Michael: Now they feel because the robots are there, it can measure, it’s more measurable. I disagree with that. ‘cause unless the humans are gone, in which case we’re all screwed, there’s going to be that element of us doing what I think is good work to actually decide: Path A and Path B, what is good architecture? What is good thinking? What is critical thinking? Why does this not work? I’m curious why it didn’t work. Let me research that piece. Everything I just described off the top of my head is our brain doing that thing of trying to find the best path and trying to be productive or whatever it is.

Rebecca: I’m Rebecca Murphy, and this is Engineering Unblocked. Engineering Unblocked is brought to you by Swarmia, the engineering intelligence platform that’s trusted by some of the best software companies in the world, including startups like Superhuman, scale-ups like Miro, and companies from the Fortune 500.

On today’s episode, I’m talking to Michael Lopp, also known as Rands. He’s a serial and seasoned engineering leader, and the author of the classic Managing Humans as well as the book that I forgot to write down, but we’ll get to that too. And he’s also, I know him because he’s the creator of a 35,000-37,000 member Slack. The Rands leadership Slack community, which has been a really powerful community for a lot of people in this industry. So Michael, welcome and what else should I have said in the intro? You can tell me the name of the book that I forgot.

Michael: There’s three books soon to be more and the Managing Humans is the one that is the first one that is still around. I’m working on a fifth edition for that, for next year. And I had to restart it ’cause of AI, which we’re gonna, I’m sure we’re gonna talk about over the next period. It was like I was writing this one chapter and I’m like, if I don’t say AI right now, I’m gonna just tone down. So I have to rewrite the whole damn thing. Anyway, so managing humans. And then I did another one called the Art of Leadership, and then I have another one that’s coming up next year for more seasoned leaders, title TBD. So yeah, I write a lot and I’ve worked at a lot of different companies. I’ve worked at Slack, Pinterest, Palantir, Apple, Apple again, Netscape. So it’s a … serial manager is the correct way to describe it.

Rebecca: And you even had a little dotcom adventure I saw on your LinkedIn.

Michael: Yeah, I did. Right after Netscape, I went to a startup you’ve never heard of and just rode that pets.com wave that we all kind of like “this is amazing!” And they’re like, and it just vanished. And it was like this horrible nuclear winter for a couple of years. That’s actually when I first went to Apple. Was that time just to kind of hide, ‘cause it was not a good time.

Rebecca: One of the things I wanna talk about today is that time, the late nineties and early two thousands. On the one hand, how different everything is from then, obviously a lot has changed, but also to what extent are we going through another time like that with AI because it feels like all the same wonder and excitement. But maybe a little bit of pets.com extent.

Michael: Well, I mean, you’re right. I think it feels …. the thing that’s different to this one is that, and maybe this is just me getting older, but it feels like it’s going faster than the other ones. The other ones were like, cool …

Netscape was probably the first one for me, but actually even before that, the Windows arrival and the GUI arrival was one of those things where, but that was really the arrival of PCs in general to mainstream. But still to me it was like, oh my God, everything’s different. And interesting. So, but the pets.com one was, you know, that whole wave, that internet bubble, it just was so … and again, similar to now, there’s so much enthusiasm and so much like, and everyone’s like, well, there’s a bubble, it’s gonna explode. And it’s got all the same vibes. It’s just going so much faster. Like, we would not be having this conversation six months ago, right? It would’ve been like, “oh, have you heard about AI?” I’m like, “yeah. I’m kind of thinking about it” as opposed to like, “whoa, everything’s about to change.”

Rebecca: Well, and even I wrote a blog post about this last week about how you could experiment with all of this for absolutely free in 1995, 6, 7. Right. Eventually you needed to buy some hardware to run your stuff on. But I remember my kid brother running a server out of our computer room. Right? And the barriers to entry were, there were very few financial barriers to learning. Now, creating product was different, but there were very few financial barriers to learning, and I feel like that’s also really different.

Michael: Yeah, it is different, but the barrier to entry right now is also just, I mean, it’s expensive from running all these servers and whatnot, but to me, what one of the things I do like about it is I’m paying 20 bucks or, you know, OpenAI or whatever. I’m just sitting there using this thing to do whatever. I was just talking about a fantasy hockey league this morning. And then last night I was working on some code, and two nights ago I was wondering why moths are eating my sweaters.

So it’s just this general and it’s accessible to everyone, which I think that’s cool. We’ll talk about a lot of the downside, I’m sure, but the bar to get there is so low. For right now, it doesn’t seem ad-supported. So there’s, I kind of like that. I am paying for this or you can use it for free if you want. But I’m paying for this and I’m not the product. I mean, maybe I’m the product. It’s probably learning.

Rebecca: You’ll be soon enough.

