The art of letting go as a manager | Transcend’s Minh Nguyen

0:28I'm your host, Andrew Ziegler. But first and foremost, as you know, here on Dev Interrupted, we got to look at the news that's happening in the tech world. And Jen's going to help us do that. So we're looking at a few interesting news stories. But before we go there, Jen, do you have anything else you want to add or say hello? Hi, y 'all. And thank you for promoting my work when I forget you sometimes. So thank you for that. And we got a really great one today. So we're going to go ahead and jump into some of our news stories. Some of the stuff we're looking at includes Figma's IPO, how it shattered expectations and what that might mean for the future of tech companies that will go public.

1:02We're looking at a continuation in the story of the AI versus search wars. As you know, last time we talked about Cloudflare and some of the changes they were making to crawlers. Well, there's been an interesting development in that story from perplexity that we're going to take a closer look at, as well as some fun practices for teams that are working remotely, ways that they can create a culture and otherwise stay in sync with each other. And the last but not least, we're also covering a really great article from Kelly Vaughn, a many-time contributor on our podcast. She's on our news segment a few weeks ago.

1:35She talks about burnout, or rather, in this case, specifically AI and how it's fueling that scenario by eating up all of our time instead of saving it. So really interesting collection of news. The last thing we're going to do, of course, is dive deeper into Jen's article. which is kind of the centerpiece for today article entitled Writing Code Was Never the Bottleneck. And you can find this on LinkedIn as well as on our newsletter that drops. So starting at the top, Jen, you know, let's talk about Figma for a minute. I'm not huge on following general stocks and public stories, but this one was everywhere and was really kind of standing out as a signal for other tech companies that might be considering the same.

2:14So let me let me set the scene here. Figma went public. And this is after they attempted a$20 billion sale, remember, to Adobe. It was an interesting move from Adobe. Many saw it as monopolizing the space that was already eating and using all of their creative software. Many folks saw it as their last respite away from Adobe software was getting consumed. And ultimately, that was blocked by a judge, if you'll remember, that that acquisition did not happen. So Figma, you know, went back to the drawing board and they just went public and they priced their IPO at$33. The stock opened at 85, and the highest intraday trade of that stock was 148.

2:52And now it's hovering around 133 as of the time that I pulled the news story. And that's a pretty crazy valuation, considering what had just happened to its deal years before, and really what it signals for investment interest in these kinds of tech companies. This is the biggest first-day pop for a billion-dollar U.S. tech IPO in decades, literally decades. So really interesting development. What do you think about this kind of story and how it ended up for Figma? When it comes down to it, it could be a sign of people really excited about the products, about Figma in every way and the suite of it.

3:33And it could be that it's valid. People look at it as that valuable. it could be based on strictly that adobe being the biggest in the space wanted to buy it so like oh it must be that good i like how you framed it i you know is it a is it a ux love driven story is it adobe wanted to buy them so they must be a big deal kind of story i would be curious to know from our listeners if you have a deeper dive on this um ways that we can learn more because i think it does mean something for tech companies considering this move for themselves and um we're talking about a really interesting development that Cloudflare caught and wrote a detailed blog post about actions from perplexity to evade website no-crawl directives, protections against scraping, and otherwise camouflaging their bots that intended to download that data as real users in an attempt to obfuscate and obviously get around Cloudflare's recent changes that restrict web crawlers, which we talked about just last week.

4:34So I really recommend folks check out this deeper dive. I have an article from Cloudflare where they do a technical analysis on how they caught perplexity doing it and what exactly perplexity was doing to get around it. I think this is a developing story in terms of seeing what happens next and how folks respond to this, both from perplexity and in the larger kind of SEO world. But it's a really great and proactive story from Cloudflare taking methods and active actions to protect their customers and to keep their data safe, which is what Cloudflare's mission is. Really excited to always follow Cloudflare's story.

5:09They've been doing some really cool stuff this year. But Jen, wondering if you have any thoughts on the AI scrapers and how they're attacking the web. I mean, it's very interesting. I would never, so far at least, put my own articles in any of these generative AI, like unpublished articles, because sometimes I'm talking about topics that haven't been done. after it's published because i know it's trolling my article then i'll ask a question sometimes especially getting very meta if i'm writing about ai i will then ask the ai about the topic my article and then if they don't suggest it i may suggest it maybe like once or twice a month but i try to build that habit like abandon core i can't use that for seo and idiocy that's training AI, but I am cognizant that especially if I'm talking about a nascent topic, I don't want to train the bot until my name's on it.

