Balancing Operational Excellence and Innovation in Engineering with Meir Wasserman

00:00:08 - 00:00:32 Stephen Koza
Here's something most engineering leaders won't say out loud. It's the operational stuff. The tickets, the firefighting, the stuff that keep the lights on. It's not fun, but it matters just as much as building the new shiny things. But if you let it consume your team, you lose the people who came to build things. And my guest today is spent a number of years figuring out that balance.

00:00:32 - 00:00:52 Stephen Koza
It's some of the biggest names in gaming, from EA Sports to Amazon Games and to 2K games. He's an author and a conference speaker and has an obsession with operational excellence. Welcome to Tech Pod Talks. I'm Stephen Koza, and today my guest is Meir Wasserman. Hey, Meir. Welcome.

00:00:53 - 00:00:54 Meir Wasserman
Hey, thanks for having me, Stephen.

00:00:54 - 00:01:21 Stephen Koza
Oh, it's my pleasure. I'm glad we were able to connect for this. So let me do a bit of an intro on you. Meir is an engineering leader who's built in skill teams at some of the most demanding environments, and tech and gaming. He spent years of Electronic Arts working on EA Sports titles. I spent a number of years at Amazon Games driving operational excellence across their teams, and today is head of engineering at 2K games.

00:01:21 - 00:01:37 Stephen Koza
And as I mentioned earlier, he likes to write. He's got a blog. You might catch him as a speaker at a conference. And so he's got a lot of interesting things to say, and I've enjoyed getting to know him over the last couple of years. And so great to have you on. Welcome, Meir.

00:01:37 - 00:01:39 Meir Wasserman
Thank you, thank you. Let's chat.

00:01:39 - 00:01:54 Stephen Koza
Yeah, let's do it, man. So I gave a little bit about your career arc, EA, Amazon Games. Maybe you could just walk us through the thread and what connects all of that. What's the through line through some of your experience?

00:01:54 - 00:02:18 Meir Wasserman
Basically I've been in the games industry my whole career. You know, as a child, the growing up, at the time I did there was there was not really a guaranteed future in video games. There was lots of like, you'll never make money, you know, playing video games type of thing. That's clearly no longer true. Games as an industry is bigger than music and movies combined, and it's been that way for probably almost 15 years now.

00:02:18 - 00:02:45 Meir Wasserman
And so I'm excited to, like, have made a career in games, started as a software engineer, making console football video game and worked, you know, both in video game game companies on the developer and publisher side and also a big tech at Amazon. And like these things, although games has been the through line, there's been lots of technical challenges and hurdles and across the different companies and the different like sizes of problems that we've been dealing with.

00:02:45 - 00:03:12 Stephen Koza
Nice. I, I did not know about the size of the gaming industry relative to the other ones you mentioned. That's a that's a good takeaway for me. Tell me a little bit about your your interest and desire to drive change. I've, you know, as we've worked together, that's obviously something that made an impression on me. I would bet that you've probably been like that in other roles and other companies before we got a chance to meet.

00:03:12 - 00:03:18 Stephen Koza
How do you think about driving change and what does that look like day to day when you come into a new organization?

00:03:18 - 00:03:45 Meir Wasserman
It's actually it's hard. Like driving change is hard. Not because identifying what needs to change is hard. That's, I'd say sometimes obvious, like at least for fresh eyes to an organization, but because it you're mostly dealing with people problems, most or most problems we solve as technical leaders certainly are people problems. It's not what's the technology choice we should make?

00:03:45 - 00:04:13 Meir Wasserman
Or you know what works best in this moment? It's how do we get a team or a an organization to align that this is a problem that we need to solve and then go ahead and execute that which that's what I when I say people problems, that's what I mean. It's not like a person behaving badly. It's like just dealing with the kind of inertia and, you know, of human nature and the desire to stay the course more than to make drastic change.

00:04:13 - 00:04:37 Meir Wasserman
So, you know, first step is like identifying problems. Yes. But then you have to form relationships and you have to socialize the problems and you have to make them understandable to non-technical crowds. Generally. That's like not always a skill set that everyone has is like, take this deep technical problem and like up to level at to a point where it makes sense but doesn't lose its impact and understanding and accuracy.

00:04:38 - 00:04:52 Meir Wasserman
All those things like have to be true before you can even be can to start to institute change. And then once you have, you know, the charter to make that change, you actually have to go execute, which is a whole other set of problems. But it's much more about like, you know, the technical side of the job.

00:04:52 - 00:05:15 Stephen Koza
Yeah. One of the things I find myself saying to sometimes is most technical problems don't have a whole lot to do with the technology. There are more people problems, and I always try to qualify that and say, I don't mean bad people, I just mean processes and organization and structure and culture and how people go about using the technology.

00:05:15 - 00:05:16 Stephen Koza
Would you agree with any of that?

00:05:17 - 00:05:37 Meir Wasserman
Yeah. I mean, yes. Yeah, I would agree with all of it. And and I agree it's not about bad people. It's it's like the intentions are almost always fine, even good. But there's also just like human nature that we, that we have to struggle against. We're all subject to it. Me, you, everyone. Like, these are things that we all struggle with.