Michael: Yeah, but it’s still, I’m kind of liking the lack of being bombarded with ads and all that other stuff that’s still a thing that’s happening. They gotta pay for it, obviously. But it’s still a byproduct of that pets.com age.

Rebecca: Yeah, that’s very true. I’ve used this for my garden, I’ve used this for like, how do I clean the stain? Right. Totally. So you, like you said, you’ve been in this, not just in engineering, but in leadership since 1998, is that correct?

Michael: Yeah, I think that’s right

Rebecca: Wow. Three decades, right? I mean, I’m also of a certain age, let’s say. I started working right around 96. But I didn’t get into tech for several years later. But I was working and using the internet in 1996.

Michael: Me too.

Rebecca: But what’s the same and what’s different? What is like, how do you get to write the eighth edition of managing? What’s the same and what is different in your eighth edition?

Michael: It’s interesting, I think. I think one of the reasons for my writing, a lot of it is evergreen, although I have updated it four times now, is I think a lot of the lessons about how humans are human are the same.

The world is certainly evolving rapidly, but I still think there’s just basic things that are required to be a good leader. And if you know my stuff, it’s very human focused. I’m not gonna tell you how to re-architect the stack. I’m gonna tell you who is the right person to pick and how to make sure she is that person and how to give her clear guidance or to ask her for clear feedback or whatever.

It’s all those human rules and we’re not evolving as fast as the internet right now, I guess is probably the thing. So a lot of the things are the same, like one-on-ones. Again, it’s a philosophy and there’s a lot of different leadership philosophies out there.

So not everyone, it’s not everyone’s cup of tea, but I am a small team, high powered small teams, so everyone has something to do. Lots of communication, one-on-ones. This is just the Rands playbook that I’m just reciting here.

A lot of empathy, a lot of kindness, a lot of people first before the product piece. That’s another thing that you’ll find. And so that’s not really, that’s me not changing. Does that work for everybody? No. There’s a lot of different ways to run a team there. I haven’t really changed core philosophy. I’ve changed some things. We can talk about that, but I haven’t really changed core philosophy in terms of how I wanna lead a team and inspire a team during that time. So, but the technology is just moving so fast. It’s so fast. That’s both exciting and terrifying to the planet Earth right now.

Rebecca: Yeah. And hard to know like who will I be paying $20 to a month in December? I don’t even know. right? I’m definitely not signing up for annual plans or anything, that’s for sure.

So the empathy thing is really interesting to me ’cause I grew up, I started as an IC, and I kind of grew up in tech during the time when people were very eager to hire, let’s say. Most of my time in tech has been not in the dotcom. I was working at a newspaper during the dotcom bubble. But so not during that, but I started doing tech stuff, 2005, 2004, 2005 ish.

And I got to see the emergence of the servant leader and a lot of the stuff that you were writing, the stuff that Lara Hogan was writing and all that kind of stuff, which at the time it seemed like, well, this is just how you do engineering management. Like I’m learning, today I learned this is how you do engineering management. And that was really true in 2014, 15, 16, and psychological safety and all of it. Of course you’re going to say that all of this still matters today, and I agree. But how has that conversation changed now that it’s a buyer’s market, basically?

Michael: Well, let’s go back to AI for a second. The enthusiasm is through the roof, and it is not a flash in the pan to anyone who’s like, ah, it’s just a flink. I’m like, no, no, no, no, you are wrong.

And yes, it does hallucinate and yes, you should not trust it to do things without verifying, but that’s the same thing you do with humans. But its ability to zero to one a thing is unparalleled. I’ve never seen anything like it in my life. And it’s trained on a vast amount of information, which you’re working on your garden. I’m trying to get moths outta my sweaters, or I’m writing bot detection software for my Rands leadership Slack. I’m not writing a thing about that. I’m just telling it what to do and it’s doing it for me.

So that is, you’re smiling and ‘cause it’s exciting ’cause it gives us these superpowers, but all of that was potentially work I would’ve given to someone else or hired a contractor to do. And I’m super happy.

Rebecca: Or not done at all, right?

Michael: Not done at all, right? Absolutely. A hundred percent. The majority of my projects are hanging in and I just like, oh, that’s cool. I had the robots do it, didn’t work out. Bye!

I think the enthusiasm is way high, but the reality of what it can do is real. And I think there’s gonna be a downward pressure, especially for folks entering the workforce right now because a company XYZ is gonna be expecting X percent more out of senior engineer, blah. Right? Whatever that math is there.

So if I were entering the workforce right now, I would be AI-…, whatever the right way to describe it, I would just be like, it’s a tool that I need to know how to use. It doesn’t replace critical thinking. It doesn’t replace curiosity, but it does allow you to produce so vastly much more than what you could before.

So I think that’s something that’s changed right now is just that getting in is gonna be a lot harder. Are you coming in as whatever AI-capable means? And that’s a loaded term, but you have to be able to do that because that’s what everyone else is doing right now. They’re figuring out where it is and it’s not a hundred percent improvement.

