AI & The Job Search w/Engineering Leader Anna McDougall

Oh, I gotta do one more thing.

Go.

What's going on, everybody? Good morning. When I say good morning, no, no one's here yet, but say hey, uh, if you are.

Um, I'll say hey.

Hey. Yeah.

Um.

See you, Han.

Sweet.

Good morning, everybody. Good morning, good morning. I forgot to do one thing. I'm posting it in our Torque Discord. Alrighty, sweet, cool. Good morning, everybody. Uh, this is the only live for the week. This is all we got. We got Anna this week. It's, it's kind of crazy. I've been traveling a ton. Um, obviously been documenting the whole thing. Um, like, like one does. And, uh, this week is like all leadership meetings in the mornings during my livestream. So like tomorrow we have the Torque, uh, all-hands meeting, and then Thursday I have, uh, Ronstadt's leadership meeting. So I gotta present. So, and then I don't do Monday streams. So Anna's it this week. This is, this is the only stream for Guidance Council 2.0. So Anna, thank you.

Yeah, you know, it's, it's quality over quantity this week, Taylor. That's what we're going for here. It's funny. Yes, 100%. Yeah.

I only bring on quality like Anna. But it is interesting. I've done this thing for 5 years now. Yeah. Right. I mean, you and I met 3 years ago, 4 years ago.

I mean— I think actually it was 5 years ago.

That's crazy. That's crazy. That's why I love that.

That was back in Twitter days. Back in the Twitter days.

That's nuts. Yeah. So I've known Anna for a long time, and this is her second time on the show. And one thing I've kind of talked about internally with my team is pulling back because I've done it for so long and that My job now is a ton of meetings, you know, VP, blah blah blah, whatever. But, and as you're aware of this, and we kind of talked about backstage, the higher up you go, the more meetings you're in, yada yada yada. Um, so, uh, so I am, I am pulling back a little bit. I, you know, my goal is to do 2 streams a week. Um, but, uh, but again, super pumped to have Anna here. Um, there's people already trickling in. Please say hey, would love to hear from you. Anna knows the drill. This is a networking— yes, virtual thing. Um, you know, the guests are incredibly important, the top's incredibly important, but what I've realized— and Anna probably speaks this too— like, meetups have changed. People don't go out of their house as much anymore. Um, you know, and I really truly think like these virtual streams— that's why I'm so passionate about Torque. We have all these streams going off all week, every week. Um, is that like, like, this is the new meetup, right? You connect with a person, you send them a DM, hey, it was good to see you, you go back and forth on LinkedIn, and all of a sudden you become BFFs, right? So it's like I tell people, please network on the— off the back of me. Connect with Anna, tell her, hey, she's Anna's wonderful. She's a wonderful resource. So please go follow her.

This is false advertising.

Yeah, false advertising.

Yeah.

So, uh, so yeah, please say hey. We don't know if you're watching unless you say hey. Um, and, uh, a few housekeeping things real quick. Um, first off, I want to give a shout out to Appwrite. So, um, I'm on their, uh, Heroes program. Um, I'm the only probably non-technical person on a technical Heroes program. My pitch is Appwrite has good people. I think they have a really good product because a lot of people use it. You should try it. That's my pitch. There you go. So go, go, go try it. Um, if you talk to anybody at Appwrite, um, Jason Torres, my colleague, he's an Appwrite hero as well. We've done a lot of meetups with Appwrite. We love the Appwrite team. So go check them out. If you want to get involved in the Torque community, right, we do a lot of things, right? We have a mentor program. We're about to launch our first cohort of our mentee mentor program. If you want to check out everything we have going on, go to torque.community, check that out. Please jump in the Discord, which is a link there. If you want to know when I go live with guests, text me. This is actually my number. It's not a bot, it's not AI, which we'll talk about AI, but this is not AI. This is actually me. So please text me, text live, and it just lumps you in. Um, so I send out a text every day before I go live with a guest. Um, all right, let's see here who we got here. Mark, Chris, good to see you. Oscar, My guy got my cold brewer. I got it right here. This is more iced coffee. I need to be more specific in my terminology, um, but cold brew is my jam. But it's good to see you, Oscar. Elliot, it's good to see you. Chris, Elliot, yes, meetings, yuck. Absolutely. Jalen, um, Jalen, my guy, I love it. Uh, it's too hot. It's very hot. It's very hot. I think it's like 30 degrees Celsius right now.

