Sparkles McTwinklefeet and the Insanity of Job Interviews | Chaos Lever

[00:00:00.10]
Ned: I mean, I'm still going to stay on Windows because I'm too lazy to learn another daily driver desktop operating system.

[00:00:07.15]
Chris: Right.

[00:00:09.04]
Ned: You just reach a certain age where you're like, I don't need to learn a new operating system. I need it to get the fuck out of the way.

[00:00:16.17]
Chris: Yeah, I mean, that was me when I was seven, so I completely get it.

[00:00:20.29]
Ned: Back in 1892.

[00:00:21.25]
Chris: When Columbus sailed the Ocean Blue.

[00:00:29.12]
Ned: Hello, alleged human, and welcome to the Chaos Lever podcast. My name is Ned, and I'm definitely not a robot. I'm a real human person who has desktop problems and concerns and worries and general slowdowns and weirdness. With me is Chris, who is also generally weird. Hi, Chris.

[00:00:55.02]
Chris: That was what I was going to say. I will handle insulting me, thank you very much.

[00:01:04.04]
Ned: What am I going to do for the next half an hour?

[00:01:07.12]
Chris: Take a nap?

[00:01:08.23]
Ned: No, I wouldn't mind that at all. Has this happened to you where you drink a full cup of coffee then immediately fall asleep?

[00:01:18.21]
Chris: I don't have that particular... Well, I think it's a gene, actually, because my grandmother was like that. She could have like Espresso right before bed, and often did, and would wake up fine, whereas I'll have a coffee at 11: 00 AM and still be awake at 11: 00 AM the following day.

[00:01:39.15]
Ned: It's not that caffeine doesn't affect me because it absolutely does. But drinking a warm beverage when you're already tired, the caffeine takes a certain amount of time to hit you, and in that interim, I pass out. And then when I wake up, I've started to metabolize the caffeine, and I'm like, I'm ready to go. It's usually about 20 minutes.

[00:02:02.02]
Chris: Power nap. That's the definition of a power nap.

[00:02:05.08]
Ned: It is exactly like, I feel like I discovered something that everybody else already knew.

[00:02:10.24]
Chris: Well, I mean, that's just society, right? Why learn from the group education when you can do everything on your own.

[00:02:19.01]
Ned: Yeah, I think that was the title of my autobiography. Or that joke is not going to get old anytime soon.

[00:02:27.07]
Chris: No, it sure won't.

[00:02:30.20]
Ned: So you appear to have feelings about a thing. You want to talk about that thing?

[00:02:35.18]
Chris: This has been entertaining because I thought this was going to go in one direction, then it went in another direction, and then I took a nap. Then I woke up, I said, I wrote a lot about horsies. And I was like, That's the wrong show. Okay. All right, so just a really quick aside, and then we'll get into the Kentucky Derby or whatever was this week.

[00:02:55.25]
Ned: An event that I could not care less about.

[00:02:59.15]
Chris: I mean, I I generally am against the pointless abuse of animals that serve no purpose other than to entertain rich people for two minutes a year. But the question I have is, why are all the names always so serious? The one that won was called Sovereignity. The most famous horse of all time was called Secretariat. It's Secretariat's father was called Man of War. To quote a great philosopher, Why so serious? If I was a billionaire, my horse would be called Sparkle's Mc Twinklfeet, and that motherfucker would be taking home trophies. Everybody in horse racing would have to say that name for a solid three years.

[00:03:39.23]
Ned: I love everything about that. How can we get you more money? I don't know. Should we start a Patreon?

[00:03:44.07]
Chris: How can we get me more money? That is a wonderful question. I'm going to put that on a LinkedIn.

[00:03:52.11]
Ned: I do not understand horse names. They're always stupid and ridiculous, even when they're not serious. So I think the best approach is to make it as non-serious as possible. I applaud you, sir.

[00:04:06.07]
Chris: And then Sparkle's Mc Twinkle's feet, obviously, would have siblings, and his favorite friend and brother would be Jeff.

[00:04:14.22]
Ned: Just Jeff.

