the candidate is motivated, demonstrates their ability to learn and internalize knowledge, and to utilize that knowledge under pressure.
If those are things you want to screen for and have a high volume of talented candidates I can see a use for them.
I've been really impressed with how much a of performance lift working on leetcode with AI is. It's so much easier to focus on developing rapid problem decomposition skills and working with an interviewer during the problem.
Unfortunately it's also necessary to improve this process because the current standards for the companies still doing leetcode interviews are getting pretty wild these days. Meta requires 2 med-hard question solved in 20 minutes or less each for the screen these days! Even if you have solid algorithmic thinking solving and implementing solutions that quickly requires you to be insanely prepped.
i totally agree otherwise, there are a ton of other good proper ways to prepare for an interview using AI. for example his resume, im sure he asked for some refinements about how he was wording certain things, and who cares at all that its not word for word grammatically from his mind. getting past the resume screening process is a huge part of the battle, and all the scam attempts and bad candidates will be optimizing their resume as well. The info within it should still be relevant about your ACTUAL technical skills or you are just also falling into the scam/bad candidates category.
Of course your example is a solid one, which ive done myself as well for leetcode stuff and plenty of other stuff.
IF his experiences where actually real and he used AI to simulate an interview based on them, thats a fine use case for AI, so i guess this article likely should have used a more clear way to condone this candidates preparation.
i got a high paying job at meta once i started see it as 'presentation' and not a 'conversation' .
I play this stupid ass game to make money
If they say they don't remember, that's a red flag. If they can't describe how something works, that's a bigger red flag. You're not looking for photographic memory, but it's very obvious once you do it a few times who is real and who is lying.
It's common sense, if you don't put in at least a tiny bit of effort in your hiring process, you can only expect to attract similar low effort candidates.
I haven't written non-proprietary code in a decade.
It is surprising to me that folks looking for a new job would not do this proactively.
Is it? I can think of projects I've worked on that have come up with friends that I have no idea how they worked anymore, just barely if at all. If the project was within the last 2 years, then yeah, but if its 8 year old plus code, I don't expect anyone to remember. However, they could have looked at it when they sent it over and refresh their minds.
If I just have to give a code example of mine on the spot during an interview with no prep, I'm sure as hell not going to remember why I took a certain approach unless there are comments.
1. Don't use blur to redact documents. Whatever blur was used can probably be reversed.
2. Don't try to hide the identity of someone you're talking about by redacting a few details on their resume. With the prevalence of public and private resume databases, that's probably easy to match up with a name.
I’ve given a lot of interviews, candidates will always try to come up with the best story as an answer to your question because “I can’t think of an example” is not an acceptable answer. It’s a demand you’re placing on them.
Also having experience puffed up on a resume happens around 100% of the time. The point of the interview is to figure out how much real relevant experience the candidate has.
OP was right to end the interview as they were an unprepared candidate and a bad fit, but low-key threatening someone with “word gets around” who’s trying to find a job and probably starting to panic about not having one doesn’t make him the good guy in this story that he thinks it does.
OP could have just told them not to use AI in future, but even that’s unnecessary as the lesson’s already been learned.
(I’ve also noticed that towards the end of the post OP mentions this, but it doesn’t line up with the actual call as described unfortunately)
Maybe I am the rounding error. I have zero puffery, exaggerations, embellishments, stolen credit, or lies on my resume.
Agreed on the blur thing, though. Blur tools should come with warnings.
I just got mosquito noise when I sharpened. Are you confusing blurring with pixelating?
As long as the blur is strong enough, there's no way to get the text back.
But a deconvolution filter will. You can't do it in Photoshop but you can with a dedicated tool that tries different deconvolution kernels until it finds one that matches the exact original blur function.
This is how you can remove motion blur from a photo due to camera movement, for example. It's wild how much information is still there, in the exact precise levels and shape of the blur.
There are limits of course, but they're much further than you might expect.
I think at this point we are in a world where the cat is out of the bag and it's not are you or are you not using AI but how are you using it. I personally don't care if a candidate wants to use AI but be up front about it and make sure you still understand what it is doing. If you can't explain what the code it generated is doing an why then you won't be able to catch the mistakes it will eventually make.