Thank you for the Celsius.

Yeah, yeah, thank you so much. I'm trying to get better going to Mexico and going to Europe. Like, I'm trying to get better with Celsius. Seth, it's good to see you. Another great engineering leader. If you don't follow Seth, please follow him. He's been posting about his job search. He's been posting about places he's submitted to, interviewed with. We drop his post in our Torque Discord to help others find opportunities. So he's a wonderful engineering leader here in Nashville. And Ricardo, it's good to see you. If you are in Nashville, it's my last thing. I'll let Anna introduce herself. I'm speaking tomorrow at Nashville JS Meetup on how to leverage social media to find a job. So Uh, tomorrow I'm gonna be back at my old stomping grounds, Vaco, which is hilarious. Um, so I will be, uh, I will be there speaking at, uh, 6 o'clock, um, tomorrow after tomorrow evening. So, all right, Anna, let's give the people what they want. Who are you? What do you do? And we'll go from there.

Oh, you know, I'm just a stranger who wandered in today. No, Taylor and I—

I have no idea why I'm here.

I don't know why I'm here. I'm just— I'm sitting at my desk, this thing pops up. No. I've known Taylor for a long time now. Back in the day, I think it was 2019 or 2020, I think, when we were on Twitter together at the same time. We kind of— our worlds collided, I think, through Danny Thompson.

Yep.

Who was our mutual friend. And kind of from there, the rest is history. So Taylor's known me since I was pre-junior engineer looking for my first job. And it's It's been a wild ride. I'm currently engineering manager at a mental health startup called HelloBetter. We are basically like mental health courses on prescription in Germany and soon in France as well.

There you go.

So we cover like panic, panic, depression, vaginismus, sleep problems, all kinds of different things. Yeah. So you can go to your doctor, say I have a sleep problem. You can't get a place with a psychologist. It's okay. You get a prescription, you get our app, and you get chat support. Support with a psychologist. So it's a really cool program, kind of fills this gap that we have here in Germany where people are waiting sometimes a year to get a therapy spot. Um, so we're like hoping to kind of fill that gap and help people, um, before they can see someone one-on-one.

I love that. Yeah, we do something like that. It's called like Blueberry here or something like that. My wife uses it for the kids. It's kind of like a chat on call for the kids if you can't get in touch with the doctor. It's— I, I really like that stuff. I feel like I don't know. I feel like you don't always have to go into a doctor, or if you do, to have that more support. So I love that. I love that.

Exactly. All right. And yeah, and why I'm here, sorry, just to finish up the whole thing. Yes, please. I wrote this Medium article and people liked it. And Taylor posted, is anyone here talking about AI in the interview process or AI in hiring?

Look who it is. Oh, Danny Thompson.

Oh my God.

Yeah, the legend himself. Danny, congrats on a great meetup. There you are. Yeah, yeah. Danny threw a banger meetup last night in New York, so shout out, Danny.

Nice. Of course, he's the meetup king. And yeah, so I wrote this article about what I call the McDougall Method. A little bit egotistical, but hey, I love it. And gotta own it, right? You gotta own what you do.

Absolutely.

Where we basically embrace AI as one of the tools that engineers can use in the interview process rather than trying to fight against it and and have this battle that I just don't see any way you can win that battle. So if you can't, if you can't beat them, join them. And yeah, I came up with a new kind of structure for how we can do that as engineering managers to appropriately and fairly assess candidates.

So I love that.

That's why I'm here.

I love that. All right. So lots to cover today. Also, if you want to send me— I don't know if you can quickly find a link to that. I would love to link to that.

Oh, yeah, sure.

Of course. So as, as you are finding that, let me kind of tee up the first questions. So again, I've kind of been a little bit on this, kind of this thread of AI and the job search. And no, I think, Anna, what, are you Australian?

Yeah.

I was Australian. Yeah. I love how he said New Jersey. Jason, come on.

New Jersey? We get Boston a lot because apparently, here's a fun fact, apparently Bostonians and Australians say car park in exactly the same way.