[00:04:15.16]
Chris: The Nickelodeon show writes itself. God, it really does. Anyway, we're here to talk about technical things.

[00:04:23.28]
Ned: Are we? All right. I guess so.

[00:04:27.01]
Chris: So the topic that I decided and finally landed on is not actually technical in itself. It's more about working in the technical environment, in particular, getting jobs in the technical field. Because really, the reality is you could work in any industry you want. There's going to be an IT job. You can work in pharma as an IT person. You can work in construction as an IT person. It doesn't really matter. When you talk about the major jobs in particular, but honestly, as we'll see, this extends across all industries and all job values and locations and sizes and scales. The technical interview process is hopelessly broken.

[00:05:08.12]
Ned: Oh, has been for years.

[00:05:11.04]
Chris: Yeah, I mean, I don't even feel like we need to do a whole episode. We could just stop right there. Everybody's already immediately nodding.

[00:05:18.23]
Ned: I have encountered many different technical interview processes, and they were all broken in slightly different ways. So I think that's where the interesting point is. It's not that everybody has the same process, and it's the same process that's terrible. It's that everybody has a unique way of doing it badly.

[00:05:38.06]
Chris: I think everybody recognizes that that's the case. People have been trying to figure out ways around it. An example that was very, very front-page news a few weeks ago was a student from Columbia University. A big deal, you might have heard of it. He started a new company that would help people, cheat on everything, based on his negative experience trying to pass technical interviews.

[00:06:07.22]
Ned: He's not wrong.

[00:06:09.22]
Chris: So Cheung-Li, who goes by Roy, was a sophomore computer science student at Columbia University, when he decided that his natural next step was going to be to work for one of the Fang companies. We've talked about them before. That's the Facebooks, the Amazons, the Microsoft, the Apples, blah, blah, blah, blah, tech conglomerate Monsters of the Universe, the big guys. By his own account, Roy realized that getting one of those jobs was going to be difficult. So good start.

[00:06:42.13]
Ned: Yeah.

[00:06:43.04]
Chris: Again, by his own account, he spent, quote, approximately 600 hours attempting to prepare for these interviews, a process that he claims made him, quote, hate programming.

[00:06:55.16]
Ned: Yeah.

[00:06:56.06]
Chris: I've already got a couple of notes because if he thinks that 600 hours is good enough, that's very cute. The number that we throw around in society is 10,000 hours for a reason, but that's a different conversation.

[00:07:06.15]
Ned: Well, 10,000 hours to become an expert. But I think spending 600 hours preparing for a job interview does feel a little excessive.

[00:07:16.17]
Chris: Yeah, we're going to hit on all those points in a second. I just want to continue with the Roy Lee's story for a little bit here just to finish providing the background, and then we're going to talk about the background of the background. Then we'll get into the background of the background. You get it. You get it. I get things. Anyway, Roy, getting frustrated like only a sophomore in college can, decided that instead of continuing on this path, continuing to study, making connections, watching YouTube videos, maybe do practice interviews, all the things that you would expect, he just thought he would cheat. To his mind, these interviews don't really assess anything meaningful to the position anyway. It's whether you've seen the problem before, memorize the solution, and can It's not like this is your first time seeing the problem. The answer to a lot of these problems is so algorithmic. They're also just not representative at all of what you do as a programmer on the job. On the one hand, I get this, as we will talk about in a second. But you're talking about doing algorithms and doing programming. That's what you do as a programmer.

[00:08:22.01]
Ned: I mean,. You're never going to re-implement quick sort. Somebody wrote a library, you're just Who's going to use that?

[00:08:31.24]
Chris: Or will you? I heard that writing your own encryption is a great idea.

[00:08:36.22]
Ned: Everybody's doing it.

[00:08:38.16]
Chris: Sadly, that's correct. Roy ended up writing what looks like a pretty rudimentary program that takes screenshots of the interview questions that are asked of him, converted them into text, and then dumped them into, you guessed it, ChatGPT. The AI would then generate the solution, which he would then enter into the interview application or just say out loud as the answer to a question, however he needed to use that information.

[00:09:04.16]
Ned: Right.