What we do instead is send out a test - something like a mental ability test - with hundreds of somewhat randomized questions. Many of these are highly visual in nature, making them hard to copy-paste into an AI for quick answers. The idea is that smarter candidates will solve these questions in just a few seconds - faster than it would take to ask an AI. They do the test for 30 minutes.
It’s not expected that anyone finishes the test. The goal is to generate a distribution of performance, and we simply start interviewing from the top end and make offers every week until we hit our hiring quota. Of course, this means we likely miss out on some great candidates unfortunately.
We bring the selected candidates into our office for a full day of interviews, where we explicitly monitor for any AI usage. The process generally appears to work.
On a different note, things are just getting weird.
- 0 effort on your side - very stressful for me - completely unrelated to job - ridiculous definition of someone being “smart”
Actually, I would not even do the test most likely and I bet many others neither.
Unpopular observation: Many people say this, but when they actually want or need a job they change their mind quickly.
I've lost count of how many of my peers went from "I will never grind LeetCode!" to working their way through LeetCode challenge lists as soon as a recruiter from a big tech company contacted them.
I talked to one hiring manager at a company who tested their mobile developer applicants by having them make an entire demo app with some non-trivial functionality. I assumed they wouldn't have any applicants, but his current problem was that too many qualified applicants were applying for every position and begging to do the test.
Definition of being smart is to be quick at mental math and logic, but the puzzles are represented visually. And yes, both those skills are needed in the course of our work.
Contrary to what you might expect, over 80% take the test. I suppose during next hiring season, we could A/B against random selection to compare what % go past our interview.
…If you used AI and can still explain to me why code works and what it does, even better. You have learned how to use new tools.
(have not tried the randomized question approach to compare, but I’m curious to try it and see what happens)
However, the coding assignment was a really good filter and allowed us to dismiss the majority of candidates before committing to a labour-intensive face to face.
I haven't interviewed anyone since AI took off, but I am assuming that from now on the majority of candidates that would usually send us crap code will send us AI code instead; thereby wasting our time when they finally appear for the face to face.
Have you encountered that yet?
General-purpose "mental ability tests" are typically illegal for hiring in the US.
https://www.reddit.com/media?url=https%3A%2F%2Fpreview.redd....
I probably should have figured out how to request an ADA accommodation... oh well.
What I don't understand is, what did the candidate do with AI? Did they use the AI as a coach? Did they use it to suggest edits to the resume?
---
I once interviewed a candidate who was given my questions in advance. (I should point out that it was quite time consuming for me to design an interview, so I couldn't just make up new questions for every candidate.)
When the candidate started taking the "schoolboy" tone of a well-rehearsed speech, I realized that they had practiced their answers, like practicing for an exam. I immediately threw in an unscripted question, got the "this wasn't supposed to be on the test" response, and ended the interview.
The second part sounds like areal curveball unless you made it clear that the questions sent out were only representative/samples of what you’d ask.
> The second part sounds like areal curveball
That was the point. The candidate wasn't supposed to know the questions in advance. Once the candidate can practice / memorize, there's no way to evaluate the candidate.
However, this to me would be a red flag because they somehow try to blame Ai for misrepresenting their experience. So they can’t even take responsibility for that.
Ask each candidate the same questions?
The consistency lets interviewers compare across candidates, and avoids the cognitive pitfall of defining a rubric after-the-fact that lets us hire the candidate who appealed to our lizard brains.
Even at startups, questions are also usually tested on several existing employees before it is used on the first external candidate, for calibration. Companies put a lot of time and money trying to hire for actual competence.
BTW, it's industry normal for companies to come up with a programming exercise and reuse it.
So marketing works in the company's favor, and not the candidates? Its a tough pill to swallow, but bending the truth and lying seems to be the way folks get jobs now.
Perhaps not lying... But I've thought about the 1pt font white on white mega-tech-list attached to Workday resumes to get past THEIR ai-slop filters. And even had my SO get insta-rejected when whatever AI term wasn't explicitly there.
As a candidate, the market is horrific. Ghost jobs, fake jobs that gather market intelligence, scam jobs, blatantly lying candidates, AI blusters, and more. I can look at the usual places, or even HN. I've even applied to my share of HN jobs without so much as a 'no' as response.
It puts us who actually want to be honest at a pretty severe disadvantage.