It’s more, for complex things it’s five to 10%, but five to 10% is huge. I’m making these numbers up, but that’s huge improvement to a team’s ability to produce. And I think there’s a lot of parts that are a lot higher than that if it’s tactical, unit test writing, blah, blah, blah. It’s hard work, but there’s things that the robots are just amazing at right now. So that’s a thing that’s really changing here. I’m a bit far from your question though.

Rebecca: I wanted to talk about the junior engineer and that challenge because it is not unheard of for a team to have one or two more senior people and then a decently large number of juniors. But like you said, when those juniors are producing volumes of code that they wouldn’t necessarily have before, you might have a junior who can do two tickets a week by hand, and now they’re doing two tickets a day.

What burden does that put on the rest of the team, and especially does that senior engineer spend all their time reviewing mediocre pull requests from junior developers rather than contributing their own code that’s probably of higher quality?

Michael: Yeah. It’s a great question, and I think this is the swirl that’s going on right now, which is, I think for me in my journey, I’m past the enthusiasm, like “this is magic!” And now I’m into like, “hmm…” I did the classic thing. And I’ll get to your question in a second of just saying, I was building this weather tracking app and I was just throwing things at the robot and saying, “oh, do this, can you do this? Is there an API for that? Where would you get that information?” And this went on for a day or two.

But Rebecca, it was spaghetti thinking. It was poorly designed and at one point I’m looking over as it’s kind of merrily doing whatever the heck it’s doing, and it’s just … it’s lost its mind.

And by the way, this is no one’s fault except my own. This is not the robot. The robot has no intent. The robot’s literally doing exactly what I told it to do. But I was looking at it, it’s like, okay, now I’m gonna refactor this. Refactor, what are we doing? So it’ll absolutely do exactly what you tell it. And in the absence of context, it’ll guess. And you’ll be happy if it guesses right. And then you’ll be sad if it doesn’t.

So I think that senior engineer that you’re talking about — he or she is perhaps happy to have this infinite set of interns that can help them in the robots. But there’s the other people here too, and they’ve gotta figure out whether the stuff that’s coming in from these junior folks that are perhaps using robots to get this done or whatever. Is it garbage or is it good? And what are they doing to actually, especially when they’re being asked to do more, what are they doing in terms of quality control on that piece there?

So when I think of the efficiency that robots are going to give us, that just keeps coming down every time because it’s like, “Cool, yeah, I got a bunch of an infinite set of interns that are generating mediocre code because it knows anything that happened in GitHub that’s not actually relevant to my product. Well, you know, now I’m sitting here dealing with that mess as opposed to dealing with the opportunity of training engineers”, and that sort of thing.

So that’s the swirl of the AI productivity swirl we’re in right now is how’s this gonna shake out in terms of being the way that we build products and services?

Rebecca: At Swarmia, we have people, we have customers or prospects coming to us, and they’re saying we wanna understand the per developer impact of using AI in our organization. And we wanna know that if we estimated something was gonna take nine hours and then it took seven and that was because of AI, we wanna know.

Michael: How many times have you been asked that version of the question in the last 30 years, right? It’s like, tell me how productive this engineer is. And we bristle at that because it’s not magic, it’s more art than science. And yes, you can count code commits and blah, blah, blah.

But everyone works a different way and great, now they feel because the robots are there, it can measure, it’s more measurable. I disagree with that. ‘cause unless the humans are gone, in which case we’re all screwed, there’s going to be that element of us doing what I think is good work to actually decide. Path A, path B, what is a good architecture? What is good thinking? What is critical thinking? Why does this not work? I’m curious why it didn’t work — let me research that piece. Everything I just described off the top of my head is our brain doing that thing of trying to find the best path and trying to be productive or whatever it is. And the robots are phenomenal, but they don’t do much of what I just described.

Rebecca: Yeah, it is interesting. There’s so much that I have to say about this, but I’ve said it on other podcasts, but on the productivity front: five to 10% is exactly, that is my intuition as well, for day to day software development. And it’s almost more of an ergonomic improvement. It is an improvement in experience and an improvement in experience leads to an improvement in productivity via some line of dots. But it’s not like I used AI and therefore I can turn out twice as much value.

Michael: Yeah. And then there’s a bunch of other non-measurable things in there, which is, I use it to make better decisions, right? I ask it a ton. I asked it 50 questions this morning alone on some on productivity, some on random moth problems, right?

So that’s the other piece as well, is I think the quality of decisions is, I hope if people are doing the right thing with it is improving, which, how do you measure that? The fact that I’m choosing a better decision to start the whole damn thing, right? So there’s a cascading effect there as well. But again, how to measure? Good luck…

Rebecca: And yet that’s my whole job is to figure that out. But yeah, somebody on LinkedIn today was saying, “how are we gonna measure the impact on outcomes?” And I was like, “well, how are you measuring the impact on outcomes before today?”