[00:09:05.07]
Chris: And since these interviews are, for the most part, 100% remote, this was super easy, barely an inconvenience. And he did this, most notably in a tech interview at Amazon, which ended up leading to a job offer. Cheaters always win, I think, as the expression goes.

[00:09:28.15]
Ned: That's what I've heard ever since a little boy.

[00:09:30.29]
Chris: Now, this is the part where I want to stop. If this had been the lightning round, I would have come to the dramatic conclusion. But this is the part that made me start to think a lot about the topic at hand, and that is the actual interviews themselves. So A lot of people have probably heard about them. Even if you're not a technical person, you have probably heard some of the famous questions. What was it? Microsoft one was, How do you move Mount Fuji?

[00:09:56.13]
Ned: There's always something like that. How do you You have a plane that's full of green and red marbles, and you need to figure out how many tons of green marbles you have based off of some arbitrary conditions, and you only have 10 minutes and tweezers to do it.

[00:10:16.13]
Chris: Right. And you're just like, I'm trying to be a receptionist. What are we doing here?

[00:10:21.07]
Ned: Pretty much.

[00:10:22.26]
Chris: So to be fair, I have not done a technical interview at a programmer level or a DevOps level or anything like that for of these companies, but I have done one for an architect level position. This was about four to five years ago, and my therapist said, It's best to talk about your traumas out loud.

[00:10:44.17]
Ned: I agree.

[00:10:46.08]
Chris: Again, four to five years ago, your mileage may vary if you're trying to map my experience then to what you might be doing now. Exercise caution. But anyway, my experience was applying for an AWS Cloud Architect position. The cloud was all the rage. I had just picked up a number of surprisingly difficult to attain certifications, and I thought, take a shot. The role is, or at least was, part presales, part delivery, and all focused on technical problem solving, utilizing, if you can believe it, cloud computing.

[00:11:25.07]
Ned: Oh, who knew?

[00:11:27.08]
Chris: So AWS Cloud Architect, the interview process went something like this.

[00:11:32.10]
Ned: Okay.

[00:11:33.07]
Chris: So it was all remote. It was an eight-hour day. So yes, you're doing that math at home. You are correct. This is insane. With the exception of a one-hour lunch break, I was online, on camera, and talking. The process required me to have three separate one-hour interviews, then a panel interview, 15 minutes to cry in the corner, then a presentation that showed my problem solving skills and real world experience. So this was the panel that I talked to before, and I was also expected to defend the solution like it was a master's dissertation. And then finally, a managerial interview with my prospective boss and their boss. The questions, from what I can barely remember, are ranged from reasonable to bizarre. But I will agree that even from an architect position, a lot of these had little to the job itself, and I was oftentimes not really sure what they were going for.

[00:12:37.06]
Ned: Okay.

[00:12:38.02]
Chris: Now, to be fair, this was an eight-hour day. I was also borderline hallucinating by the end of it. So it is totally possible that I am not completely remembering everything correctly.

[00:12:49.29]
Ned: Yeah.

[00:12:51.06]
Chris: But I did some basic research, and my experience does seem to mirror other people online. And if I'm going to make a judgment of this experience and how it may or may not have been helpful in qualifying me for that position, I mean, come on, man. That's a lot.

[00:13:09.26]
Ned: It is a lot.

[00:13:10.21]
Chris: That was an endurance challenge more than anything else. Now, hilarious, Amazon assigned me a, mentor to help understand working at Amazon and to help navigate the interview process. And they were like, this is optional. Would you like to connect with him? And I was like, of course, I'm going to connect with him, you banana.

[00:13:30.16]
Ned: Yeah.

[00:13:31.19]
Chris: Now, in retrospect, I should not have called the recruiter a banana. When I did talk to the mentor, his advice was limited to, I can't really help you. I joined Amazon through an acquisition, so I I can go through this process myself. Good luck, though. So that's helpful.

[00:13:51.16]
Ned: How do they select this person to be a mentor?

[00:13:54.06]
Chris: You're asking some wonderful questions. Okay. Next question. Now, I don't want to make this just about my own experience because, of course, I'm a one of one.