00:05:37 - 00:05:54 Meir Wasserman
And as a leader and a technical leader, you have to find a way through anyway and, you know, get people on board. These are. This is like a never ending part of the job. Like I did it today as part of my job. I did it last week a lot, and I'm sure I'll do it tomorrow as well.

00:05:54 - 00:06:22 Stephen Koza
Yeah, totally. Well, since we're talking about leadership and teams, let me dig into that a little bit more with you. I'd love to hear your philosophy on how you try to keep your teams focused on what matters, while balancing that with doing the stuff and doing it well, the operational stuff, the day to day, the low that kind of has to get done.

00:06:22 - 00:06:50 Meir Wasserman
Okay. This is this is a constant challenge. There is a never ending literal, never ending infinite amount of operational work that a team could do. And there is always a seemingly infinite backlog of like customer value. You could also provide, right, new features or bug fixes or something like that. And so as a leader, you have to kind of pick through like, well, what what do you choose to do?

00:06:50 - 00:07:12 Meir Wasserman
You can't kind of do all one or the other, but you also have to have a point of view on like what what actually makes sense. And so the first thing I try to do is like, take the if you're an engineer who is like unduly affected by a piece of like an operational burden of some sort. You're going to want to fix that.

00:07:12 - 00:07:32 Meir Wasserman
And maybe you do. Maybe you do fixed on your own time and life is good, but you're going to, like, at least want to advocate for fixing it through, through your management chain. And that happens a lot. And it's it's a very powerful motivator for the people who are feeling the pain. But as a leader, you also can't just say yes to all that because it will it will consume all your time.

00:07:32 - 00:07:59 Meir Wasserman
So I think the answer is like, you have to basically fix the problems that have demonstrably caused a bad time in your system, like an outage of some sort. Or you can prove that the time, the cost to the team is like really slowing down innovation in some way, some metric that you can show. This slows us down enough that's actually worth fixing.

00:07:59 - 00:08:27 Meir Wasserman
And everything else, even though it feels bad and does, is kind of a paper cut. You just acknowledge, like, we're not going to fix that and it's because we want we'd rather drive customer value and do new features. So that's that's kind of the heuristic. It's like if this is this problem actually caused a like a production issue, an outage or any sort of like negative customer impact, we should probably go fix that and we will.

00:08:27 - 00:08:48 Meir Wasserman
And if we can show that like, yeah, this other type of problem didn't cause an issue, but it slows us down enough that over time, if we don't fix it, we will deliver less value that we should also fix, and everything else we just kind of say, yep, we know it sucks and it's not amazing, but also like we'd rather we'd rather deliver value.

00:08:48 - 00:09:12 Meir Wasserman
I'll add one bit though, which is to say, I think AI agents are changing the the math here. I think because we can start to give AI agents like problems and have them autonomously solve them. It will probably we talk about it right now in terms of like, hey, this is like we can build more features and deliver more value, but we can also like fix more operational debt.

00:09:12 - 00:09:20 Meir Wasserman
And so that's pretty exciting and cool. And we're not there yet. But I'm very excited to start leveraging AI agents for that.

00:09:21 - 00:09:53 Stephen Koza
Love it. Yeah we share your excitement on that. And maybe we'll come back to that when we dig into AI a little bit. I love the framework. I was actually going to ask you about that. So how do you know when something matters enough to focus on it? I'm curious when it when you suspect a team is doing too much on the toilet, the low and not paying enough attention to the stuff that does matter, the stuff that kind of meets that threshold for you.

00:09:53 - 00:09:58 Stephen Koza
What do you look for? How do you identify that and if you see it? What do you do about it?

00:09:58 - 00:10:18 Meir Wasserman
There's a perfect world where you have you've collected enough data that it just jumps out at you. But I don't know who lives in that world, I certainly don't. So. So I think actually what you look for is, you know, are you are you delivering customer value at a fast enough pace that you think as a leader is acceptable based on prior experience?

00:10:18 - 00:10:38 Meir Wasserman
That's that's one bit of it. And the second is you have to pay attention to your anecdotes from your ICS. For me, you know, engineers. And if they're if they can tell you, like, yeah, this thing is a constant problem. It's like always breaking or it's slowing us down. You may not have captured that data yet, but you probably have some pretty smart people that can tell you that.

00:10:38 - 00:10:55 Meir Wasserman
And I know that flies in the face of the previous thing I said where they don't complain about everything and you'd have to pick and choose. You know, it requires high judgment, right? Like, this is part of the job we have as leader of any kind is like, you have to pick through what you're hearing and and apply high judgment.

00:10:55 - 00:11:09 Meir Wasserman
And the longer you're around it, the more you start to see some of the patterns there. I think the answer is like, if you have the data, use it. And if you don't have the data like try to seek it out, but there's there's not going to be an empty list. It will be a, a set of some stuff that you have to sort through.

00:11:09 - 00:11:11 Meir Wasserman
It's never nothing.

00:11:11 - 00:11:39 Stephen Koza
Yeah. Like that. So it sounds like it's kind of a combination of some instinct and some walking the halls, so to speak, and as much data as you have and can take advantage of. But data by itself doesn't do it. Let's talk about doing things at scale when it comes to focus and operational discipline. So I know you think you were credited on a bunch of games when you're at EA.