Yeah. We, our boss, me and Jason's boss is heavy Boston and trying to talk to him is crazy. Is crazy. So, all right. So kind of this theme around AI, I've had a bunch of guests on and I'm really glad to bring Anne on cuz she brings kind of a little bit of a different kind of tenor and view being an engineering leader. I feel like right now where we're at with AI is where we were with JavaScript when React kind of came on the scene, Vue, Angular, In that there were some people resistant to it. jQuery is great, right? Vanilla JavaScript is great. And I feel like there was a huge shift that people who got it got it and the people who didn't, didn't, right? And then those people kind of fell behind and they had problems finding a job. This is just me. Validate me or don't. Is that where we're at with AI right now?

A little bit. Yeah, I think it's a little bit different in that it's not a whole new language or a whole new framework, but in terms of being a skill that developers have to learn, I really do see it that way. You know, we've always said, you know, since you saw me on Twitter in 2019, we said, you know, the whole thing about being a developer is you have to keep learning new skills.

Yeah.

And I don't think many people, when they talk about that, they're often thinking about frameworks, they're talking about packages, they're thinking libraries, but they're not necessarily thinking like your tooling. I think that's one of the things that developers who have been in the industry for decades, they understand it almost better than the younger or the less experienced developers, that tooling also changes. IDEs were not always a thing.

Yeah.

DevTools were not always a thing. These are actually part of the developer experience or part of the developer tooling that you have to learn and you have to update and upgrade with. How much or how little you use it or how much you take on, how much you decide to leave to the side, is fine. That's up to the individual. But in the end, it's a good tool to have in your metaphorical tool belt, I would say.

As an engineering leader, and again, the comments are going off. Danny's comment here, I have to laugh. I knew you mentioned me. I have an AI bot that stalks Taylor. Listen, I'm not going to lie, Danny. I would not— based on how aware Danny is everywhere, I would not put it past him for this. But Danny, I appreciate you, brother. And then, yeah, again, definitely please connect with Seth as well. Um, he's also in the Torque Discord. But so what are some tools? And again, like, I'm all over the place. Welcome to interviewing with me. But like, it will kind of go all over the place here. But what are just some tools? So if someone watches this or listens now in the future and they're like, all right, great, I gotta learn AI. There's a lot within AI, right? There's like, yeah, you kind of have your prompts and your, your cursors and all that stuff. I saw somebody did a breakdown of video the other day. I should have shared it in the Torque Discord, but I forgot to. But And then there's like your hardcore engineering, NLP, ML stuff, which I don't think we're necessarily talking about that in this moment. What are just some basic tools that people should get comfortable with? And again, this is just coming from you, right? I could ask the same engineering leader, the 10,000 engineering leaders, and they're all gonna have different answers, but you're on the show, so.

So no pressure, but give me a perfect answer. Yeah. So I would say the The one that you should really get to know is some sort of AI agent. So my developers used GitHub Copilot agent. That's relatively new. I mean, everything is relatively new. I know, everything. I know. But I would say choose one, right? Like with the programming language, you know, it goes back to the content from my book, which I don't recommend to anyone anymore because it's quite outdated.

Isn't that crazy? Isn't that crazy how tech works?

I wrote a book.

It's not valid anymore because of the way tech is.

Yeah. So one of the things I say in that is like, just pick, just pick one, right? So I would say if you can pick one and get good at using one and kind of get comfortable with it, then that's more important than like, oh, okay, I've tried this one, I've tried this one, I've tried this one.

Yeah.

Because I think, um, I think getting just a bit of experience, getting used to it being part of your normal workflow is the most important thing. So I'd say that's, that's kind of the starting point that I would recommend to developers at this point in time if they haven't already.

Is that something— I mean, is this something you ask in the interview? I mean, is that— I mean, when you're, when you're hiring or evaluating talent, if somebody has the Copilot experience, if somebody doesn't, are you going to gravitate towards a person who has the Copilot experience?

I would say not at this point.

Okay.

Not at this point, um, purely because it's so new that I wouldn't want to like have it be anything that sways me one way or the other. If someone's an excellent developer and they've never touched an AI tool, probably they'll be good with AI tools when they start using them. So I'm not really concerned. Um, in about 2 years, I would say, like in a year to 2 years, I would start expecting most interesting developers to be having that.

People don't need to freak out right now. Like they don't need to freak out that they're behind right now. Okay.