[00:14:05.07]
Ned: Indeed.

[00:14:05.25]
Chris: But another example that was true enough because it was published on LinkedIn, and this was for a job very recently, that was of a more technical nature. According to an unnamed user who secured a cloud support engineer at Microsoft, the process of interviewing went like this. Step one was an online screening where technical questions focused on debugging a Linux-based issue using CLI tools and general network troubleshooting. Part two, which they called the first round. So I guess the first round was actually the zero-with round because we count in computer and everything starts at zero. The first round was a deep dive into Azure cloud concepts, Linux system internals, and some DNS HTTP questions and troubleshooting.

[00:14:59.12]
Ned: Okay.

[00:15:00.07]
Chris: On to the second round, which was operating system fundamentals, more networking, TCP/IP deep dives, subnetting, etc. Scripting, bash, and Python. Amusingly, they didn't talk about PowerShell for some reason. And a live problem solving exercise. And then the final round was behavioral questions, emphasizing customer obsession, which is a phrase that makes me nervous. Ownership, handling pressure, plus more technical discussions about monitoring tools and incident response. So if you're keeping score at home, then yes, that's a complete four rounds of interviews for a support role on basically the help desk.

[00:15:42.24]
Ned: Okay.

[00:15:44.10]
Chris: That's resolving issues for customers and answering trouble tickets and escalating.

[00:15:50.06]
Ned: It seems like a lot.

[00:15:51.18]
Chris: The candidate reported spending months of focused preparation, let's assume 599 hours. Sure. Practice Practicing troubleshooting scenarios, mastering the OS, mastering networking basics, which is a contradiction in terms, developing scripting and automation skills, and studying documentation for a help desk position. So this was posted on LinkedIn a week ago. You might have even seen it. It was a pretty popular post. But come on, how is this going to help? I get It's the part where it's asking about a lot of these basics and fundamentals and stuff. But we're talking about a help desk position, which is about escalating tickets. That's more about finding keywords and sending the ticket onto the right person.

[00:16:41.07]
Ned: I mean, to a certain degree, you're correct. And the skills that actually help you, and this is right in my wheelhouse because I started on help desk. The things that actually help you resolve problems on the help desk is listening to what the person is saying and then in interpreting what they're saying to what is actually going on, because those two things are not usually the same. That's not something that takes a lot of technical skill necessarily. It's more people troubleshooting. Because you're going to get an unreliable narrator every time.

[00:17:19.27]
Chris: To quote a great philosopher, Everybody lies.

[00:17:23.18]
Ned: Even if they don't know it.

[00:17:25.02]
Chris: Right. Well, and I think you're making a great point. The reality is, tests like the one I talked about above, whether it's for this job's description or another one, actually end up being irrelevant. They consistently fail to demonstrate that they do a better job weeding out candidates than a resume scan and random chance. It's not just now that they have been questionable to the point of pointlessness. If you search, technical interviews, broken on your search engine of choice, hopefully, Kageh or Start page based on our previous episode. Or was it the one before that?

[00:18:05.01]
Ned: They all-I'm certain that our listeners have heard all of them. Right.

[00:18:10.00]
Chris: And they listen to them all continuously.

[00:18:12.18]
Ned: Indeed.

[00:18:14.17]
Chris: You will find thousands upon thousands upon thousands of examples of people complaining about this. I picked one just as an example, and also because it is older. I wanted to make this not just a current, but also a historical argument.

[00:18:31.26]
Ned: Okay.

[00:18:32.15]
Chris: So we're going to wind the clock all the way back to 2016. We don't have to think about everything that happened that year.

[00:18:40.27]
Ned: Let's not. But okay.

[00:18:43.15]
Chris: Sahat Yalbukov talks about the broken hiring system on a long post on medium. And he goes off on the Google interview process in particular, being, optimized for hiring Booksmart are academic candidates who know their algorithms and data structures cold. So my expectations weren't very high to begin with. There was a very popular hacker news post a while back called Google, 90 % of our engineers use the software you wrote, which is Home Brew, but you can't invert a binary tree on a whiteboard. So fuck off. Now, this is where it becomes a little bit of a Russian nesting doll. Sahat is referencing an older example where the guy who wrote Home Brew, which is a program that still exists, is still incredibly popular and useful as a package management system across operating systems used by literally millions of people. That level of success was completely irrelevant when trying to get a job at Google.