Michael: This is a really important thing and you’re gonna giggle when I tell the punchline on this thing. But the whole hallucination bit about is it lying to you or not? And the thing I always say in written is it doesn’t lie ’cause lying requires intent and it has no intent, right?

But it does hallucinate, which is the nice way of saying lying. And the question is, well, what are you going to do if you get some bad information? And I’m like, are you on the internet? Do you live on the planet Earth? Your job as a human is to know when you’re being lied to or hallucinated to, that’s part of the gig.

Whether you’re a leader or not, is trust but verify or always verify or whatever your policy is. So it’s alarming. I asked it to quote me to me about something and it said, “oh, Michael Lopp said blah, blah, blah.” And I said, “what?” And I have Claude and OpenAI sourcing everything that gives me, so anytime it gives me a fact, it’s supposed to give me a footnote. The footnote was totally fake. Like, made up. And the only reason I know that is because it’s my website. And I’m like, this isn’t an article I wrote. And I got into an argument with it, and it was like, “yeah, no, no.” I’m like, “you lie, you don’t lie, you hallucinate.” And it’s like, “oh, sorry.”

So the enthusiasm is fun when it’s like, good question. And then you’re like, good question, but here I’m gonna lie to you for the next 20 minutes. Or hallucinate.

Rebecca: You are absolutely right. I didn’t do what you said. That’s my favorite. I think going back to the junior engineers, I’ve been doing lots of toy things with Claude Code. Because I have been an engineer, I can sniff pretty quickly, like, oh, not like that. And I think experienced engineers who are paying attention can also do that. They can be like, no, we don’t need to reinvent Git, let’s say. We’re good.

But how do we build that skill in juniors? And I mean, I do think this is going to go all the way back to how we educate juniors in school. But yeah. What are those skills that people need to develop now that maybe they didn’t need to develop pre-AI, those junior developers?

Michael: I don’t think it’s changed at all in terms of what the person coming into industry, whatever the industry needs to do, I don’t think those skills have changed. What has changed is that there’s this Gen Z meme about ‘cheating is okay’, and they’re not actually saying cheating like you and I mean cheating. It’s like what we would say is hacking, right?

We’re like, cool, how do I hack this interview? Or how do I, I don’t need to do this work. So, ‘cause the robots help me get things around it. I obviously have a knee-jerk negative reaction to the word cheating. Because cheating is, again, intent.

And I don’t think it’s where we’re going, but this is like a meme. This is coming from Gen Z right now. This like, cool, I’m gonna, this is not your mother’s or father’s, whatever. So that’s an issue there. But I think the immune system of how people are seeing the work from people ending the work is unchanged.

You’re gonna be like, cool. This is a B, or a C or an F. And because you’re not thinking about the problem. You clearly can ask a robot to do the work for you or the research for you, whatever, but if you don’t understand what you’re building or why you’re building or whether it was built successfully or not, that cognitive ability, that critical thinking skill — that’s the thing they have to learn. And whoever that is, and if they don’t learn those muscles, they’re just gonna be in this echo chamber where the robot’s giving them stuff that’s a C or something like that.

And I think the folks, the senior engineers, the senior leaders, whomever, the ones who are there to say, “hmm, does that smell right?” Our job is as important as it ever was in terms of, cool. This is just a big pile of spaghetti that does what I told you to do, but it’s a C right? Whatever a C means.

So again, I think they can use these tools to accelerate what they’re building and what they’re doing and how they’re thinking. But it doesn’t absolve them from thinking. It doesn’t. It’s just an echo chamber. It’s just a very interesting echo chamber of, “Cool. Yeah, that’s a great idea. You should do that thing!” Why? Why is it a great idea? You don’t know, robots — I know what a great idea looks like.

Rebecca: I was asking Claude if it could use itself in a GitHub action. And it was like, well, someday when the Claude command line tool exists, you’ll be able to do that, but not right now. I am using the Claude command line tool right now. They’re like, “Oh, so sorry. So sorry. I didn’t know I existed.”

Michael: I was doing something and I’m using GitHub to track everything. And again, I’m just letting it do its thing ’cause I’ve learned to combine tasks and just like a document and I say, “hey, go do this thing.” As opposed to one off spaghetti stuff.

Anyway, I look over in the other window. And it’s like, “Okay, cool. All right. Oh shoot. I just deleted the entire repository.” And it’s like, “okay, let me roll that back.” And I’m just looking, going like, I didn’t say anything about GitHub or whatever. I think I was working on a GitHub script or something like that, but it literally blew the repository. But it also, it’s like no problem keeps a shadow around so you can get it back. But when it’s just merrily doing its thing, I’m not really quite aware of what it’s doing.