No, we're still in the stage where it's easy enough to catch up. And truth be told, people are going to be learning these skills for years or decades to come.

Sure.

Like with everything, the best time to start was yesterday. The second best time is today. You might as well just start now and slowly stay calm, but just slowly chip away at it. In terms of the interview, so when we do this McDougall method for the technical interview, one of the things we say to them is you are able to use any tool you want. So we encourage developers to do the technical round. It's more like a pair programming session with part of our codebase that we kind of pull out and we make it, we anonymize a lot of stuff and we, you know, we get rid of any keys and this kind of stuff. So we create like a little mini fake codebase based on our actual codebase.

Yeah.

We give that to them 1 hour beforehand and then we give them a task as if it's their first week on the job.

Yeah.

So what is your first week on the job? You're gonna be using your normal tooling, right? You're gonna be coding in the way you code. And so that's what we encourage people to do. We encourage them to just solve the problem like you would if you were actually at work and you were on a call with one of our existing senior engineers. You're new, you don't know where certain files are located or how certain things work. You can ask whatever questions you want. You can use Google to look up docs. You can use an agent if you want to. However you want to do it, whatever your normal workflow is, that's what we want to see. So that's, that's kind of how we integrate it into our process.

I, I, I love that. Yeah, Danny in here, I agree with this take. I've said in the past, right now I believe we are in what we call the AI forgiveness bubble, but this will change. Things they let slide will become expectations. Great, you 100%. So how— I am curious with this, right? And I've had— I mean, in full transparency, I'm going to kind of ask you this question so I can answer questions better when I get asked, if we're being honest here.

Yeah.

How has the interview process changed With AI, right? Because there used to be take-home tests. And I mean, here's the deal, like, which, which is hilarious. I always say this. When I first started recruiting in 2011, I was told that— I'll never forget this— that developers who Google are bad developers. That was the information I was taught, which is way wrong. You, you, you can still Google, you can do all this stuff now with using AI. How have you changed your interviewing process? What does that look like? You're doing take-home, like, is it code? Like, what does it look like?

Yeah, so we have several different rounds. So I'm not— when I'm talking about this McDougall method, I'm talking about the technical round. I'm only talking about one of those rounds, okay?

Right.

We've got— we're already— by this point, you've already spoken to me or another engineering manager and we already think you're cool. So like you've got that already kind of ticked off. But at this stage, what we do is essentially we, yeah, give them a code base. We spend about 15 minutes kind of catching them up on— it's just 2 engineers and the candidate, by the way. We spend about 15 minutes catching up on their experience, where they're at, then they have about 40 minutes to do one, maybe 2, maybe more if they get through it, little exercises, bug fixes, or whatever in this fake codebase that I mentioned before. For us, it is, as I said, that idea of replicating that first week on the job to give us an insight into how they work. This is rapidly different from how things were. I say in my blog post, and I'll say it again here, I think a lot of engineering managers, VPs, CTOs, they're stuck in this thinking from the '90s or the early 2000s when you were physically in the office, standing in a room with people sitting at a desk and you had a whiteboard and they said, how many golf balls fit in an airplane? And you had to show them your working out, you know, to kind of solve this problem and display your logical prowess or whatever.

Right.

We can't do that anymore. For one thing, when we're remote, as most technical rounds are now, I would say 99.9% of our technical rounds are remote.

Yeah. No one's going to the office and doing whiteboards.

No one's going to the office to do a technical round. So you can't control the candidate's environment, right? They're not in a controlled environment. You don't know what screens they've got there. They could have 5 laptops all around them, right? You can't tell. You don't know. So this, this idea that they are operating purely off their own brain, this, this desire to see purely what their brain can do, it's quite old school now and it doesn't really work. And the reason it doesn't work increasingly is because of AI. I would argue it didn't work before that either, but let's not go there. Let's just stick to the AI stuff.

Yeah.

Because now we have things like deepfakes.

Now we have things like ChatGPT that can be actually listening to the conversation And then you have that startup out in the Bay Area, is it Cluey, who like raised $15 million? Um, and, and their whole thing is like cheating. Yeah, like, like their entire platform is how to cheat real time.

Correct. But, but yeah, this is the thing, the interesting thing I always say is when I, when I hear about AI cheating, I'm always saying, is it cheating if you know that it's happening? Like, if we say to the candidates We want you to use AI. We want to see how you do this. Get the help. Like, okay, I have currently a team of 6 engineers. Am I sitting next to them watching them code all day?