[00:19:50.18]
Ned: Wow.

[00:19:52.05]
Chris: So if you think about this from the Roy Lee position, if having a stellar resume like the Home Brew guy doesn't give someone a leg up for technical position in the interview process? What chance in hell does a college sophomore have?

[00:20:08.29]
Ned: Not great.

[00:20:11.17]
Chris: The rest of Sahat's article is interesting as he talks about the interview process at many other companies, some of which he names, some of which he keeps anonymous. I won't go through all of them, but it's variations on a theme. I will note that when talking about the frustrations he had at, quote, Big C, where they asked him to whiteboard a breath-first search algorithm, he said, It's like testing your dentist skills based on their ability to balance a redox equation using the Oxidation number method from their undergraduate chemistry studies. What relevance does it have? Absolutely none. I am including that dentist line, not because I understand a word about what he just said, but I feel like the frustration is three-dimensional. It just leaps off the page.

[00:21:01.29]
Ned: Yeah.

[00:21:04.17]
Chris: For both, or really for all of these cases, it makes sense that Roy would get frustrated and try to cheat. Here's the other part. It's absolutely super easy to cheat on these interviews. What Roy did is just an extreme example of what's been happening all through history. Cheating on interviews, cheating on tests. Cheat on everything is the shitty way that Lee introduced his productized version of this software. But the fact is, people have always done that. It has ever been thus.

[00:21:43.04]
Ned: Yeah, I mean, all of my certifications have expired at this point, so I don't mind saying it. But there was certainly a time where I needed to achieve certain vendor certifications in order to maintain a partner status at a place we both worked. The pressure was thus that I needed to pass them very quickly on products that I was not very familiar with. I would do the logical thing, which is find a test, a brain dump of the test questions, memorize the answers, and then go take the test and pass it. I was still horribly unqualified to work on any of that software or hardware, but we maintained our partner certification, and that was what was actually important at the end of the day. So yes, cheating, absolutely a thing that people do on tests and on interviews. I have an interview story, too, but I'll hold on to that for a little bit later, I think.

[00:22:43.10]
Chris: Well, I'm glad that you do because that's exactly where we're going. Okay. I have been on the other side of the table doing the interviewing, dozens of people at this point. I've seen a couple of tactics over the years. The first one is and has been for a long time, just have someone else do the interview for you.

[00:23:04.15]
Ned: Yep.

[00:23:05.19]
Chris: This one is also notorious in testing circles. So you've got the brain dumps on the one hand, but what if you just send somebody else in to do the test for you? Somebody that knows all the answers already.

[00:23:19.09]
Ned: Yes.

[00:23:19.27]
Chris: This particular strategy was made most famous by the TV series, Suits. It's in the first episode. They don't dwell on it, probably because all the lawyers are like, Don't tell people that, you fucking idiot. But that shit is based on reality. Isn't that refreshing?

[00:23:37.25]
Ned: It is.

[00:23:38.18]
Chris: There is a non-zero chance that your lawyer never even actually passed the LSATs on his or her own.

[00:23:45.05]
Ned: And arguably, it might not even matter all that much. My wife and I have been watching Suits. We're on season 2 now. First time through, really enjoying the show. It's excellent. The interview story that I wanted to mention was I've also conducted a lot of technical interviews, and one in particular that we were conducting was over the phone, and we started to notice that there was a lag between when we would ask the question and when the person would answer. We started to suspect that what was happening is that he was wearing an earpiece and on the phone with another person who was feeding him the answer after we asked the question. We never fully proved it, but it became very obvious about halfway through the conversation that this kept happening. He would take this 10-second pause or five-second pause before answering the question, and you could almost hear something talking in the background. And at the end of the interview, we just concluded and said, Don't call us. We'll call you.