It is one thing to frame your experiences in ways that are relevant to what the job is looking for: it is not only unethical to fabricate experiences, it is counter-productive. I will be checking references, and if their reports of the role you played on a project don't match yours I will not be hiring you. If you don't have references who can speak to the work you did, I also won't be hiring you. All you have done is waste my time and yours.
The sheer number of applications from auto-submit-to-every-job application processes have completely broken the system. There is simply no way for every recruiter to consider ever candidate, which is what they are now being asked to do. I know that is frustrating, and I am sorry you are in that place, but lying will not help.
We will eventually figure out how to defeat these candidate-spam bots. In the meantime the only hiring pipelines that are still functional are human-to-human individual networking.
For example, in 2003, I was fresh out of college and the job market was slow. I applied at a retail store so I could have some beer money. I was honest that I was looking for a job in tech and that I wasn't going to stay forever. Then I said I'd probably be there for 3-4 months.
I was there for 2 weeks, and I don't list the job on my resume.
Was I telling the truth when I said 3-4 months? I certainly gave them the longer end of the estimate in my head.
Was I telling the truth when I left the retail job off of my resume?
I think the message here is: don’t ask for the moon, you are not Google.
The pagination example seems like a perfectly reasonable thing for both sides to want to talk about, and which becomes relevant at a level of scale much smaller than Google.
Hard to argue with their interview process when it successfully unmasked someone who didn't have the basic experience to discuss a simple topic.
Candidates who rely on AI seem to just be totally turning their brains off. At least a candidate who was embellishing in the old days would try to BS if they were caught. They could try and fill in the blanks. These candidates give plausible-sounding answers and then truly just give up and say "ummm" when you reach the end of their preparation.
I've been interviewing for 10+ years across multiple startups and this was never a problem before. Even when candidates didn't have a lot of relevant experience we could have a conversation and they could demonstrate their knowledge and problem-solving skills. I've had some long, painful sessions with a candidate who was completely lost but they never just gave up completely.
Developers I've worked with and interviewed who rely on AI daily are just completely helpless without it. It's amazing how some senior+ engineers have just lost their ability to reason or talk about code.
Alternatively, there are people who haven't been promoted but think their AI-fu is so good they obviously should have been, without realizing that "senior" is actually a different role, with additional responsibilities.
I've found asking about their pedagogy when coaching junior engineers is a great sorting strategy right now. It isn't something a lot of people have written about so ChatGPT's answers are full of useless platitudes, and mid-level engineers often don't even know that it is part of the job.
Interviewing is hard. Over the years the one thing I have learned is that for a technical role you want to interview people for how they THINK and REASON. This is hard and requires a time investment in the interview.
Back in the day when interviewing people for roles in networking, data center design, etc. I used to start by saying I am going to ask you a question and unless you have seen this very specific issue before you will NOT know the answer and I do not want you to guess - what I care about is can you reason about it and ask questions that lead down a path that allows you to get closer to an answer - this is the only technical question I will be asking and you have the full interview time to work thought it. I have people with 4+ CCIE family certs (this is back when they were the gold standard) and 10 year experience have no idea how to even reason about the issue. The candidates that could reason and work the problem logically became very successful.
For coding at my company now we take the same approach. We give candidates a problem with a set of conditions and goal and ask them to work through their approach, how they would go about testing it, and then have them code it in a shared environment of their choosing. The complexity of the problem depends on the level the candidate is interviewing for. For higher level engineerings besides the coding, we include a system architecture interview, presenting a requirement, taking the time to answer any questions, and then asking the candidate how they would implement it. At the end we do not care if it complies, what we care about is did the candidate approach the problem reasonably. Did they make sure to ask questions and clarifications when needed. Did their solution look reasonable? Could they reason on how to test it? Did their solution show that they thought about the question - IE, did they take the time to consider and understand before jumping in.
Anyone can learn to code (for the most part). Being able to think on the other hands seems to be something that is in short supply.
> but it had been some time ago, and they never worked on any of the features
It appears that the candidate might have actually worked on the daycare app, but not on what they said they worked - i.e., the ratelimiting and pagination. It appears that they might have been working on the frontend, and took the liberty of "expanding" their role - this used to be extremely common in a big sample of the resumes, and I'm guessing it still is. They might have used AI to prep - they used to use google earlier, but the prep was (and is) still inadequate if you've not actually worked on and implemented it. I don't think it was an entirely LLM created project...