Of course not.

No. Like, I'm trusting them to use whatever tools they use to deliver good stuff. And in the end, we check that stuff and we talk about that stuff and we have conversations about what it requires and all that, all that jazz. But I'm not sitting there watching them. I'm not like policing the tools that they use. I'm giving them suggestions. And so in the end, I don't know why we're doing that for candidates. We want to see if they can deliver and if they can communicate about their ideas and what they're doing and if they understand it. And I think that's— that bit there is the tricky bit.

So how do you— yeah, this is, this is fascinating. How do you figure out to get in their brains on how they understand it, right? Because I feel like that seems to be like with all the tools and all the frameworks and all the AI, at the end of the day, it sounds like it's getting in their brains or getting it out of them how they understand. So how do you facilitate that type of environment during an interview?

Well, for one thing, you have to establish trust. I mean, this is, this is the thing, okay? One thing that I'm gonna just kind of do a little bit of a segue to answer your question. Just just stay with me.

Yeah, all over the place. Yeah, yeah, I love it. Love it.

Love it. That this is what psychological safety is all about, right?

Interesting.

If you go into a interview, your body—I mean, you—you know candidates. I mean, they're like your body. It doesn't know if you're in a technical interview or if you're being chased by a bear. It's—it's—it's reacting the same way. Right? It's reacting the exact same way. So our job as the interviewing team, right? So I, as I said, by the time they get to the technical round, they've already spoken to me. I reassure them this is not a pass/fail thing because it's not in our case. We, we use this as one criteria, right? The technical round is not a pass/fail. It's like a, we're weighing up the pros and cons and the benefits and strengths and weaknesses of a candidate across a whole range of stuff, not just technical skills.

Absolutely.

And so that's one thing we say kind of like, It's okay if you flunk it, you flunk it. Let's just talk about it. Right. Secondly, the engineers, I actually train them on how to, how to be approachable, how to be friendly, how to seem open. You know, it doesn't come so naturally to everyone. That's fine. But in an interviewing environment, you have to be aware that how you come across will influence candidates and their sense of safety.

100%.

So you can bring that stress level down just a little bit. That's what these first 15 minutes are for in the, in the McDougall method. If you can bring this level of stress down in the early stage, then the hope is that they can focus up a little bit more when it comes to the actual task itself.

Yeah.

Then you're asking, how do you assess how they think? Okay. Someone actually said it in the comments. They got ahead of me there that you're looking at if they're using AI, for example, what prompts are they using?

Okay.

What are they accepting? What are they not accepting? What bits of the code are they taking? What are they not taking? Then at the end, once this 40-minute coding part, pair programming part is complete, the idea is that the rest of the interview is just talking about it. It's saying, okay, I saw there the AI suggested you go with, I don't know, that you store things as objects and you went with an array instead. Why is that? Can you talk me through why you decided not to take that suggestion? You use the AI as a tool for conversation and you see how they think. I love that. Because you might disagree with what they chose. That's fine. I'm a believer in culture add, not culture fit. I don't think everyone should think the same way. What I'm interested in is seeing, like, why did you choose that method? Can you explain it in terms that other engineers can understand and say, okay, yeah, I see where you went with, like, why you went with that. Or, okay, yeah, we can talk about that. Maybe that's a better way. You know, these conversations can start. So for me, it's more about that. It's trying to see not only can they implement, but can they understand what they implemented and why they implemented it that way? That's through this conversation. I just want to say before we go on.

Yeah.

The reason we do it this way is because if we're in the moment and they're accepting it in that moment and you say, oh, why did you choose that? People will go on the defensive again.

Always.

We don't want to ask those questions during the coding. During the coding, we just let them code. We help them, we answer their questions, They use their tools, you know, we just let them do their thing. And then afterwards we start asking the questions to get their sense of understanding because we don't want to break that potential flow that they could get into if we get lucky, you know, in the technical round.