[00:24:58.28]
Chris: Yeah, that one definitely happens. That's the older way of doing it. The new fashioned way is to use ChatGPT the old fashioned way, which is to just sit there in front of the camera, type the question in, hit Enter, stall, and then regurgitate the answer when it finally comes up on your screen. It's very similar in the way that you described your story, the way that people handle this, especially if they're not used to multitasking like that. What will happen is I will ask, and this This is based on a true story that happened four or five months ago. But I'll ask a question like, what's the difference between segmentation and micro segmentation? The guy will talk for 12 seconds about his grandma's favorite cookie recipe, and then all of a sudden, he'll start talking about blast radius. I'm like, I wonder what happened in that 12 seconds.

[00:25:51.25]
Ned: Yeah.

[00:25:53.19]
Chris: Obviously, I'm over-exaggerating, but I think you get the idea.

[00:25:56.26]
Ned: Yes, I do.

[00:26:00.01]
Chris: So, yeah, just to put a pin in that, I mean, like I said, there's no guarantee, no matter what process you put in, whether it is a massive, complicated, eight-hour, just absolute slog, or if it's 15 minutes and a good feeling and a handshake. The real reality is the only way that a company can qualify somebody for a job is to let that person try to do the job.

[00:26:28.29]
Ned: Yeah, and Sometimes it's not so much whether or not they are actually capable of doing the job from a technical perspective. It's whether they're able to handle the day-to-day responsibilities and tasks of the job and work well with the rest of the team. Those two things are far more important than the technical skills that you're trying to assess. In fact, just from hiring and also working with other people in office environments, you and I both know this, there were people that interviewed amazingly, and then they were an absolute fucking nightmare to work with. And it was little personality traits and things that did not come out in the interview. But within two weeks, you were like, Oh, my God, how soon can we get rid of this person?

[00:27:14.23]
Chris: Right. And I also want to put in a note that this is not just about technical hiring. This is hiring all over the place. Yeah. You hear stories, famous stories from Hollywood along these exact lines. My favorite one is that Johnny Depp was not... He passed and got the job and started to do the work on the Pirates of the Caribbean movie. And the people in the office saw the dailies and they wanted to fire him. They were like, This isn't working. This is not going to happen. And you could include that as like, he turned that character into one of the most valuable properties in the history of Disney. Whereas if you went just by his quote-unquote interview, he wouldn't have been hired in the first place. Right. And the reverse is true. Has anybody seen Valerian in the City of a Thousand Sons? I bet those two actors crushed their interview.

[00:28:13.06]
Ned: I got about 10 minutes into that movie and gave up.

[00:28:17.04]
Chris: That is nine minutes and 45 seconds too long, my friend. I apologize.

[00:28:21.02]
Ned: Thank you. I accept your apology.

[00:28:24.26]
Chris: Anyway, back to Roy.

[00:28:27.13]
Ned: Oh, yeah. That Roy guy. Remember Roy? Yeah. Yeah.

[00:28:30.29]
Chris: So what happened? After he took that Amazon interview, he was offered a job, and according to him, at least, he declined it. He was referred to the Columbia Disciplinary Committee, a. K. A. Somebody tattled on him, and he was going to get suspended as a student, and instead of fighting, he decided he'd just drop out and start a company. He took this little idea that he developed for himself of cheating on this interview and expanded it into a product that would, for $60 a month, allow anyone to, quote, cheat on anything. And this is no bullshit. He expects you to use his tool for interviews, sure, but also for just general life. He says that he built it so that people would, quote, never have to think alone again. Every single quote that I have makes it sound more terrible and more dystopian, does it not? The website states that, quote, If there's a faster way to win, we'll take it.

[00:29:28.18]
Ned: Awesome.

[00:29:29.13]
Chris: I told you it kept it's getting worse.

[00:29:30.27]
Ned: Yeah.

[00:29:31.17]
Chris: It's extraordinarily gross. And naturally, Roy has gotten something like five and a half million dollars in seed money because everyone else is also the worst.

[00:29:39.23]
Ned: Specifically, the people who invest in seed companies or seed rounds.