Not being to remember small details about certain projects is also perfectly fine for people who have worked for more than a couple of years. Unless you can discover a pattern of lying like the author supposedly did then I would just be perfectly fine moving on to another topic.
Notes, notes, notes. Then review them before an interview. Not bullet-point notes of things that happened (that's fine too, but not just that) but make stories when they're very fresh, like, right after they happen. You won't be able to turn raw bullet points into a story later, you'll forget too much.
Then take some time to match stories to common interview questions. That's your prep document. Feel absolutely free to fill in gaps where needed, most folks' "real" memories of these things are half wrong anyway, and there may be times you literally couldn't have an acceptable answer to a common question without making some of it up, because you didn't take useful-enough notes. What are you going to do, fail every interview that asks that question forever? No, just make the story you need, connect it to reality as much as possible, and move on. But do it ahead of time. And you only need to do this once per such question. Perhaps you'll even manage to take notes on a less-invented story later (I've found that nearly all of these stories need a little invention, though, even if you have perfect notes, to fit into the acceptable range of responses)
* Generate/improve this resume to appear very experienced.
* Generate/improve this resume to be a good candidate for this job description.
* Ask typical interview questions about this resume, and provide good answers.
AI allowed them to add plausible work to their resume that they couldn’t have come up with on their own.
I'd actually say that _not_ using AI to prepare for an interview is mistake, putting you at a major disadvantage (and there are plenty of honest ways to use it).
You can practice with AI if you want, but it is definitely not necessary. I would much rather have someone say "I don't know that one" (and have hired many people who did), rather than have someone provide some content ChatGPT gave them the day before.
That was typical before some students got handed a lot of dotcom boom money.
(And then somehow most interviews throughout the industry became based on what a CS student with no experience thought professional software development was about. Then it became about everyone playing to the bad metrics and rituals that had been institutionalized.)
You can ask questions based on a resume without them disclosing IP, nor the appearance of it.
That resume-based questions thwarted a cheater in this case was a bonus.
He proudly said they don’t ask questions based on resume, because they don’t care where you worked or where you went to school…as long as you know your stuff. In fact he only looks at the resume after the interview.
I wonder how long they will stick to this stubbornness.
So why not just have a lottery instead of a hiring process?
/s But only slightly.
* they don't trust their interviewers to be professional and objective, or
* they're trying to have a EEOC CYA paper trail that says they make efforts to be unbiased, or
* DEI motivated (e.g., not everyone has the advantage of good past experience as a starting point for conversations), or
* some other HR theory?
I assume the folks at kapwing are monitoring the responses, so if you're really open to ideas then i offer the following for your consideration:
The best interview I've had to date has been a live debugging challenge. Given an hour, a printed sheet of requirements, and a mini git repo of mostly working code, try to identify and solve as many bugs as possible, with minimum requirements and bonus goals for the ambitious.
This challenge checks all the boxes of a reliable and fair assessment. It cant be faked by bullshittery or memorized leetcode problems. Its in person so cheating and AI is out of the equation, but more importantly it allows for conversation, asking questions, sharing ideas, and demonstrating, rather than explaining, their problem solving process. Finally its a test that actually resembles what we do on a daily basis, rather than the typical abstract puzzles and trivia that look more like a bizarre IQ test.
Stumbling upon this format was such a revelation to me and I'm stunned it hasn't been more widely adopted. You'll meet many more "Sams" as your company grows - many will fool you, some already have. But a well designed test doesn't lie. Its up to you and your company to have the discipline to turn down cheap and easy interviewing tactics to do things the right way.
Job hunting has become a game of shotgunning your resume while employers cast the widest net, and this has been hugely detrimental. Internships, junior positions, and onsite training are disappearing across the board. Everyone instead wastes time shopping around without any real evidence that this way improves outcomes.
People have been lying about their experience since time immemorial. You don't need an AI to do it, you can just ask a friend with experience to invent a few plausible projects you could have worked on, and solutions you might have found. Or just look at a bunch of resumes online and read a few blog posts of people describing their work.