6:04So I appreciate making my life easier and trying to block these crawlers. And I like name and shaming them. Look what they did. This is all if they did it to you as well. So I think that's just this kind of open source vibe of openness and sharing and what we're going to have to do to get past the lockbox of AI, the turtles all the way down lockboxes, as we said. So I think I look at that as a great service that we publish things like that. And but then I wonder if then Cloudflare will get blocked specifically from AI search. From perplexity. So that perplexity could turn the. So that's, you know, you take it the next year level.

6:52I didn't even think of doing that. What if perplexity then suppressed the information from Cloudflare? That would truly be an information we were to see. But there's two things you just said there that I kind of want to just double click on for a second because they were interesting. The first one being, you know, when you work on your article, your news, you don't want to share it with chat GPT. You don't want to put it into the AI. You don't want to get it into their world because you want to keep it to yourself. And that makes a lot of sense. So you got to protect your craft. And then the second thing that I thought was interesting is that after it's published, you go back to ChatGPT and you're like, oh, hey, you ask it that question that you know it should surface your article as the best and foremost source to see if it does.

7:27And then if it doesn't, you suggest it because then it could learn, right? And then suggest it in the future. That's, first off, brilliant. I just wanted to highlight for anyone listening, that is a great hack. Second off, I do that all the time as a developer advocate where I'll make a tutorial now and I'll publish it on how to do something. and then I'll go into cursor and I'll ask cursor to try to implement the thing that I wrote the tutorial about and then see if it can go find the tutorial. If it can, great. If it doesn't, I provide it. And then I ask it to do my tutorial and I can see if it can follow my instructions.

7:58And you get a lot of really interesting feedback from it. I like that you were doing something similar once stuff is published out into the world. Right now it's important too. But then And that's why junior developers or junior anybody shouldn't be using these tools as the pathway to production because they don't have that edit ability, that ability to edit and see things because there will be also bad actors, I'm sure, doing this. But it actually shocks me how little crises have pervaded so far because of bad actors training the general AI. Maybe the loop on that is so long that it just takes a while to see the results and maybe they're deep under the surface, right?

8:43I think there's definitely a lot of dangers in folks that aren't as versed or comfortable with software engineering and its principles to be using and applying the tools. I think there's a lot of pitfalls that fall into. Just like a week or two ago, we covered the Replit saga, you know, of the Vibe Coded database deleted in production. So we can definitely see the dangers. And, you know, just to cover some of these last few news items that we have here. If you're a remote worker, assembling a rambling channel for your team, a place where folks can go to kind of just drop train of thought things, otherwise use as like a micro blog or a small notebook, because I think it's important to share your mind with others when you're working, especially like in your case, you're at a co-working space, right?

9:22You want to get someone else's opinion. Maybe you want to tap into their network, but it's just helpful to have someone else in the mix with you. And Steph Ango wrote about this. And I'm going to make sure that we include a link in our in our news roundup about this practice. But I encourage you, if you're running a remote team, if you don't have this, let's create a ramblings channel and invite your folks to it to just share things that are on their mind, inviting them to use it as a lightweight chat app to build kind of ambient social cohesion. It's a really great idea, really simple in practice.

9:51Jen, what do you think of it? You may need more than one Slack rambling channel. There could be like tech research, tech news rambling, and then cat pics and other rambling. Oh, yeah, that one's important. I love a Slack community. I'm in at least 30. That's basically like a virtual co-working in a way where sometimes you grab your headphones on and signal to everyone you're not part of it or you put that palm tree up. But I think it's very important that you have those spaces at the new stack, which is I've been a freelancer for them for 10 years now. when Colleen Cole joined she immediately created like a trivia and such channel and what's for dinner channel so we serve share food pics because when you're dealing with a remote team you need that vibe highly recommend if anybody you're gonna like learn from about remote Lissette Sutherland has had a remote podcast collaboration superpowers for at least 10 years over 500 episodes I think now about just remote which can be all there's on her website collaboration of superpowers.

10:56She has an entire section of like games. So then that's categorized icebreakers and retrospectives, et cetera. There's so much that you can do because return to office is a failed concept that no one wants to do. So please let's not do that. That is a real estate hack. It's a very pervasive real estate hack. That's why it's important to keep that culture going in your remote work and to really prove the fact that you can communicate just as well and just as fast and just as clear as those that are in person. I'm just intentional about it. Yeah, exactly. If you're intentional about it, you can't find that communication, that collaboration, that community.