[00:29:46.21]
Chris: And this goes back to the last part of Navel Gazing. And honestly, we'll have to stop here because time is a factor. As we go forward into an inevitably AI-drenched society, We're going to have to find some answers to an important question, and that is, what is the line or limit of desirable or even acceptable AI use? Again, going back a decade, we had this issue come up already when the Google Glass first came out. You remember this? They were glasses that had a little camera. They could see what you could see, and everybody was immediately terrified of them. I specifically remember going to a dive bar after a hockey game and seeing a sign on the wall that said no Google Glass.

[00:30:36.07]
Ned: Or you get glassed.

[00:30:38.25]
Chris: And at that point, I was just like, that's it. Skynet has already fucking won.

[00:30:42.21]
Ned: You were 10 years Too early, but yeah.

[00:30:48.21]
Chris: But the question, I think, is a valuable one. Was it okay then to have a computer seeing what you see and whispering in your ear? Is it okay now with Zuckerberg's mass surveillance dystopian sunglasses? These are not questions that have easy answers.

[00:31:07.05]
Ned: No. I'm thinking about my wife and I met through Quizzo, and I can only imagine that- Wait, I thought that you were a mail order bride. Okay, so for the sake of the story, we met over Quizzo. At the time, this was before the first iPhone came out. The idea of doing an internet search from your table at the bar was still... You would need a laptop and a WiFi connection, right? I can't imagine how Quizzo works now. I guess we just all have to agree not to cheat because it would be trivial to win Quizzo.

[00:31:54.23]
Chris: I see what you did there. Using your phone.

[00:31:58.25]
Ned: You like that? But that's a low-stakes environment, right? Mostly, people get real upset about shit like that, but in terms of the larger world.

[00:32:10.13]
Chris: Yeah. The first thing I want to state is that it is clear and abundantly true that people will cheat at absolutely everything regardless of the stakes.

[00:32:19.24]
Ned: It's a valid point. When at any cost. What did that website say?

[00:32:23.08]
Chris: If you would ever like to see this in action, play chess online.

[00:32:26.23]
Ned: I have been I'm aware of the controversy of chess cheating, but I haven't really read into it because I don't play chess, but I know that it is a major issue.

[00:32:42.05]
Chris: I think that's the thing. Like I said, a lot of this has probably been in the zeitgeist forever, we're just now saying the quiet part out loud. Ai, as an assistant to you doing your job, for example, seems to be coming more and more normalized. I mean, this week, I think Satya Nadella put out a quote that said something like, 30% of Microsoft's code is written by AI. Now, this is a nonsense statistic that is impossible and is absolutely not accurate.

[00:33:11.27]
Ned: And actually not what he said either.

[00:33:14.25]
Chris: Details. I read the headline, that makes it true. Fair. The point that I'm trying to make is, clearly, there's no shyness about AI being used in this manner anymore. I mean, if it's good enough for the CEO of Microsoft to be proudly talking about it, doesn't that mean it's good enough for the rest of us who are just chunking through yet another frigging RPO script? Or is this the fun part of the slippery slope? Where we go from AI being merely an assistant or a helper into being a crutch, and then finally being an outright replacement for personality or actual independent thought of any kind. Am I being dramatic? I mean, probably. I'm doing the thing with the voice. Consider the difference between these scenarios. Okay. All right. On the first hand, a student uses AI to help understand a difficult concept, asking it repeated questions to drill down and explain things in simpler terms, provide additional examples, do some... What does that thing call? Where you make a thing sound like another thing, metaphors or similies, or both, to help with learning. On the other hand, a student inputs an assignment prompt and submits the AI's response as their own work.

[00:34:43.23]
Chris: Both of them use the same technology, but with profoundly different ethical implications and implications for your actual understanding of what you just put your name on. Even if you end up with the same result, a past interview, an A in the class, blah, blah, blah. If you're the second student, you have not done anything. You have not learned. And thus, and probably most importantly, you will never be able to reproduce without the full process of AI getting it done, you will not be able to do it again. Now, my math teacher might have been wrong about me not always having a calculator with me, so I should learn my freaking times tables. 7 times 7 equals 65. But she wasn't wrong about the importance of learning the basics of mathematics the hard way. Because in that way, I can actually, I'm going to put this in air quotes, understand math.