Rebecca: It’s something I yell at it a lot to commit often atomically. So, yeah. I wanna go away from AI, and we may wander our way back, I don’t rule it out. But you have had a lot of experience leading people and leading teams and leading things that people have heard of, like Slack.

And so I’m curious, you talked about empathy and all of that, but how do you assess the health, especially if you’re up at director level and not directly working with the ICs? How are you assessing the health and the psychological safety of your teams?

Michael: Yeah, I kind of did it three times. Palantir, Pinterest, and Slack. And that was kind of described in different ways, but that was job number one. I literally went to Palantir as a free electron. I didn’t have a role. They were like, “hey, we’re kind of worried about the culture.” And the job was like, with all respect, a corporate therapist and it was just, and so without like, “hey, you gotta deliver this product or hit this revenue number” or whatever it was.

So that was kind of fun for like three months or so. But I use that playbook for all three companies. There’s a very obvious path to figuring that out. And then there’s the listen to what they tell you path.

The obvious path is like, okay, who are the senior leaders? Go meet with all of them and then go, who are their directs? And sometime in that first layer to the next layer, you’re gonna start hearing the story. Whatever that story is. Like, oh, I’m making this up now, but like “Our CEO thinks that he or she is vastly better designer than he or she is.” And I’m like, “oh, that’s interesting. So what?” I made that up. I’m not talking about anybody there. Whatever those narratives, they start to show up and in terms of understanding the health or the psychological safety of the team, you start to pivot and say like, oh, oh, oh.

A common one is for especially for rapid growth startups is people, the team doesn’t feel they have a concrete growth path. All three in every team ever. But that’s one that shows up because it tends to be something during rapid growth. You’re like: hire everybody, get them here now we got a lot to do, we gotta show that hockey stick growth and the humans get a little passed over because there’s not really a clear path.

And it tends to be who knows who and does so and so, like so and so, okay, now she’s a director, blah, blah, blah. And it’s just, it’s a mess because it’s not principled and it’s not fair either, by the way.

So you find whatever ails the team and focus to a set of folks. Which maybe are not obvious, not the directs of the CEO or the directs of the directors or whatever that side of it could be, “oh cool, this team over here is the one in crisis or has the biggest issue”.

But it depends, and it takes a lot of talking to folks and listening and finding what is the source set of, first set of things that you need to focus on, while also learning what the health is.

Startups are just bonkers. They’re just, it’s rapid growth. They weren’t a sure thing and by the time I usually show up they’re kind of a sure thing, but there’s just so much go, go, go and hustle as opposed to deep thoughtfulness, so there’s lots of opportunities to fix stuff.

Rebecca: Yeah, I love that. I realized this midway through my career that, and what I’m doing now is a little bit different, so I don’t know if it holds, but it held for a long time that I need enough code for there to be a mess. I need there to be enough people for there to be a mess. It’s not interesting if it’s all smooth sailing for me. I like those. I especially like productivity problems.

Michael: Yeah. For me it’s like if you look at the last three, for me it was, I showed up around 150 [people] and there’s some interesting things that have happened at 150. 150 means you have a business like this isn’t 20 and you’re month to month or every quarter to quarter. So it’s real and it also is this inflection point, it’s this probably magical number of Dunbar’s number where there’s a social cohesion thing that breaks down because everyone can’t know everything anymore. So they start to feel like, who is that or what does she do? And that’s alarming to this organism that has since maintained internal coherency. So that’s when it’s kind of start bringing in the big guns.

Rebecca: I joined Indeed, I think around 250. So it was still… and what a time. What a time. I think that our biggest mess was that we hired lots of people, but we didn’t hire managers nearly fast enough to manage the people. And so you ended up with very inexperienced managers managing people.

Michael: Or there’s the other version of that, which I’ve experienced multiple times is they knew to do it, but they just promoted from within. Which is fine, because those are the people that you know and you trust. That sounds human to me, but they’ve never done it before. And when you’re at more established companies, when you’re hiring these directors or VPs at a startup, you’re like, “cool, you’re VP at blah.” And you’re like, “great, that’s awesome. How many years have you been doing that?” and they’re like, “three.” “oh, cool. Three years as a VP” and they say “I mean like three years as a leader,” and I’m like … You are not a VP. And that’s not a, that’s facts in terms of a skillset that I think is important for a senior leader versus a director versus a frontline manager, blah, blah, blah. So there’s a lot of that going on, which leads to just, and there’s also this AI, this unbridled enthusiasm, like, well, I’ll just figure it out. It’s like, yeah, love it, go for it. You don’t know. So you’ll figure it out quickly, I’m sure, because you’re smart, but you’re gonna screw up.