00:11:39 - 00:11:54 Stephen Koza
Big company, big catalog there. What did. What did shipping at scale teach you about keeping your team focused and running? Good operational discipline and excellence. What would you take away there that you've carried with you since.

00:11:54 - 00:12:25 Meir Wasserman
The takeaways I've had are less about like as you as your organization scales like what are your opportunities to like continue to drive operational excellence through every part of it. And it's more about as your organization scales, how do you stay in touch with it? So what I've recognized in my career is that you start your career. Maybe you're managing a team, maybe it's 5 to 10 folks and you're you kind of know everything about that team because it's small enough that a person can know everything about what's going on.

00:12:25 - 00:12:48 Meir Wasserman
And so when you're asked to, you asked questions about what's going on. You have like full depth of knowledge, essentially. That's pretty normal. But we all know that doesn't really scale if you're running several teams. So 5 to 10 people, because how can one human possibly know all that. And the transition between those two is not there's not a moment where it's like where you feel as a leader.

00:12:48 - 00:12:59 Meir Wasserman
I now have this big of an organization, and so I'm no longer expected to know every detail about my organization. And never there's not like that that one moment in time where it's like, I am free from that burden.

00:12:59 - 00:13:03 Stephen Koza
Yeah. No, no, nobody sends you a notification that morning and says you're here.

00:13:03 - 00:13:28 Meir Wasserman
And it's just it's just this transition that, you know, probably is organic from for a lot of folks. And so then you find yourself in a position where you still feel compelled, you're still be asked questions that require depth of knowledge about a much larger organization. And you'll feel compelled to know all of this. And you kind of have this, this moment where you either have to recognize, like something's got to change in my system or you don't recognize it.

00:13:28 - 00:14:02 Meir Wasserman
And I think you just kind of fail as a leader, or at least you're ineffective as a leader because you're just trying too hard to stay connected to everybody in detail. And the the mechanism there is like, you need a series of mechanisms that are you need a way to feed data through organization that gets you broad knowledge about what's going on and then, when needed, lets you dive in to any, any piece of your organization with high trust, with an understanding that this is to give you the context so that you can surface the context more broadly across organization, all that type of stuff.

00:14:02 - 00:14:20 Meir Wasserman
So you need kind of constant, broad knowledge that's expected. Now once you scale and then you need the trust to dive in without that being like disruptive or like, hey, my ship levels here like this is a problem. Like none of that. Like it should be also very normal. And if and that those are the mechanisms I've certainly created.

00:14:20 - 00:14:40 Meir Wasserman
And then the last bit because again like you just can't have full knowledge about every little working bit is you have to get really comfortable translating the broad knowledge you have into kind of a non-technical language, and also going a little bit deep on that without having done every single deep dive. That seems to happen a lot too.

00:14:40 - 00:14:45 Meir Wasserman
And that's part of the scaling problem. That just doesn't happen when you run one team of 5 to 10 folks.

00:14:45 - 00:14:58 Stephen Koza
You mentioned skip levels. Any other tips tricks, techniques? Once you get to that magical transition point that nobody tells you you've arrived at, that you've started implementing or using over the years.

00:14:58 - 00:15:17 Meir Wasserman
The first time I became a manager of managers, I continued to have monthly one on ones with like every senior engineer in my org, just like a bunch of folks that felt like I want to stay connected so they, you know, they have an outlet and also can get a different point of view. And the volume was too high and then eventually concluded, like, well, I can't do that.

00:15:17 - 00:15:39 Meir Wasserman
But also, if I just cancel all of these, like I'm missing a key element of like organizational leadership, which is actually being connected to the, you know, to the people in the organization, not just your directs, but like key people throughout. And so I tried something that is called an N on one. It's just me. And then through every direct I have, it's all of their directs.

00:15:39 - 00:15:54 Meir Wasserman
So like like three people might be like eight people, whatever. It's like me. And then all of them in a small group setting once a month. It replaced a monthly one on one, but it's still like a tight group in this. People who are appears so it feels very tight and it just ends up being and it does not include their manager.

00:15:54 - 00:16:25 Meir Wasserman
So it's that person is not in the room. And it's just a way for like, like people to still feel connected to their skipper level, get a different perspective, maybe on whatever they're hearing about or whatever they're doing, but just kind of get like the one level of perspective, and that's nice and very useful. But I think what it also does is it creates the environment where if you then need to talk to them, maybe their managers on PTO, or maybe it's, you know, for whatever reason, you have to deep dive with one of these skip levels folks.

00:16:25 - 00:16:30 Meir Wasserman
They don't feel threatened. They don't feel like about it. You've already kind of got their relationship going on.

00:16:30 - 00:16:51 Stephen Koza
One of the things I realized at some point was not everybody feels comfortable just openly communicating, you know, two levels up. You know, some people do and it's great. But then it's like, you know, I've never talked to this person and that's kind of weird. And then you reach out and they welcome it and it's appreciated. And I like that idea.

00:16:51 - 00:16:55 Stephen Koza
And and on one I may have to adopt that.