[00:35:46.05]
Ned: Well done.

[00:35:48.19]
Chris: And I will understand it at a level that's a little more deep than just asking the calculator to do some mathing for me.

[00:35:56.09]
Ned: I'm going to bring up my wife again. This is like a hat trick. This is like the wife episode. I'm so sorry, but she's a math teacher, so it's relevant. She often says that kids that become overreliant on their calculators lack number sense. They don't have an intuitive feeling for how the numbers should go together and what a reasonable result of an operation is. And so they tend to just accept blindly what the calculator spits out without thinking through, Okay, I multiplied two single-digit numbers, and When I got a 14-digit number, I probably fucked something up somewhere. And I mean, that's an extreme example, but this is the thing. She was asking a question that had to do with finding the result of two perfect squares. The kids were putting the squared numbers into their calculator instead of just reducing them down. It was such a simple operation, but a lot of the kids just didn't have the basic math sense to look and go, Oh, 81 is 9 times 9. I take that down to 9. I think we can expand that concept more generally to anything where being able to perform the minutiae of the operation is not the important part, and we have computers and AI to do that part for us.

[00:37:19.14]
Ned: But understanding what the larger picture is and what the expected result should look like and the overall flow of things, that's what's actually important. That is what allows you to figure out when somebody makes a mistake or when an AI makes a mistake. It's like training a junior admin. I may not remember every command line switch, and that's not what's important, but I know that you should be backing up in this way. When I see you not doing that, I can point that out and go, Okay, well, what happens if you lose a backup? Now, because you're doing differential backups, We've lost almost the entire week of data. Maybe you want to do incremental instead, and here's why. There's a teaching moment there.

[00:38:08.12]
Chris: The other thing about the fundamental understanding at that level is that you are much better at recognizing maybe you don't know exactly how something is incorrect, but you know that it's wrong. Yes. You can then dig into it more. To stick with the math example, 8 times 8 does not equal 65. Not because the numbers don't add up, but they're both even. You can't get It's an odd number from multiplying two even numbers. If you don't learn that math at a certain age and you don't understand, then you don't have that inferred knowledge that 8 times 8 equals 65 is an automatic problem. That should be a red flag.

[00:38:43.25]
Ned: Yeah. I think you're going to have one type of person who wants to understand things. For them, AI is a tool to help them better understand and learn how things work, and they will use it as such. Then you will have people who don't give a shit about how it actually works, and they just want to move on to the next thing. So they will use AI in that way. I think the job of someone who's conducting interviews is to try to subtly detect which type of person you are and hire the one who actually cares about learning.

[00:39:21.28]
Chris: Right. And you can't do that with just programming on a whiteboard. No. So that was good. Like I said, we went We went into a lot of different directions with this. But I think the two main takeaways are technical interviews are still broken and always have been, and don't give Roy Lee any money. Don't do that. Please don't.

[00:39:43.12]
Ned: Well, hey, thanks for listening or something. I guess you found it worthwhile enough if you made it all the way to the end. So congratulations to you, friend. You accomplished something today. Now you can go sit on the couch, fire up ChatGPT, and take as many technical interviews as you want. You have earned it. You can find more about the show by visiting our LinkedIn page. Just search Chaos Lever or go to our website, chaoslever. Com. And hey, if you're feeling extra grateful because we had such an excellent discussion, why don't you go leave us a review on your podcast, Catcher of Choice. We don't usually ask for this, but I really appreciate it. Maybe. We'll be back next week to see what fresh hell is upon us. Ta-ta for now.

[00:40:31.20]
Chris: I think we should also charge a little extra for this episode, considering all of the math skills that we showed off and the expertise that we imparted.

[00:40:41.02]
Ned: I am expecting a job offer from AWS almost immediately.

[00:40:45.05]
Chris: As we all know, if you add a perfect square to a perfect square, you get a perfect circle.