I'm not surprised this happened. I'm surprised by why the author was surprised. Maybe "Sam" was exceptionally bad at "faking it" in person, but I've done tons of interviews where the candidate had exaggerated their experience and couldn't answer basic questions that they should have been able to.
Honestly, this is why some companies do whiteboard coding interviews before getting to the interviews about experience, because it does a decent initial job at filtering out people who have no idea what they're doing.
I personally wrote that I had experience in a programming language I didn't, back for an interview in 2010. I got called out on it too..!
My wife has run a couple of marathons and her friend called her up to hear about her experience, because she was putting it on her resume for a job. She got it (probably not because of her running experience).
* had someone make a giant cheat sheet with interview questions and taped it to the wall behind their computer. Part way through, the tape gave out and covered him.
* had someone attempt to lip sync the answers. The guy talking and the guy on screen were not the same guy. There was a bit of pretend 'oh just lag' for a while.
* Person we interviewed was not the same one who showed up for work. Great answers, great experience on the interview. Asked about some things we had talked about for quite a while - and he could not recall anything. Came to realize not the same person.
* the glorious mechanical keyboard furiously googling for an answer.
* the sample project they were asked to create as starting point for the interview, they had never run before. They sat and read through what was likely AI generated docs to run the app. Took them a while to realize they needed something other than Java 8 installed to run the sample.
I enjoy remote work but I wouldn't want to start working for a company where I had never met anyone. It seems like a great way to get scammed.
One real-life interview would surely be beneficial for both sides.
I at least look up toward the ceiling while thinking so maybe that’s sufficient to not give off cheating vibes.
I've also had an AI cheater during phone screen, but they were pretty clumsy... A question of form "You mentioned you used TechX on your resume, tell me more what you did with it" was answered with a long-winded but generic description of TechX and zero information about their project or personal contribution.
Another thing that I can take away from that is "take home project" is no longer a good idea in AI times - the simple ones that candidates can do in reasonable time is too easy for AI, and if we do realistic system, it's too hard for honest candidates.
Take-home projects were never meant to be evaluated in isolation.
It was common for candidates to have their friends review the take-home or even do it for them.
You had to structure the take-home so the candidate could then explain their choices to you and walk you through their thought process. When you got a candidate who couldn't answer questions about their own submission, you thanked them for their time and sent the rejection later that evening.
At some point it feels like it would be easier to just get good at programming, and yet...
Just have a 1 hour or 2 hour call with candidate where you guys go through the project.
After some back & forth I was able to (politely) prove their feedback was not correct, which actually granted me a follow-up interview.
Unfortunately, this was a unicorn, most companies don't give feedback, let alone admit they were wrong.
But, take-home is preferred, I want to use my IDE, with my keyboard shortcuts etc.
Then there are take-home timed challenges on systems like hackerrank / leetcode etc, which are horrible in terms of accessibility and access. Not to mention that they are a pass/fail, and focus purely on speed, not quality.
Next to that they don't allow you to work in an environment you're comfortable in. No debugger, etc. When an HVAC company hires a new tech, do they tell him/her to do a 1.5 hour repair with only a hammer and a lighter to diagnose and fix an issue? No, it's stupid. Why do developers have to do this then?
And the same applies to live coding exercises. While there is an opportunity to explain yourself, you're still in an extremely uncomfortable environment. Why is there such an emphasis to put people in an environment where they are not set up to succeed?
HVAC has certifications you can get. We should strongly consider this in our industry. I don't think its an unreasonable compromise, especially now with the advent of LLMs.
> Then there are take-home timed challenges on systems like hackerrank / leetcode etc, which are horrible in terms of accessibility and access. Not to mention that they are a pass/fail, and focus purely on speed, not quality.
“If you can do the job under these constraints, imagine what you can do under optimal, normal conditions!! Hired!”
If we expect people to use AI, and it is available in most companies now, then being able to appropriately refactor, test, and sense-make of AI-generated code is even more important. The key is raising the bar on quality beyond mediocre, and not relying on those take homes to test skills they are no longer testing.
What compels you to play lingual games with peoples' livelihoods?
You have to match the level of prep to the jobs you're pursuing. You don't need to grind LeetCode to have a SWE career. Most people never do that.