Rebecca: I have one more question about AI and then we’re gonna talk about other things again. I said to this group last week that I can’t think of another product that people have bought, that software organizations have bought, without a clear hypothesis of what would change or how the productivity would be achieved. Because it’s a very, there’s the very naive, like, well, people can type faster and so they’ll produce more code, but more code doesn’t lead to more productivity.

Michael: The thing is, what you just described is pretty much, I mean the technology is different, but it’s pretty much exactly the same as what we saw with the internet bubble. It was like, oh, we’re gonna just disintermediate the middle man, and we’re gonna do this, this, and that. And it was. That pets.com thing that we talked about earlier, that was all just like, cool, well this is the new shiny and it’s gonna give us stuff without any thesis. And by the way, it did result in amazing stuff, but how many of those companies are gone now? Like 98%. Right? Whatever that number is. I made that up.

Rebecca: But e-commerce wasn’t a fundamentally terrible idea. No, exactly. Chewy.com exists today, right?

Michael: But I think that e-commerce thing for AI, whatever that is. I don’t know what it is, is still to be defined is that we’re gonna, and three years ago we’re like, oh yeah, it’s the AI engine that does this thing, which is blah. We’re still defining that. And again, I think the enthusiasm is it’s a little less feverish than it was six months ago, but I don’t think we’ve actually realized what it is yet.

Rebecca: No, and I also think with the internet, I remember trying to read New York Times dot com on September 11th, 2001, and … not so much. It was a thing that they were doing to see what the internet was about, but I don’t know that they had any big transformational ambitions around.

And of course it has become a transformational tool for media and for lots of other people, but I feel like there it was more of an experiment and now there’s pressure to make AI show value, whereas New York Times could have been like, “yeah, we tried that for a few years and we’d have to put up a paywall and we don’t wanna do that”, or whatever. But it was definitely, it felt a lot more experimental and a lot less profit or revenue driven than, you know. Now I look at people embracing AI, like, “this is gonna fix everything” or whatever. And the internet was never fixing anything. It was just this new thing.

Michael: It is a time of very high enthusiasm. I agree with that. But it is, it’s not what it is going to be. But again, there’s a set of folks when I’m talking with folks about this and this happens less and less ‘cause it’s permeated the membrane of the world right now.

But there was a time where I would talk to people and I’m thinking of those folks that are pressuring you to find a number or whatever, where they’re just, they’re using the barest of facts or opinions or the thing they heard from Bob over in the other team to say like, cool, we gotta get on the AI train, or this sort of thing.

And when I talk with these folks, I call them the fakers. I talk to these folks, and again, engineer, I’m not trying to be judgy about it. I’m just trying to be a person who thinks about stuff, but I say, well, what’s your concern?

And if I get an answer, which is sort of… “I don’t understand”. What they’re saying is they don’t understand, but they don’t say that they just ’cause they’re scared of not knowing and blah, blah, blah.

Whatever it is the fakers need to be, and, sorry, I’m making it sound derogatory and I mostly want to end people mostly. They need to know right, what it can and can’t do even if they’re not technical and whatnot.

We’re lucky in that we have curiosity and we’ve been around technology for so many years that we are aware of what it can and can’t do, but that’s the minority. The majority needs to understand what are the art of the possible here. And they’re not there yet. It’s still early. It’s still early days in terms of understanding. It’s our job to educate them though.

Rebecca: Yeah. Again, away from AI for a minute, although, oh, we’re trying so hard. Even this, there’s echoes of AI if you want to hear them. But you’ve been around for a while. You’ve been through downturns before. One of the things that I had not, I got laid off once in 2006 or seven, but otherwise my tech career has been mostly of the variety of ‘hire, grow, retain is the job of a manager.’ But now the job of a manager is to sometimes keep a team running after half of that team is gone.

And so what advice do you have? I think there are a lot of leaders who didn’t grow up with this experience and now all they know is the heady days of the 2010s and 2020, 2021. What advice do you have for that cohort? Who is maybe, I know I would be questioning whether I wanted this job anymore if that was what I was spending my time doing.

Michael: What a good question. Let’s start with the hard part. Let’s just say you laid off half of your team, for whatever the reason is. Well, number one is, do you understand why we are here now? You, the leader, because you gotta know, right? We are, whatever that is, we are this much revenue. This is what we’re doing right now. This is the next year that we see, these are all facts in a spreadsheet or wherever it is that exist. And they should frame how you’re talking to this about your team and you’re asking more human question than what I’m saying right now, but I’ll get to the human part of it in a second.

You have to be able to tell that story, and I don’t mean story in a negative way. You gotta be able to explain it. And by the way, just, this is not your question at all. If someone can’t explain that to you, if you can’t understand it, then I’m one of those folks that needs to get the heck out because something else is wrong there. And I’m not thinking of any particular company. But if it’s like, “well, we just gotta do this thing here” I’m like, “you are not telling me the answer to this thing. You are hallucinating the answer for this strategy.” So that’s number one.