00:16:55 - 00:17:10 Meir Wasserman
Yeah. And like as an engineer like it kind of just scales because no matter what level you're at, it doesn't get too big, you know, unless you have like 50 tricks and then don't do that. But, you know, you have like a normal, like set of direct reports, like it's just it kind of just works at every level.

00:17:11 - 00:17:22 Stephen Koza
Yeah. Jensen Wong and Nvidia's has openly talked about how he doesn't do one on ones, but I think he has like 50 direct reports or something insane. You can look it up. It's more than eight. I'm positive.

00:17:22 - 00:17:38 Meir Wasserman
With that. Yeah. It's yeah. Whatever the number is is large and I agree like that wouldn't work. But when I hear like I don't do one on ones, I don't think like, oh that's wrong. I just think like, okay, you know, it's fine. Like that's a different way to lead. That's not a problem. But like for me, I still derive a lot of value out of that.

00:17:38 - 00:17:59 Stephen Koza
Mark Zuckerberg's talked about it as well. The difference with those guys is, you know, the people that report to him are very senior, very capable. The thing that Zuckerberg has said is something like, if you need to have a one on a regularly scheduled, recurring one on one with me, you probably shouldn't report to me. I was like, okay, actually, that makes a lot of sense.

00:17:59 - 00:18:11 Stephen Koza
Cool. Let's, let's, let's keep going into leadership a little bit more. What's a leadership lesson that you learn the hard way that maybe you wish somebody had told you way sooner?

00:18:12 - 00:18:33 Meir Wasserman
Good question. So first I'm going to amend this question a bit. I think every lesson has to be learned the hard way. Every one of them. Right. It's like, you know, you're told not to touch the hot stove, you're going to do it anyway. And then once you do, you don't do it anymore. And I and I think, I think even though it's like a story for children, I think it's just like mimics human nature.

00:18:33 - 00:18:59 Meir Wasserman
All throughout our lives we can consume lessons, you know, through a blog post or a book, or maybe someone listening to this right now is like going to get a lesson. But until they experience the hardship associated with it, it probably doesn't stick that well reframed. But either way, like I think, I think I have some that here, which is I was a bit surprised when I kind of moved into leadership and got got my feet under me.

00:18:59 - 00:19:23 Meir Wasserman
I think, you know, like leadership, you kind of understand, like it's about having a vision and helping a team get to that vision. And that's not just like it's not just like helping them see the vision, but also like execute against that and like push it forward. That that wasn't surprised by what I was surprised by is that a leader is required to create urgency.

00:19:23 - 00:19:47 Meir Wasserman
And I feel like that no one told me that. And you can just kind of see like the organizational inertia in corporation, the corporate environment, you can kind of see like how all the little things add up to this, like this big problem where we're work is slow. It might be like stakeholder alignment or like socializing the messaging across broader folks or whatever it is.

00:19:47 - 00:20:07 Meir Wasserman
And that might exhibit as like, I set up a meeting or I sent a slack to a person and they didn't get back to me the same day. And so now it's the next day. And what do I do if I'm trying to be a nice human being? I might wait a few more days and then ping them again and say, hey, not to be a bother, but did you get that?

00:20:07 - 00:20:42 Meir Wasserman
That message. But if I was really trying to accomplish a goal, I wouldn't wait that long, right? And so I would actually just next day like check in and advocate for my priorities and try to make sure that, like, whatever step in the process that is, it's not blocking me from moving forward. And so leadership, being a leader and that you being the impetus for urgency in your organization was definitely a learning that like it caught me off guard a bit when I kind of reasoned through what was happening and put it in like these terms, I was like, okay, I can do that.

00:20:43 - 00:21:03 Meir Wasserman
Like I can be the person that drives urgency. And then as I've gone through and expressed that several times throughout my career, I've realized like, yeah, actually good, good leaders recognize this and do this. And that's how you get results. That's how you get value out to customers, or you fix some operational problem or whatever it is like, you create the urgency around it.

00:21:03 - 00:21:21 Meir Wasserman
And then I'll just put a big asterisk here and say, it's not like, no, no one's saying like create a fake deadline or create urgency about things that don't matter. There's plenty of work I do that I push along at the right pace, which is not like the hyper thing, but for like your top priorities, it will kind of only move as fast as you make it move.

00:21:21 - 00:21:23 Meir Wasserman
And so it's up to you as later to make it move fast.

00:21:24 - 00:21:52 Stephen Koza
Couldn't agree more with that one. I think my team would probably vouch for me on that. And like you said, everything can't be. Everything can't be a priority. Otherwise nothing is. The Frank Superman, the former CEO of snowflake, through their Hypergrowth and IPO, wrote a book called AMP It Up. That was my takeaway how to create urgency. There was some good nuggets there, but like you said, I probably didn't absorb enough of it because it was a book and not a hard one.

00:21:52 - 00:22:17 Stephen Koza
Lesson. I probably appreciated the book because I've screwed it up in the past, so I like how you think about that. How has your style evolved throughout the years as you've worked across different companies, different kinds of teams and orgs? How would you say you've grown, adapted, change? What do you think your team would describe you or describe your leadership style as?