However if you're trying to get the more competitive jobs then some prep is necessary, as you already discovered with your Google interview.
The reason so many people do interview prep is that the ROI can be extremely high. Spending 100 hours grinding LeetCode sounds like hell to most people. Spending 100 hours doing practice problems to get a $100K raise for a job where you stay for 3 years suddenly becomes a $3000/hour career booster. That high paying job opens doors for more high paying jobs in the future, so the real number is even higher.
That's why people do it. You don't have to do it and it's not guaranteed to get you the high paying job by itself, but for people in the position to take advantage of it, the ROI is huge.
But my job is very demanding and I have 4 hours after work to spend with my wife and kids before I have to start all over again. I'm just not in a season where interview prep (which may as well be a university 16-week course) is reasonable.
Since the day-to-day job rarely requires it, and I've gotten jobs without it, there's little incentive to change unless I want to.
Nothing wrong with that - nice position to be in actually.
I have a list of past projects I'm comfortable talking about. I can go to great lengths talking about any of them in detail if prompted. I'm also comfortable talking about technical topics including those I'm not intimately familiar with - that's part of my job after all. But most importantly, I'm confident enough that I can say "I don't know what that is, can you elaborate?" and "I'd need to look into that and get back to you".
I've you're going to leetcode me, I'm going to underperform. I've never had to do leetcode for a job. I also don't typically apply to the kind of companies that think leetcode is a good filter. Why should I waste their time and mine to apply to a job at a company I'm probably going to hate working for?
Was it really necessary to take the moral high ground and lecture the candidate? As if companies are honest and well-meaning in interviews. You caught him and that's the end of it.
When caught in vulnerable positions, some people are very open to sincere remarks, but the situation is fragile. Not wounding the person further is the key.
I always try to remind myself, that I don't need to cut with the sword of truth. I can (and shall) point with it, too.
What happened though was the candidate decided to paste the entire challenge prompt into cursor and I watched cursor fail at completing the assignment. I tried to nudge them to use their own skills or research abilities, but alas did not come to fruition, and had to end the interview.
The crazy part was they had 8 years of experience, so definitely have worked before not using AI, so it was very strange they did that, especially since they remarked that the challenge was going to be easy
Is that really good advice?
If you have the wisdom of knowing when to embellish and when to blur, then you're more likely to get a job and more likely to fit in.
I'm a spectrum, and generally I'm over-truthful and I notice my habit regularly affects me negatively.
Saying something that is untrue is completely different from blurring or glossing over some of the details. The interviewer can always ask follow up questions if they want to hear more details: lying removes the opportunity for accurate understanding.
Saying something that is untrue might sometimes help someone fraudulently land a job: if it is believable, if they can back it up when asked, if the company never finds a way to check and if they never contradict themselves at all.
But it is just as likely that the answer "I don't have any experience with that, but I would google '<phrase>' and start from there" would have done a better job with no possibility of being summarily and appropriately dismissed if they get caught.
Twilio indeed can't handle batching of SMS requests -- even to this day several years after I asked them to :)
To be specific, what I want is what sendgrid offers, copy + replacements, so I can send the copy I want to send, a list of recipients and a list of replacements for each recipient in a single request.
It's still a good idea to try to bullshit candidate on topic he claims to know well.
I would probably have been fooled by the applicant's screening interview, but it would have rapidly come apart, in the ensuing steps.
My team was a very small team of high-functioning C++ programmers, where each member was Responsible for some pretty major functionality.
This kind of thing might be something they could get away with, in larger organizations, where they would get lost in the tall grass, but smaller outfits -especially ones where everyone is on the critical path, like startups- would expose the miscreant fairly quickly.
Yes, developers use AI in 2025 and this will only increase as the technology gets better. Shaming the use of AI is like taking away a plumber's toolbox because you'd prefer they work with thier hands alone. Developers at all levels have a use for AI, and given two developers with the same skill level why wouldn't you prefer one who could use AI as a tool.
If you are already hiring an engineer on their output over their comprehension, rate the output that they give you
I can't stop repeating it, just invite the candidate to your office. That's it, that's how simple the problem is solved.
Just in case anyone else in the audience is curious, this is what self-justification of egregiously bad behavior looks like.