But the human side of it is, and this is more the actual process of going on. Just understand it’s an open wound and humans, it’s a professional open wound. It’s not like you’re bleeding, but you are metaphorically bleeding. Those people that were sitting there in that office or on that Zoom are gone. And it’s just the mourning cycle and the grief cycle. And yes, you can say, okay, cool, let’s focus on these tactical things, but you just gotta let the humans go through that process.

It’s just gonna be awful for a couple of weeks and keep the meetings, they’re gonna feel empty and whatnot. But anytime I’ve tried to be like, “okay, cool, back to work, everybody!” I would never do that. But when you’re trying to, not trying to force what is ultimately a human emotional process, you’re gonna just get, they’re gonna be like “what the heck?” And there’s gonna be a point where they take a deep breath and they’re like, “okay, what can we do here?”

And then it goes back to the first part, which is, how do we get out of this? What do we need to do as a business to, for that initial story, to be able to do that. Is it quality? Is it more customers? Is it products? I don’t know what it is, but you gotta have that dance card ready to go for the team, which you can always point back and say, this fixes this horrible thing that we just went through, or this is progress toward this. It could be little stuff, but it’s that means by which you say, cool, this is within your control to do this right now. Let’s do this now. But doing that the day after layoffs, bad idea, there’s this swirl and healing that the team needs to go through. Pretty good answer.

Rebecca: The story that you just told of how to recover really requires a team that feels some sense of autonomy and ownership, right? Because if everybody, if half the team just got laid off and you still have no autonomy, you don’t get to decide what you work on, you’re still taking tickets, then that’s a very different,

Michael: Yeah, I mean the culture going into that real, but hypothetical in this scenario case is really important. And to your point, if they don’t feel the autonomy, if they feel this is just happening to them and they don’t understand why you’re in deep trouble, right? You’re just in deep trouble ‘cause those people are going to leave those,

I mean, who knows what’s going on in the market hypothetical scenario, but if they don’t feel like, “hey, cool, here’s the dance card and his is the thing we need to do.” And you’re telling them that and they’re like, “I was doing that before. And now half my team’s gone. So why is this going to be any better?” Anyway, it’s a complicated one, but it’s time, patience and small wins is my strat.

Rebecca: And I do think it’s really interesting how, again, in 2019: hire, grow, retain. All of our leadership meetings were all about, what’s our hiring pipeline and who’s accepted offers? And that was it. That was what we were talking about all the time. And then clicks that just turned off. And I know for me it was an awakening of, I need to care about things that seemed taken care of.

Michael: Yeah. Well, there, you cannot say this to your team, but you say it to yourself, which is that failure sucks. Especially when it’s happening and the team’s gone or whatever, psychological safety has been violated, but it is one of the more unique crucibles for you as an individual to be able to say: What happened here? How did I get here? What was I too trusting? Was I, did I have bad judgment? Whatever it is. It is truly, and again, hearing this when it’s bad is the last thing that people wanna hear. But you are, I’m like, “hey, by the way, the thing I’ll say to you in three months or six months is, do you realize how much you’re about to learn?”

It’s a crisis and you feel bad, but you are about to learn quite a bit and it’s the hardest way to learn stuff. I would wish we had other ways that we could do this without going through it, but in my head, for leaders, for managers, there’s these badges that you get when you go through these things and they are the worst parts of the gig.

And by the way, when you do them, you realize all the things you did wrong to be in that situation, you’re like, cool. I am never gonna let myself get there again. So a lot of my writing, one-on-ones, empathy, all these things are about early detection of things that could turn into horrible things, right?

It’s all about early detection because the thing is every time that some team is celebrating this win in the sight of a site outage or something horrible happened. And when they’re all clapping, patting each other on the back and going, “good job, we saved it!” I’m like “But it happened. Do you understand there was a disaster that we had to save and I respect everybody, but we screwed up. This is a screw up. And the save was amazing, but it doesn’t excuse the screw up.” So you gotta learn from those.

Rebecca: Two more questions before we wrap up. You mentioned earlier that you have changed some of your beliefs or understandings about management over the years, so tell me more about that.

Michael: Well, it’s really timely. It’s about AI again. I was in a delegation. I don’t know when this was. Maybe it was Palantir. Or maybe this was the first Apple gig. Anyway, the point was this, I’m like, hey, manager, stop coding. I wrote that. I wrote that a couple of times. I took a lot of heat for that. It was super bad advice. It’s bad advice for engineering leaders.