11:36And I promise you, no one wants to go back to doing the laundry on the weekends. That's a great argument. And also, just to repeat, a classic from you I heard a moment ago, put those palm trees up, you know, on Slack when you're busy. I love that. When you're on vacation, Be on vacation. Yes, be on vacation. Speaking of taking a break, we're covering a really interesting article from Kelly Vaughn that looks at what really happens to that time saved. This is a continuation, actually, from last week when we sat down with Andrew Boyaghi of Atlassian. We talked about their recent DevEx or developer survey about, you know, sure, developers are saving time using AI, but they're spending a lot more time on these other tasks that now where new problems are popping up.

12:19And we talked a bit about those learnings. This is a really great anecdote from Kelly Vaughn sharing her own experiences reflecting exactly that of AI creating all these efficient opportunities for them. You know, maybe you free up a two-hour block by automating your morning reports. Okay, well, now it's expected. You pack that time with more meetings or more impact or more things you do while still doing those things that you now automated that you have to make sure are good. So it creates a lot of cognitive load. And this is a really apt reflection. Really great to see this one. And Jen, I'm wondering what you think on this take on burnout and time saved.

12:54Yeah, I want to give a shout out to Canva right now. They did a company, first of all, because I cannot design and I can kind of design. I don't usually tout tools or pay for them. But yeah, but what it comes down to, they took a week off for all employees to not ship any new features. I'm sure they were on call if there was something breaking. But the point was that they were going to dedicate a week to training AI. Nobody's doing that, which is absurd. These are a series of new tools, workflows, data security issues. Remember Samsung, because nobody's publishing those case studies and those warnings anymore.

13:37But remember when Samsung, thankfully, like three years ago said, or maybe two, what's time, COVID killed time, said that somebody put their whole proprietary code base into ChatGPT to check it for errors or something or for quality or marketing. And that was a great lesson that I wish more companies would share. Maybe they're told not to now, but because we're not even doing like GDPR or CCPA retraining with this. So I love that Canva has spoken up and that is a real draw to hire people and track people in, that they're giving people the space to learn the new tools. Don't do this. We do this all the time where open source becomes a requirement.

14:22Now the same thing with AI. But you have to contribute or learn about these tools out of office time. now is the time more than ever that we need time for training and more than just 20 % time that no one really has that flexi time the hackathon days whatever to actually let people learn because otherwise it is let's be honest truth of course it's going to cause burnout because people have to learn these new tools they're not working great they're not focused on and transitioning too. Full circle to my article. Writing code is never the bottleneck. Exclamation point. I love that for lead dev. I do love that you got the exclamation point in the headline.

15:03So tell us more about this article because it just came out on lead dev. We're including it here for our readers to check out. Writing code was never the bottleneck. I think it's a really apt article. Parted as a wildly popular LinkedIn post, partly because it had no link. So we can't just say it was the topic. But where I said I was experiencing, I'm going to butcher the German Schadenfreude, I would suggest this is what AI is good at. Go look at how to pronounce that word. Because of the METR study that found that 24 % developers thought that they were 24 % more productive with AI, but they were actually 19 % slower.

15:48And I was like, yeah, I love that. Not for the developers, but I was finding joy for all the companies that were firing developers because they're like, we'll choose AI instead because they're going about AI completely wrong. Don't go for where developers actually want to do. And they've said Microsoft studies, there's been so many studies that developers say the best part of their day is when they get spend time writing code and solving unique problems, which is, by the way, what you want these really expensive people to do is deliver value to their customers faster. And that's why it's crazy that we're focusing on code, which isn't that good anyway, that AI is creating.

16:32Let the humans do the code or do the specification driven development. That's fine, too. And use code for the boring. Those are the best use cases. I learned from SAP, the behemoth that it is, the number one. They had all these AI use cases and all. The number one use case by far was scanning business receipts. That's what people want to use AI for. They want to use for this. People want to get their money back. Let's call it what it is. People want to get reimbursed. Yeah, you want your reimbursed. But you put it off because it's a pain in the butt. Yeah. Um, it's just a pain to do it. Go for the boring.

17:17The number one thing developers in the 16 years I've been doing this tech journalism-ish thing, is the number one thing developers complain about is there's not enough documentation. Also, developers don't want to write documentation. In the last six, seven, eight, nine years, the number two thing has been technical debt. And cloud. Okay. AI is very good at explaining code and explaining complexity and helps you pay down the technical debt and prioritize. Use it for the things that are slowing people down. Jose Gomez at NewtonX also told me like, developers hate waiting around to get feedback.