Man, there's, there's so much gold. I mean, literally, if every engineering leader hired like you, that would be ideal. It would make the recruiter's job so much easier too. What? Oh man, just, I have so many thoughts. First off, shout out to you for like that safe space. I think that's huge. I think More engineering leaders need to create, you know, an environment that allows the person to be kind of their true self, right? So, so huge on there, huge on that. I want to affirm that. I think the next thing that I want you to— so, so what does it look like with, with you? And, and you may have answered this already, but I've been too busy. I mean, the comments, I'm trying to— there's so much comments that we need to get to, but— Like, are you on a Zoom call? Is it you? Do you give them a problem? Are you— do you just watch them? They share their screen, are you— and you're just taking notes so you don't interrupt them? Is that—

I'm not there because I'm the hiring manager and I freak people out whether I want to or not. Okay. Right. I could be— you could be the friendliest person in the world, but if they know you're the one who makes the call, they're gonna get freaked out. So I'm not there. I let the engineers do it. I train the engineers on the method. We actually developed it with the engineers. So the full list of people who were involved is in the blog post. And then we, yeah, I trained up the engineers who weren't there on how to do it. And there's 2 engineers and the candidate.

Yeah.

So I've actually had a round already with the candidate at this point. So I already think, okay, I think that they'll be a good addition to the team. I think they'll bring something new or interesting to the team. So that's fine from my side. They've got my kind of like, tick, maybe there are some strengths and weaknesses, but I want to see how their technical skills are again, so we can weigh all of those things together and go from there.

Okay.

Yeah, 2 engineers and they share their screen, the candidate shares their screen and we send them the repository 1 hour beforehand so they have a bit of a chance to look through it. If they've used AI tools to look through it beforehand, that's also fine. We don't mind. They can do whatever research they want. As I said, we're pretending that it's their first ticket in their first week on the job. That's our approach.

Then is it like an actual business problem that you guys are trying to solve or is it just like a random problem? How do you guys frame that up?

Yeah. As I said, the codebase is based on one part of our app. Again, the developers create it. I trust them with this. It's a ticket like any other to create this codebase. They choose one part that they think is interesting or particularly complex or particularly not complex as the case may be. Then I get them to create 5 different levels of difficulty. So for a frontend position, for example, level 1 might be change the color of a button and level 5 might be completely refactor this function that's got an endless loop occurring or something.

That goes way up.

Or it's just super verbose and we want to see like how you would make it look better. So there's everything in between. And then based on the level that we're hiring for, we start people at different positions along that scale. So if we're hiring a mid-level, we might start at the junior. We're hiring a senior, we'd start at the mid-level. We kind of give them a chance to kind of get an easy win early.

Yeah. I like that. I like that. All right, let's, let me get to, I can't believe we're technically at time. Can you go a little longer? Is that okay?

Yeah, I can go a little longer. That's fine.

All right, hold on. Let me, this is, this is nuts. Y'all are blowing up the chat. This is crazy. All right, we got this. So Chris, Chris, I believe that's it. Companies want engineers who are more efficient with the AI. Does that in most cases. Let me, okay, let me ask you this. Have you seen increased output? Like, like, like, have you seen, like, do you have enough data points in your brain where it's like, yes, a developer with AI is moving faster and pushing more releases or whatever terms we use in the developer space than a developer without AI?

I would say that right now I have individual examples, but I wouldn't say my dataset is particularly huge. Like, I'm going off of self-reporting and this kind of stuff. I'm also quite wary of trying to measure this stuff too early. I'm, you know, I'm a big believer in experimentation, and I believe that when, especially when a tool is new, that it's kind of like a— what you want to see is a kind of exponential curve, right? That at the beginning it should be very slow because it's hard to learn something new.

Sure, right. Sure, sure.

And then at some point it picks up. I have some engineers on my team who use it for everything and work really fast, but maybe they would've worked fast without it. I don't know. Like, it's not—

Yeah, right.

It's not a randomized control trial, so it's hard for me to know. But I would say that I think there's a lot of potential there for it to speed things up, especially in getting ideas. Like, if you're first encountering a problem, I saw Danny mentioned MCPs, you know, if they're getting a handle on that, if we as an organization are getting to the point where we have the ability to have our codebase be a bit more accessible by the AI agent, then that also enables a lot of things like how, how are some ways I could do X, and then kind of get some ideas and kind of use it for ideation, or of course, for creating quick prototypes for product. That's another thing that we've heard a bit about. So yeah, I wouldn't say that I have enough data points for it to be like set in stone, but I do see a lot of potential there. And I think at this point, you know, it's worth investigating for yourself. And again, another good reason to choose one and run with one for a while rather than trying to flip-flop around.