00:22:17 - 00:22:37 Meir Wasserman
Oh, I, I don't know how I could I'm not the right person to ask how someone else would describe me because, like, I'd probably be too hard on myself, but also like, it would be embarrassing to be like super wrong. So I'll just say, like, I think in terms of evolution, it's I've come to like appreciate that. Like really strong leaders.

00:22:37 - 00:23:00 Meir Wasserman
They, they can they can really boil complexity down to some simple something simple. It doesn't have to be like oversimplified, but it's like something you can get behind and deeply feel and understand. And I, I've worked with several leaders who were good at that, and I recognized like, oh, I could try to be good at that too.

00:23:00 - 00:23:20 Meir Wasserman
And then I tried to adopt that. And that's I think been a big evolution of my style is like pattern matching, like problems, whether they're human problems or tech problems or organizational nurture, whatever that kind of pattern matching is saying, like, what is this level up to? And how do we like frame this problem in a way that gets people up behind wanting to solve it?

00:23:20 - 00:23:39 Meir Wasserman
That's been a big level up for me, and that's been a way that I feel like I've been able to really get a large group of people around some problem space to go hammer it down into its component parts and then like, begin to make progress and then move on to the next problem space after we're satisfied. Didn't start there.

00:23:39 - 00:24:00 Meir Wasserman
Definitely started on the I own a team and I'm going to be an expert on that. Everything that team does, but that's that level of expertise is like useful in very certain moments once you go up the chain. But like definitely not not at the same level it used to be. And so now it's much more about like, how do you talk to your peers about the problems you're solving and how do you make it?

00:24:00 - 00:24:10 Meir Wasserman
Put it in a framing they care about at they understand. And then also how do you rally the team around that and that? Yeah, that's probably the biggest changes happen over the years.

00:24:10 - 00:24:44 Stephen Koza
You just made me think of a similar, I guess, growth lesson that I had. So I was fortunate that early in my leadership career, I got to build a new business inside a company. It was a new market segment with kind of different kinds of products and different kinds of customers. And it was me. And then it was a few people and it grew and grew and I, I probably had the same mentality like, oh, cool, I got a team and an org chart and, you know, this cool, new, different job.

00:24:44 - 00:25:10 Stephen Koza
But the nature of what we were trying to do required selling not just customers, but internal stakeholders like the product team had to be aligned to build the stuff we needed, and the marketing team had to get behind us to, you know, on and on. And so the big lesson I took away was, I think, similar to what you just said, which was I had to learn the hard way.

00:25:10 - 00:25:33 Stephen Koza
How do you cast a vision and get people excited about it and run in the same direction and committed? And we were very successful. You know, I maybe deserve a little bit of the credit, but there was a luck and all sorts of conditions. But that was it. Like looking back, that's what made us successful. We had this very clear vision on what we wanted to do.

00:25:33 - 00:25:46 Stephen Koza
Everybody was excited about it. People wanted to join the team and on and on, and I've carried that with me ever since. It's just so critical. You know, it's one thing for you to see it and believe it. But if nobody else does. You're gonna have a really hard time.

00:25:47 - 00:26:08 Meir Wasserman
Yeah, it it reminds me of, like, you know, some terrible product. You know, real world product would go like, oh, this is super easy. Just follow these ten simple steps and you're like, no, like, there's no process. That's ten steps and easy and so and and so for like how we like create the words that make sense for people.

00:26:08 - 00:26:31 Meir Wasserman
Like it has to be actually simplified. You know like hear these. I listen to like a lot of like physics and other types of like educational videos and like the people that really resonate. Like, I don't deeply understand this stuff, but I but I'm curious and when they can put it in terms, you understand, like, you know, you know, they're burying a ton of their complexity like that is 100% true.

00:26:31 - 00:26:49 Meir Wasserman
But you also feel like you get it. That is like that is not relegated to like difficult sciences. That is like every day in life, the problems we solve there also we can also like simplify to that nature where the complexity is hidden, but it's not gone and and people will rally around that a lot better.

00:26:49 - 00:27:09 Stephen Koza
You just maybe think of something out of you know, if you played around with notebook LM the Google AI, I don't know what they refer to it as like research tool. You can plug in a bunch of sources and it'll give you summaries and slide decks, and it'll create a podcast for you based on whatever you give to it.

00:27:09 - 00:27:30 Stephen Koza
So if there's a whole bunch of information you're trying to learn or distill, it's really cool. It'll give you like ten 15 minute AI generated podcast. It is so good at taking the complex and making it simple. I've done this a bunch of times for random stuff, and they always have these like dead simple analogies to take, like take a really complex technical topic.

00:27:30 - 00:27:48 Stephen Koza
And I have a technical background, but I'm a business guy, you know, I gave up my engineering career long, long time ago. So pretty much everybody else here at our company is smarter than me, which is okay, I like that, but it yeah, it helps me grasp stuff that would otherwise take a really long time.

00:27:48 - 00:27:50 Meir Wasserman
Yeah. That's awesome. I love that stuff. Cool.