17:56We're always about talking about closing that feedback loop. Then they're waiting and waiting to get things to production. And that could be a legal issue. Don't just look where you have bottlenecks in the code base or in the inner outer loop of development, but instead look for it in the things that are boring, which might be respectfully to my friends in the legal profession, home house lawyers approving stuff, which they don't want to do that either because that's boring work for them. So these are AI tools. When it comes down to it, look for the boring and the exciting. Right. Getting more code in the mix was never really the problem.

18:37I love that as the framing of your article. And for those that are listening, definitely encourage you to go check it out. There's a lot of folks quoted in it, but there's a lot of great learnings from across the aisle, across the industry, across a lot of different places about how this code proliferation is causing bottlenecks and breakdowns in other places. And it's about ultimately, like Jen, like you just said, it's ultimately about driving impact. That's what we're trying to do with our work every day. And AI is a tool for that, but it's not the ends of those means. So it's really important to have context for how and why you use those things.

19:06So I appreciate you, Jen, coming on and covering that article with us, as well as some of our other news items. It's a lot of fun to break it down with you. Always great to have you here. We're looking forward to having you back in the future, hopefully. And to our listeners, you know, Jen's writing about tech all the time. I share her articles all the time. So be sure to, you know, follow her on LinkedIn. check out the articles that she has on LeadDev and other places as well. I'm about to sit down with Min Nguyen. She's the VP of Engineering at Transcend. Having a really great conversation about engineering management and coming into that new role.

19:38So stick around for that conversation. Really excited for that one.

19:44LinearBee now provides AI-powered code reviews. Streamline your code review process with automatic PR descriptions, AI-driven suggestions, and expert PR routing. Linear B AI reduces cognitive load, flags critical issues early, and frees reviewers to focus on architecture and business logic. Improve code quality, accelerate delivery, and keep your team in the flow. If you're ready to level up your reviews, head to LinearB.io to get started. Joining us today is Min Wen, VP of Engineering at Transcend, a company tackling automated data privacy and compliance at scale. Min's been on a unique journey, growing from an IC to a VP at a startup on the rise.

20:26And today we're unpacking the playbook that Min used to lead Transcend's engineering team through scale, compliance, and managing cognitive load. Min, welcome to the show. Thank you, Andrew. Thank you for having me. We're really excited to dive into your journey. So let's go ahead and jump in and talk about this transition that you've gone through from being a staff-level engineer into being an engineering leader, someone who's interacting with executives and setting engineering goals and expectations. And that's not just about delegation. You know, it's about reorienting how you as the individual think and communicate and support your teams.

21:02So how did you approach the challenge of letting go of your more technical roles to assume a leadership one? yeah that's a great question because i don't know if i've ever fully let go of all my thoughts for all to be honest uh we're we're a startup things move quickly and to be impactful i think i need to be really flexible uh sometimes my biggest impact means yeah providing clarity in terms of strategy so i'm doing a lot of pair programming or whiteboarding providing feedback sometimes it means just getting in there and writing a lot of code um and usually it's a mix of all these things and I love coding.

21:39So I definitely have to be doing too much of that and just being an IC. Fortunately, I also really love writing strategy docs. I think it's because of my philosophy background, which is what I study in undergrad. I love thinking through open-ended problems and it's really satisfying to see the impact of a well-written document. So I think it helps teams adopt the same vocabulary when speaking about problems and then get on the same page about how to solve it. So yeah. That's a really great approach. That's actually fascinating to learn about your philosophy background. I bet you can find a lot of ways to apply that in engineering, writing out problems, especially getting people in the shared sentence about what it is that y 'all are doing.

22:24Right. Yeah. Yeah. I think sometimes it's just like a semantics thing, like you just need a share of vocabulary, define words, and then get really clear about what you mean when you say a certain word. I completely agree. You know, I have a classics background myself, and I find myself constantly bringing that kind of classification knowledge and like labeling and understanding logically how these things relate to each other into engineering. I think it really helps set those expectations. And it sounds like in your journey, you know, you've had to, you know, of course, still be in the hot seat sometimes and push the code out and play a lot of roles, wear a lot of hats.