If you can't be trusted to work remotely, absolutely stick to in-person roles. If you think your coworkers are any less deserving of your respect and candor because they aren't in the same room as you, you definitely aren't qualified to work remotely.
You also just get a much better idea if the person will work well in the team and if they're passionate about the work.
Stopped hiring people who can't show up for a chat.
The thing with interviewing is that ultimately the questions are fundamentally unpredictable. No coach, or AI, can truly anticipate what the questions will be.
The AI was used as a tool to generate false stories, but that's not what I assumed when I read the title. It's common for people to "prepare" with LLMs by having them review resumes and suggest changes, but asking an LLM to wholesale fabricate things for you is something else entirely.
I do think this experience will become more common, though. There's an attitude out there that cheating on interviews is fair or warranted as retaliation for companies being bad at interviewing. In my experience, the people who embrace cheating (with or without LLMs) either end up flaming out of interview processes or get disappointed when they land a job and realize the company that couldn't catch their lies was also not great at running a business.
To add to your experience, I became increasingly suspicious of the "perfect fit" resumes. it's insane how so many people just put the right keywords. I think it might work to pass in larger companies where HR use automated systems to triage applicants.
For example, if you have 3 years of working experience and claim, "I know Docker, Kubernetes, AWS, GCP, Azure, Python, React, PostgreSQL, MySQL, and networking extensively," in 99% of cases, I can no longer trust anything you say.
As for the 1% hidden gem I might miss out on, I likely won't have the budget for them anyway.
I don't know if this is true per se, but many job seekers in online forums seem to believe it is. Typically, keyword stuffing is thought to placate some nebulous "AI system."
Whether such systems actually exist is unclear to me.
That way we can spend massive piles of money on cloud compute, monitoring, documentation, not to mention the constant maintenance to mitigate the security issues in the multiple layers of libraries we depended on.
Anyone paying attention has started planning accordingly for this over the last couple years. The remote work revolution has resolutely failed, and it's clear in retrospect it never had a chance.
Plenty of candidates are willing to lie and as we see here AI has made lying much cheaper. There is nothing you can put on your resume that AI couldn't have put there for anyone. But AI can't yet fake a network.
Personally, I'll put in second-degree referrals to my company: if someone I have worked with has worked with the person and is willing to personally vouch for them, I'll put their resume in and ping the recruiter (yes, it's gotten so bad even internal referrals don't break through the slush pile without a specific ping.) But I get the recruiter's attention because I only recommend people I have reason to think are actually good.
Oh my... I don't think I've ever seen a resume that didn't embellish or straight up lie about the applicant. AI does make lies more convincing and allows to go further with lies though.
Also, I'm impressed and upset that it takes so much effort to get a job doing something that sounds like entry-level Node.js / React stuff :( And the effort on the part of the applicant to manufacture this fake identity and experience to apply for this kind of job... and they are a masters student! Like... shouldn't this alone qualify you for the kind of low-stakes undemanding job?
Scamming may not be new, but a person using AI in this way is able to penetrate quite deeply into (long, tedious, time-consuming) interview process if folks aren't keeping an eye out for it (and this article, like many personal experiences, indicate that people aren't yet). Having an AI voice in your ear, rapidly providing you answers in real time is something new; at least in terms of how easily accessible it is.
It's amazing to me that folks have the audacity to come to interviews like this. I think some candidates genuinely feel that it is a reasonable thing to do along the lines of stuffing their resumes with keywords to get through the various recruiter filters. It's like hey, everyone in baseball is doping, so I have to do it to keep up!
The behaviors are obvious once you've seen them before, but as an engineer and not a "talent acquisition" person, I feel deeply uncomfortable implying that some candidate I'm interviewing is lying or cheating, so it took me a bit to speak up about it.
These types of articles need to continue to come out and the conversation elevated, if just to save some poor devs hours of interviews with candidates who were able to bluff their way through the less technical initial conversations.
Remember you try to hire a ${coder, admin, } not the next tv-news-presenter, beeing on screen is not a mandatory needed skill in most jobs.
By asking for something, that makes people uncomfortable, you will exclude a lot of likely brilliant candidates.