Sure.

Because then you'll be is stuck in this learning curve state for quite a while.

So that makes sense. So this is interesting, Danny, writing a paragraph, it's taking up our faces. We had a candidate, we were like initially and realized what they, they were using to interview. Where'd Anna go? Uh, using an interviewer tool on their phone for the replies. The responses were noticeably different from the technical interview. That stuff, red flags. And that was a big thing. But the big thing was, it was because the quality replies were so bad. I always thought it was just social media posts. Yeah. So again, I mean, like, like there are going to be— have you had any kind of horror stories?

No, no, no, I haven't. I mean, I've read, I've read some horror stories, but, um, the thing is, I, I feel like, I feel like I've read enough AI slop now that I would recognize if someone were, uh, trying to answer non-technical questions with it. Um, because that to me is kind of wild. Like, if you ask, oh, what are some of your hobbies? And someone's like, in the modern era, we experience many different things that can give us joy. Yeah, yeah.

Like, no, okay. This is a good question. I don't know if you've answered this, but what recent techniques or practices, whether technical, cultural, have truly changed the game in terms of how your team is delivering better code with AI tools? What's something that surprised you or shifted your approach? It's a good question.

Yeah, no, I don't think I have anything personally that I can vouch to as yet. I'm trying to think. I do think for us and for a lot of companies, I do think that the connection of agents and MCPs will be the thing that really brings it to the next level. I think like once enterprise level, and I think Danny mentioned this is one, once enterprise level code bases can be kind of like well traversed by an agent or well assessed by an agent in a way that enables options. Yeah, because I don't think there's ever like one correct way of doing most things. So what I would expect developers to be able to do in that scenario is to be able to assess different options that can be provided by the AI and to know and understand the trade-offs there and be able to appropriately implement those and then communicate those further to whoever the stakeholders are.

Yeah.

So that would be how I see it. Culturally, I think there's still a lot of people, I mean, and it's unfortunate, but it's true, You know, we live in this LinkedIn bubble where we're hearing about new stuff happening all the time. We're engaging with it. We're around other people who are engaged in it. It feels like everyone knows about it and everyone's doing it. But the reality is there are so many developers who are not even on LinkedIn in an average month, year, whatever. They're only there to update their CV every time they change a job and that's it. You know, they are not reading all the news. A lot of them are not subscribed to newsletters or listening to podcasts or any of that. Stuff. There are a lot of developers out there who do not actually have exposure to these topics. And it's only through sharing, through knowledge sharing, that those people will be brought along. So culturally, in our company, I can say we've put a lot of effort into sharing knowledge with each other. So people actually showing how did I use an AI agent to solve this problem, people doing demos on how they've used it, for example, our QA engineers, how they're using it to help them with testing, you know, all these kinds of things like actually sharing with each other and being proactive about that. So engineering managers, as like part of our job, is finding ways to enable our developers to do that with each other and also externally wherever possible and to encourage that. But yeah, there's still actually a lot of people out there who are not onboarded into this stuff at all.

Yeah, yeah, yeah. And I think like Emre commented, how do you scale onboarding? You kind of just shared it just then. I don't know if this question is directly related to increase output or increase good output. Yeah, I mean, I would imagine increase good output is the important part.

Yes, yes, yes. And Emery, that's, that is exactly why the technical round, the McDougall Method, is all about assessing, like, not just implementing, but then assessing what you implemented. Because someone could implement every task. Oh, I just pressed a button. That was not what I wanted to do. Someone could implement every task, no problems using AI. Dunk, dunk, dunk, dunk, dunk, it's done. But they have no idea what they just did. Right? And if that's the case, we don't want them on our team either. Right? So it's also a situation where you have to know that this person is critically evaluating what the AI is giving them, not just blindly accepting it.

Yeah, absolutely. Yeah, it's like Danny said, company internal evangelists are crucial. I think, yeah, I think, I mean, even from the non-technical side, even on our neck of the woods with what we're doing at Torque, I mean, I think the more internal, I think internal evangelists are highly slept on in terms of impact within the organization. So yeah, totally.