00:27:50 - 00:28:11 Stephen Koza
Well, let's why don't we shift into AI? Because it's come up a few times. It's obviously the topic de jour. Correct me if I'm wrong, but I think you and I have talked about maybe you having some initial skepticism or reservations around AI, or I don't want to put words in your mouth so you can correct me there.

00:28:11 - 00:28:25 Stephen Koza
But how is your thinking about AI? Not just the tech, but the opportunity and where it fits and where it's going? How is your thinking about that evolved over it's, you know, call it since the ChatGPT moment.

00:28:25 - 00:28:51 Meir Wasserman
Yeah. So November 2022, so long ago very, very was very skeptical. And it wasn't the skepticism wasn't like, oh, this isn't world changing. That was, I think, kind of clear that it was it was more like, how does the world adapt to this in a way that doesn't like destroyer society? You know, it was like a lot more like touchy feely like that.

00:28:51 - 00:29:12 Meir Wasserman
But there are also some technical concerns I had like, oh, does the I've since gotten over these, whether I so I'll go there. But it's like, does this kill our junior pipeline? Like how do I though AI is very powerful in the hands of an experienced senior, what about five years from now when that person, ten years from now, the person retires and then the junior, like, never had the chance to learn those lessons.

00:29:12 - 00:29:31 Meir Wasserman
And that was like one concern I had. And then the other, another concern I initially had was like, well, does it kind of is it kind of making this dimmer? Like if we don't have to think about the code interface to machines, then and that's obfuscated from us, then when when things go right, not a problem. It's great.

00:29:31 - 00:29:54 Meir Wasserman
But when things go wrong like are we are we is a population capable of reasoning through why it's gotten wrong and then actually solving it? That one and not so sure yet. I think the jury's out if that's going to be long term harmful. But I think the junior pipeline thing is going to be just fine, because what I've seen happen as the especially as the models have improved, is like people are using AI.

00:29:54 - 00:30:14 Meir Wasserman
There's kind of two ways this happening. And, you know, some people are like incorporating it into their current workflows and that's like the slow way of going about it. And then there's people who are just saying, like, what if I never had a current workflow? How would I use this tool and how would I? How would I then adopt, you know, make whatever I want to make whatever outcome I want to have happen, happen.

00:30:14 - 00:30:33 Meir Wasserman
And that's where I think, like the real power is going to start showing through. And the juniors coming out of college now are, you know, they're kind of grown up in this world now. It's not that old, but it's also not that new. And they're showing that they're just as capable of like critical thinking as the previous juniors.

00:30:33 - 00:30:52 Meir Wasserman
They have all the same kind of gaps in like experience and whatever. But we also can still train them and teach them the way we, you know, not the way we used to, but like with kind of the same mental models. So I'm kind of past the skepticism part, not kind of. I'm very much I would say like this is actually super cool.

00:30:52 - 00:31:14 Meir Wasserman
And where I've begun to go is like, okay, this is a provably useful a lot of people are using it for all kinds of stuff, especially with agents. And I just saw a post on LinkedIn a day or two ago about some peer of mine had like 36 agents running simultaneously, produce X number of PR like it was like impressive stuff.

00:31:14 - 00:31:34 Meir Wasserman
I think the nature of the coding job was changed. I think instead of saying like, how do I use AI to generate code a little faster and then go through the normal process we used to go through, I think it's become, I think like the way we have to think about it now is, is like, what if I never have to look at code again?

00:31:34 - 00:31:56 Meir Wasserman
AI generates it, AI reviews it, and if and, you know, tests it and if it all passes, I feel good about that outcome. And I never once looked at it. And that's where that's I think the future we're headed towards is not where we're at today necessarily, because not all the systems are set up for that. But it's like, not that also not that far away.

00:31:56 - 00:32:17 Meir Wasserman
And so then the mindset shift I'm trying to work through with folks is like, if that were true, if you weren't burdened by having to ever look at code again, how would you change how you engineer to create outcomes? What would you do then? That's that's where we want ahead. It still requires engineering brains. It still requires rigor.

00:32:17 - 00:32:26 Meir Wasserman
It requires like actually a lot of upfront rigor to make sure that you're creating the right outputs. But it doesn't necessarily require you require you to either write a line of code or look at a line of code.

00:32:27 - 00:32:50 Stephen Koza
Yeah, I, I to wonder about the entry level jobs. I'm with you. I, I think it's a solvable problem. My opinion is the jobs themselves just need to change. Because if you think about the example I always give is what is a first year law associate? Do the read and they summarize them and they write briefs for partners.

00:32:50 - 00:33:12 Stephen Koza
Don't need them to do that for AI does that way faster. And you know, probably better in a lot of cases. They do that because it's a way to learn. So okay, what's the thing they can do to learn that isn't just taking, you know, 100,000 words and making it 2500 words. So I think the jobs just need to change a little bit.

00:33:12 - 00:33:34 Stephen Koza
But I'm generally an optimist. I think the universe will probably figure that one out. I also wonder about the coding thing because, you know, the stat is like more code I think will be written the next 12 months than all of humanity. Something like that to this point. But you know, the the models are never going to do well.