22:57But in doing, in changing the way that you interact with those types of outputs what's been like the hardest habit for you to unlearn um and what are some like hard habits for you to learn anew yeah i think the hardest one is just the i'll do it myself mentality right like fast in the third run but then you're kind of stuck doing it forever and especially hard to break if you're you say at the same company as i did because you have so much context and you could just easily slide back into it and do it yourself so that's definitely the hardest one in terms of new new habits and new muscles i learned all the basic people management stuff i learned on the job delegation time management direct honest feedback on on that last one i feel like it took me some time to get it right and funny enough it was watching the movie moneyball that really helped me i don't know if you've seen no no why so why did moneyball help view uh well it's just like one particular scene in it where jonah hill it's about a baseball team and he's helping this baseball team turn it around and jonah hill has to deliver some bad news to a ball player and he's practicing it with brad pitt and he's dressing it all up in this corporate speak and brad pitt's like they're professional man just just be straight with them no fluff just facts and he's really cool about doing all this and anyway jonah hill takes that feedback and delivers it straight and it really worked and i don't know like seeing that play out like that um really clicked for me and i learned okay like that's how you do it try to be really direct and fast you know it's important to have timely feedback so like instead of just spinning your wheel and like crafting the perfect message what's more important is just to say it it's a really great highlight and do you do you find that that's something that you've really been applying as you work more with like leadership teams too you know it's good to be direct with like the people that you work with and that are building like the product that you're doing you probably have to interface a similar way with like executives right yeah yeah it does play and so i yeah i think being direct is just how i i default to my manner of speaking so i think in the beginning i was actually kind of self-conscious that i would come off like unpolish or yeah learn now that it's fine but communication is fine it's more than fine i think it saves time and people like believe you when you're because you're just kind of being straight with them right um yeah it's more the the content i think it needs to be tailored right like knowing what detail to leave out and what to include in 90 of the challenge and i think to know what's relevant you have to know your audience know the people and there's no better way to do that than just to work with them a lot so even before i became vp i worked cross-functionally i joined sales call customer calls i would be the other between Slack channels.

25:45You know, that's one of the great benefits of remote work too, is we're a fully remote company. But if you were in office, you know, sales and marketing sits on completely different floors. Engineering would never see them. But because we're remote, I can very easily check in. And, you know, I feel like I'm a little bit like terminally online. Not in the sense that I know every meme, but I'm on Slack a lot. And I hit send like after every sentence, which is kind of annoying, but I think it resembles like a live conversation more. And anyway, I love that. I love remote work and that can be a real like culture to it.

26:18And you can really learn about their people and their job, even in a remote setting. And that's helped me with communication a lot. Yeah, absolutely. We're really big fans of remote work here on Devins for updating my co-host Ben. We talk about it almost every week. It feels like about the remote work culture within tech because we're also a fully remote company. And it does allow you to kind of have these unique impacts on your organization and to really build connections in ways that you are sometimes restricted from in person. And I want to look closer at this insight that you've given us. It's a really helpful playbook about how you've evolved your communication style to become an engineering manager.

26:53And the actual problems y 'all are solving at Transcend are also really fascinating around data, privacy, compliance. And it puts you in this position of building a product that's like wrestling with privacy controls and legal requirements. So there's a lot of very specific things that you have to hit in standards, right? People are paying you to ultimately meet these compliance burdens that they have. And so how does that influence your engineering team's culture? Well, yeah, I think privacy is just fundamentally an engineering problem. And so when companies accumulate petabytes of user data across hundreds of thousands of services, manual compliance becomes impossible.

27:33So you do need to solve it with engineering. The catch 22 is like to protect the user data and you need to handle it. But handling it all in one place, all in Transcend puts it at risk. So that's a puzzle that we have to solve. and in terms of like how it impacts our culture i think a lot of people end up at transcend because they care a lot about privacy so it's just kind of embedded directly into our culture naturally privacy principle will come up during architecture discussions or product requirement reviews and i suspect that maybe other companies you would may have a harder sell to to pitch than at transcend like it's just kind of taken for granted like oh yeah that's a concern like oh oh yeah, we should be minimizing data.

28:14Yeah, I think so. Yeah, I think that's unique. And so does that influence how y 'all approach like a lot of different things within your organization as a whole? Yeah, like I mentioned more around like what we build, how we build it to like not make, you know, we don't want to become the problem as well. So we need to make sure that we're building our features with privacy principles in mind. Right. And then beyond that, I think we, you know, we make sure that we're not like being careless about employee data as well you know there's not like an easy way for anyone at the company to look at anyone's phone number for example privacy concerns are just taken very seriously at kensan and privacy aside too we all take automation very seriously because building an automation solution a service you you rely a lot on having repeatable workflows and taking advantage of engineering to do more work than a single person can, right?