Yeah, I think, I think one of, one of the struggles we have is that sometimes the people who are best at this stuff are the ones who aren't so interested in talking, right?

Yeah, yeah, yeah. Like, nah, I don't want— no, no, I just want to do it. I don't want to talk to anybody.

I'm good, I'll show you, you can show other people. I'm like, oh no. Yeah, yeah, absolutely. That's cool.

I love it, I love it. Uh, all right, we'll wrap up here. This is great. The chat has been through the roof. I love this. Do you have any AI? Do you have any content you consume that you want to give a shout out? And this kind of goes— Elliot, this is a great question because this is kind of one of my wrap-up questions. Do you have anything you're consuming that you want to give a shout out to?

No, honestly, I don't have anything specific. I just stay on top of everything that comes my way. Check on Medium. I have a bunch of different podcasts. that randomly talk about these topics. I don't have any one in particular. I'm not an AI expert. Like, I know we've been talking about AI this whole time.

I know, I know. Everyone's like, what?

Yeah, I'm not an AI expert in terms of its, like, you know, in terms of the technical details of it. I'll say that I'm good at, like, implementing stuff and knowledge sharing and cultural change and structural things. You know, that's my jam. But yeah, I will say that there is a podcast that I absolutely love that actually doesn't really have much to do with AI. But it's called Soft Skills Engineering. It is excellent. You must listen to it. Really, really, really great podcast. Can recommend.

Awesome. I love it. I love it. All right. Last few questions. We wrap up here. And again, thanks everybody involved in the chat. I appreciate you. That was awesome. I always feel like days like today where like the chat's going nuts and it's great conversation, it's like we're going down the highway at like 200 miles an hour. Uh, first off, I always ask folks, coffee or tea?

Coffee.

Hot or cold? Iced?

Hot. Sorry, I'm Janeway. I'm Captain Janeway. Coffee, black, hot.

I love it, I love it. Uh, next, next question. Um, do you have anybody just in your network you want to give a shout out to? Anybody who's doing really cool stuff? Anybody you want to give—

uh, oh my God, you can't do this to me, Taylor. I've got so many people. No, you had to warn me. Oh, gosh. Okay. Okay. Gergely Orosz writes the Pragmatic Engineer newsletter. He's amazing.

I've always wondered how to say his name. Say it again.

Well, I don't know if I'm saying it right.

Oh, well, it sounded way better than me.

Gergely Orosz. I'm pretty sure it's Orosz it's pronounced. But yeah, he's excellent. Who else? Kaveh Bulambo. She's the CEO and founder of Black in Tech Berlin. She does really excellent work around the Berlin tech scene if you're in Germany or are interested in coming to Germany, Americans, just saying. And yeah, who else? I mean, Danny Thompson, obviously have to give him a shout out. He's here, so no pressure. Yeah. And Tiancai Feng, if you're interested in data. He's the guy to follow. He is amazing. He's one of these people you were talking about, Taylor, where we started talking on LinkedIn and then randomly had decided, oh, let's have a virtual coffee. And now we talk like every month, sometimes twice a month, just catching up. Like, great friend now and just really good at what he does. He's really into data governance. So if that's an area that interests you, go for it.

I love it. I love it. Last question. I asked you this 4 or 5 years ago when you came on the pod, but we'll ask it again because it's been a minute. Uh, so the show's called Guidance Counselor 2.0 because we had guidance counselors in middle school, high school, and college, but now as adults we don't really have them to turn to. If you go back in time and tell young Anna something early in her career, what would it be?

Oh, that's a really good question. Honestly, my, my mom gave me the best advice and I followed it, and it seems to have worked for me, so I'm just gonna stick to what she always told me. Whenever I was confused about what to do, she always said, pick something you like and that you're good at, because if you have those 2 criteria fulfilled, you'll be fine. And that's what I've always done. I've done it in multiple fields now, and I have to say it's led to a really great, interesting, cool, successful life. So I am— I promise that is a good way to go.

I love it, I love it. All right, y'all, well, thanks again. Anna, stay on so I can say bye to you backstage. Y'all have a good one. Only stream of the week. We'll see. We have, we have a lot of events though in the Torque Discord, so go check that out, torque.community. Um, so I will see you in the Torque Discord tomorrow, resume LinkedIn reviews. But y'all have a good one. We'll see y'all next week. Peace. Bye.