00:33:34 - 00:33:54 Stephen Koza
Never say never. They can write great code, but there's a, there's like a reasoning and judgment element. You know, even when I'm doing stuff, Claude or whatever I'm using, I'll do it one way. But I'm like using the critical thinking to say, well, like, what if we did it this way? Or what if we organize it a different way and it's like, yeah, that's a great idea.

00:33:55 - 00:34:12 Stephen Koza
And so it can't make all those decisions on its own. Maybe it will one day. But just like the other example, I think the jobs just change a little bit. Well, and perhaps a lot over time. So it'd be interesting to see how things play out. But you can tell them generally an optimist on this stuff. And, you know, I'm rooting for humanity.

00:34:12 - 00:34:13 Stephen Koza
I think it'll be okay.

00:34:13 - 00:34:33 Meir Wasserman
Yeah, I think it'll be okay to. You know, it's funny, I hadn't heard this stat about more code in the next 12 months than before, but I'll tell you, like any engineer, that's experience will say like more codes, not better. It gives me like, no hope to say we're going to generate a ton of code. I would actually feel a lot better to say, like, we've actually found a way to generate less code and have a less surface area for attacks.

00:34:33 - 00:34:41 Stephen Koza
Yeah, yeah. The the unwritten part of that is like more code will be written that doesn't get reviewed than ever before.

00:34:41 - 00:34:45 Meir Wasserman
Yeah. I mean, you get reviewed, but like in theory by also by AI.

00:34:45 - 00:35:12 Stephen Koza
We talked about kind of how you have thought about AI, how that's evolved. Let's talk about the value, whether this is something you're actively trying to figure out. Or maybe you've got peers or colleagues. One of the things we see is everybody's experimenting. You should be. That's how you learn. I think every boardroom conversation these days is, you know, what's your AI strategy or the metrics?

00:35:12 - 00:35:30 Stephen Koza
What's the adoption? What are you doing? But there's also a lot of efforts that don't produce value, at least not yet. So do you have a framework for how you think about this stuff, how you separate opportunity from hype and value from noise when it comes to AI investment?

00:35:30 - 00:35:54 Meir Wasserman
Good. Good question. So now we're definitely in the try everything experiment moment because we're not trying to like like pre assume that like there's no value to be found down a certain path. I think we will need that framework eventually. And especially you think it's well known like the economics of AI are like not positive for the companies that are providing the service.

00:35:54 - 00:36:15 Meir Wasserman
And so will that change. If it does certainly like the expense, it'll get way more expensive. And you have to like the frameworks will be way more important to understand. Like is it worth using here? Worth not using there. You know, what I found is like when I use it as like, well, if I do like a lot of quick iterations, it's kind of expensive.

00:36:15 - 00:36:31 Meir Wasserman
But if I put a lot of thought up front and then and then give it give, you know, direct to my AI to make a bunch of changes like well thought out changes, I have traded that expense for my upfront time, but in aggregate, it's still faster than it would have been before if I had just manually done it.

00:36:31 - 00:36:54 Meir Wasserman
And I think it's a little tempting to just like, be very conversational and do quick little stabs. I think where we're headed, though, economically speaking, is it's going to be way more important to be really rigorous and intentional upfront with what you're asking so that you waste less of the AI token usage on paths that end up not working.

00:36:54 - 00:37:19 Stephen Koza
You're right. I think it's not universal knowledge what the economics look like for the frontier model companies. Fortunately, it's easy for them to raise capital, but I think the hope is that the, you know, the the cost ends up the cost to deliver it comes down. But but they have talked about having to adjust their pricing model because it's it's essentially flat.

00:37:19 - 00:37:35 Stephen Koza
Whether you're logging in a little bit or you're hitting the usage limit every single day. So my suspicion is they they shift to more of a true consumption model or maybe, maybe even a value derived kind of model.

00:37:35 - 00:38:06 Meir Wasserman
And, you know, this is not what you asked. But for SaaS companies, well, before AI came around, there was kind of a problem where they want to charge usage just for any usage, which makes sense. But it's hard as a company to want to pay for experimental usage. You know, if I had like a data platform I used and I want to do a bunch of experiments to find the right way to express my data, and then once I have that, like, I'll pay for it because, great, it's not really like a supported model.

00:38:06 - 00:38:22 Meir Wasserman
And so I can see the thing you just said I hadn't considered until you said it, but it's like, yeah, if we if it's kind of like a value, you know, it's charged based on value providers like, wow, interesting. Because then I am way more cautious about experimenting if you provided.

00:38:22 - 00:38:46 Stephen Koza
Yeah, maybe there'll be a happy medium or something. But you know, I, I think that's how Palantir prices, if I'm not mistaken, some of the newer AI native startups Sierra, which does customer service stuff, I think they charge you based on support ticket closed or something like that. Because in in SaaS today everybody's got these per seat license models.

00:38:46 - 00:39:10 Stephen Koza
And so the question is, well what happens when you've just got a bunch of agents that are using that platform. So something's got to change. I'm sure smart people will figure it out. So I'm waiting to see well, let's let's move into closing. I got a few rapid fire ones for you if you're ready. Yeah. What's the most counterintuitive lesson that you've learned along the way?