And yes, there’s CTOs and VPs and SVPs and blah blah out there that are probably not doing this. But I think losing those fundamental skills, and you retain them, it’s not like your brain empties out of knowing how to code or something like that. But I think that reminder to continually build is incredibly important. And I long before AI pivoted back to ’yep, you’ve gotta have a side hustle‘. You’ve gotta have something else that you’re doing. And it is a little bit of, ’do I smell like an engineer to my team?‘ But it also, it’s mostly just. And we just learn this big time. It’s mostly just you. It’s such a fast changing industry that if you’re not sitting on and staring at it and playing with it and tinkering with it and building with it, you’re gonna get left behind.

That’s my problem with the fakers is they’re doing just enough and I’ve been this person, this is why I can critique them. I was this person with AI until six months ago, I was like, “yep, yep.” And by the way, senior leaders, we are world class at sounding like we know. We talk about, we’re so good at that.

Rebecca: So your whole job, right?

Michael: Yeah, exactly. And that’s really bad advice. So I changed a lot on that and I always have something. It’s not, I’m not checking in code to Git for the company or anything like that, but I have things that I’m doing and I’m going to continue my education. That’s my main thing that I went back and forth on.

Rebecca: Yeah, and I find I was in the camp of not coding and AI has actually, I’m having a blast. It’s kind of wild. And I don’t need to know exactly the syntax of TypeScript types and like, but I’m still thinking and building and I love it.

Michael: Yeah. And I’m jumping into the code sometimes and I was talking to robots and it’s like, “okay, well I’m gonna build you this app.” I’m like, “okay, well,” and I eventually, I’m like “what kind of app are we building?” It’s like, “oh, it’s a Node app.” I’m like, “oh, I’ve never done that before.”

Rebecca: Woohoo! I’m building a Python app right now. Same. Same.

Michael: No, I was like, and I was like, I’m like, “okay, cool. Explain this to me.” And then I’m poking around the code and looking how it works and all that stuff. So that again, I think for folks who have been educated in computer science, and can understand when the robots can understand what’s going on. I think we have superpowers now in that we have this ability to have this, again, endless infinite set of interns helping us do stuff.

Rebecca: Yeah. Well, here’s your last chance to talk about AI or not. If you have a creative way to not to, I’ll be impressed. But what’s your, it is August 25th, 2025. What’s your hot take about the industry right now?

Michael: Ooh. Oof. Wow. That’s a dangerously simple question. You know, I always say this, it’s a little unsatisfying and I kind of said this a couple times already.

When I talk to one of the fakers and I’m trying to educate them, and I am like, “This is a big deal. It’s bigger than you. It’s much bigger than you think. It’s for real, and perhaps most alarming or opportunistically. I don’t think we’re even close to the final product.”

And I’m not talking about an advanced general intelligence or anything like that, that’s a whole other thing. But just the way that I feel that you’re going to work with your tools is gonna change and the way that you’re gonna work with your personal information is gonna change. And this app focus that we’ve been doing for 20 plus years or so. Those tools exist to allow you to go and manage a browser or to set tabs or whatever. You don’t really wanna do a lot of that. You just wanna know is the high school, what time is lunch at the high school? ’cause I don’t wanna be in town when that sort of thing. There’s this whole set of things. I think we’re years away from this, but I think you’re gonna see just a transformation in terms of how the humans and the computers are getting stuff done.

And I think in the right hands (danger) I think there’s just some really cool, I think there’s some really cool things coming to help people manage their data, get their stuff done, ask questions, get good answers, blah, blah, blah, blah, blah. It’s all very nerdy right now. Prompts are nerdy.

I love writing. I’m a writer, and prompts are fun to watch what it does, but is your average human being gonna want to write three sentences to figure out how to reconcile their calendar or whatever the heck that is? So I think it’s a long way of saying, I think it’s very early days in terms of what companies are gonna produce. And I think it’s exciting. I can see all of the nuts and bolts and how they’re going and magically doing things right now.

So I am, I went through the social [media] thing too, so I also know deeply how it can be weaponized by evil people. And it’s already happening, by the way. But I’m also, I’m generally an optimist, so …

Rebecca: Well, again, it’s like the internet all over again. We’re like editing Apache configs right now, I feel like. Yeah, that’s where we’re at.

Michael: Yeah, exactly.

Rebecca: Well, Michael, it has been, or Rands, whatever. Do people call you Rands in real life?

Michael: People call me Rands when they’re talking about the blog or I’m speaking at a conference. Got it. I know who they’re talking about. But anything otherwise.

Rebecca: Otherwise, your name is Michael. Okay. Well, Michael, this has been so great. I’m so glad I got to meet you earlier this year and that you agreed to chat. This has been awesome and I can’t wait to read your next book and

Michael: Oh, great. Now I gotta go finish writing it.

Rebecca: I gotta find my copy of Managing Humans. I was hoping to have it hold up. I know I own it, but don’t ask me where right now.

Well, that’s the show. Engineering Unblocked is brought to you by Swarmia, the engineering intelligence platform that’s trusted by some of the best software companies in the world. See you next time!