People who refuse to do video interviews may be for example: - people who value privacy, not only their own, but most likely yours too - people who feel very uncomfortable beeing watched by strangers and who think or even know that they will perform significant worse than in an audio-only interviewsituation - people who simply don't own a camera - people who use textonly computers offjob - poeple who have experienced that your 'standard'-videochat-app may not work, maybe because they use linux, bsd, os/2 or nonstandard operatingsystems - people who don't have broadband internet, yes there are still people like that - people who pay for every bit send, and yes having a not so cheap phone/internet contract is still common in some areas - people who feel uncomfortable to let strangers in their bedroom, even virtualy - people who have disabilities or cosmetic issues that they fear may distract you - people who have disabilities where moving and out-of-sync pictures distract them - people who tend to refuse unreasonable requests and who therefor regard you as unqualified to be their next employeer - ...
All of them have good reasons not wanting video interviews.
You, as an employer, may miss your best fit.
The level of trust is simply too low - if being seen for a few hours over a web camera is that much of a dealbreaker for a candidate, there's plenty of candidates to take their place.
It's not much different than choosing to interview people who will come into the office. Of course you are limiting yourself to people in the area. But employers know this.
Also, this idea that there is a single best candidate is rubbish. There are multiple candidates that are just as good as the next. And every person has their ups and downs, as well as trade offs. I also find it hard to believe that most employers are going to be able to tell the difference on such a fine scale as to not be able to choose certain limiting factors.
Not my favorite AI driven change as I think live coding is so high pressure it can give wrong signals.
Asking developers to explain why they wrote that code mitigates against using LLM coding tools - if the candidate can’t back it up then they’ll do poorly in it.
I recently had a candidate submit an otherwise average exercise that was a big mish-mash of coding styles (inconsistently using var/let/const in js, for example). When asked about it, they weren’t able to explain their choice at all and just stumbled through it.
Off topic, why have such a take home exercise then?
One candidate was absolutely stumped and could not answer why and when they became interested in technology. They couldn't say anything about themselves personally. It was baffling.
All the candidates did really well on the online intake questions and the general meet and greet over video. However, once they arrived for the in-person part of the interview, and it got relatively technical, most did nowhere nearly as good as they did on the online. Only one or two admitted to using AI.
so all I can say is fix your assessments because this whole “they cheated” idea isnt universal, and more likely matches what people do on your job already
but for anyone that didnt read this article yet, this one is just about embellished experience custom tailored to get the interview, and there was no technical assessment
So why bother with it?
The people in these positions are scared to death to write original code and then have the balls to whine about people who use AI to provide unoriginal answers.
Why are we calling these "phone screens"?
Except it doesn’t if he hadn’t stretched the truth in his bombastic resume he would never have received an interview.
I will defend him because companies do the same thing of stretching the truth.
Because the preparation ended up not being sufficient.
Assuming people doing the hiring can be outsmarted in all cases like this is part of the problem.
Maybe 'preparation' can evolve to the candidates asking AI for a crash course and way to start using it instead of talking the talk.
It never ceases to amaze me that it's surprised it's hard to BS your way through tech jobs at tech companies. Maybe it works with tech positions at non-tech companies.
I have no doubt as well, but I couldn't help but noticing, "Don't bother with take home tests," wasn't on the list of remedies.
I think you're drawing the wrong conclusions from this experience, and if you believe it's right so, it means you didn't interview before AI.
It was exactly like that. The only difference was the lack of availability of tools that can give you the answer right away, fake the voice, etc.
But even then, if it stinks, trust your guts.
It just makes me wonder about the importance that an understanding and commitment to ethics will play as people start to use AI more and more in their daily life.
Well who are they? How would the next member of the community know this is a fake candidate. I like the idea in general of finding a way to eliminate these time-wasters but how would that work? The candidate can adjust a bit and improve the AI "foo" to come up with online answers for them.
edit: I'm talking about egregious cases where the name, location, picture, and work history are false, not the exaggerations you mention. The profiles have few connections since they do get flagged and recreated with a new false identity...
> If you feel that a profile may be fake or that it is inappropriate, you can report it. A profile may be fake if it appears empty or if it contains profanity, fake names, or impersonates public figures
They may use a real name and they may have worked some of those companies just lie about their technical level, experience, what part of the projects they worked on, etc. Those may not be covered by the reporting guidelines.
Actually it would be interesting if the interviewer had an AI to counter these tactics