00:39:10 - 00:39:31 Meir Wasserman
I don't know if it's counterintuitive, but I was surprised to learn that as I as I grew up in my professional career, I worked for a lot of different leaders, as have we all. And, you know, I kind of early on, I kind of buckets them. They were like leaders that I really liked and liked the way they operated.

00:39:31 - 00:39:49 Meir Wasserman
And that would I would call them good leaders. And then there were leaders where I didn't like the way they operated. And I called them bad leaders. And and I have matured my mental model since then, which is to say, like there's actually the learning is like there's actually a lot of different valid ways to be a leader, and it's not objectively good and bad.

00:39:49 - 00:40:23 Meir Wasserman
It's more like, does a leaders prefer methodology match my preferences when, yes, I call them good and when? No, I called them bad back in the day, and now I just recognize it's more like it's more about my my preferences than anything else, not about actual leadership effectiveness. And so that was one learning, which I've then turned into like a new learning, which is, well, when someone gives you leadership advice, you're giving you leadership advice based on how they prefer to lead isn't like a universal truth of any sort.

00:40:23 - 00:40:42 Meir Wasserman
Even me. Like right now, like this is like a vice. I'm giving you what I think is right, but doesn't make it like universally correct. And so as I've kind of these like realizations, I was like, okay, this is actually helpful because it helps me recognize that I'm learning. I'm not just learning from leaders that I think are doing a good job.

00:40:42 - 00:40:52 Meir Wasserman
I'm also learning lessons from leaders that I think are not. That do not mean my preferences, but it's just a different type of lesson. It's equally as valuable.

00:40:52 - 00:41:07 Stephen Koza
Okay, last one here before we wrap everything you've seen and done across your career and getting out there. What are you most excited about right now and what are you paying the most attention to over the next 12 months?

00:41:07 - 00:41:29 Meir Wasserman
I mean, the answer is, clearly I had nothing in my entire career has matched the level of which this is changing our environment. I'd be surprised if anyone answer the question differently. To be perfectly frank at this point, or maybe I should just be surprised and it was better answers out there. But for me, it's like, it's not just like, where's I headed?

00:41:29 - 00:41:52 Meir Wasserman
It's also like, how do we incorporate it? A big problem space I think about a lot with AI is like, how do you take what's what is capable now and creating production lines, code out of it or production lines, outputs, proper scale, secure, all that stuff. We're not like crazy far off from that, but there does need to be some structure around it.

00:41:52 - 00:42:17 Meir Wasserman
You can't just people say vibe coding is kind of shorthand for like conversational coding. That's not generally sufficient to create production code. I'm sure some people would disagree and say they did it. That's fine. But I think systematically it's not quite right. And so I think of a lot about like, well, what's the how do we bridge the gap between vibe coding and production, like enterprise level code?

00:42:17 - 00:42:23 Meir Wasserman
I think those answers out there, we're kind of trying some out now, and I'm eager to see where that heads.

00:42:23 - 00:42:49 Stephen Koza
Great answer. Yeah, I as I asked it, I was thinking I wonder if it's going to be not AI. So that's a totally reasonable answer. I am most interested in seeing or figuring out what's possible, whether it's use cases or impacts that we aren't currently thinking about or not everything because it's moved so fast. You gave the date.

00:42:50 - 00:43:10 Stephen Koza
I remember before ChatGPT was a product, some some kind of podcast was talking about the model like GPT three, and I don't know how you get access to it, but I was like, hey, I remember telling people at my company, you know, there's this thing, the right sentences for you, and they're like, what are you talking about? I'm like, no, I listened to it on a podcast.

00:43:10 - 00:43:40 Stephen Koza
It's real. And that wasn't that long ago. And, you know, like a year ago, the model couldn't do math right or string together a coherent paragraph. So we all know this. It's moving super quickly and there's a lot of use cases and things that we do envision and we're thinking about and trying to figure out what we do about it and how do we get the value and security and governance and all this stuff you mentioned I'm really interested in.

00:43:40 - 00:43:57 Stephen Koza
Okay. What do we not know yet and what have we not thought about yet or what? You know, what have we not connected the dots between or envisioned yet? That's the stuff that I just think is really cool. It's, you know, that's that's the reason I pay attention to Twitter and read it and, you know, whatever else. Probably a little too much.

00:43:57 - 00:43:58 Stephen Koza
But it's interesting to.

00:43:58 - 00:44:01 Meir Wasserman
Admit that.

00:44:01 - 00:44:11 Stephen Koza
Yeah. True story. Well, Meir, this has been great, man. It's been really fun. I appreciate you coming on and hanging out for a little bit. Before we wrap up, where can people find you?

00:44:11 - 00:44:26 Meir Wasserman
Oh, LinkedIn, I guess, I don't know, I don't have a social presence, really. I have my blog at Mired Up Blog mere, but I don't market it. I just kind of send it out internally and say, hey, here's something I learned last month and I'd love to share it with everyone.

00:44:27 - 00:44:46 Stephen Koza
Well, thanks again, man. It's super fun and for everyone listening, I hope you enjoyed the episode. Subscribe where you get your podcasts. Give us a follow on LinkedIn or Spotify or wherever you like to connect. I'm Stephen Koza, this is Tech Pod Talks and we'll see you next time.

‍