The AI SRE Hype and How to Get it Right with Yotam Yemini, CEO of Causely

(Intro Narrator at 00:00:00) Today, we're talking to Yotam Yemini, CEO of Causely, about why AI SREs are getting so much hype right now and how you can actually make the benefits tangible. You're listening to Joel Beasley, Modern CTO.

(Joel Beasley at 00:00:19) Let's just start right at the top. We're going to use AI SRE a lot in this conversation. Can you just describe exactly what that stands for?

(Yotam Yemini at 00:00:28) So it's a loaded question because people are making it mean all sorts of things these days and are really reducing the role of an SRE to troubleshooting and being on call. In essence, if you think about an SRE, their job is supposed to be about how do I engineer reliability into my systems. And if you think about AI, the idea that I would apply AI to figure out how to do that makes a ton of sense. And so that's the essence of what an AI SRE is supposed to be. You see a lot of companies now are borrowing this term because it's catchy, and you can sort of immediately imagine what it may mean, to mean one of a couple things that is not really the essence of what the role of SRE is supposed to be about.

(Joel Beasley at 00:01:16) And SRE, what does SRE stand for?

(Yotam Yemini at 00:01:19) So it stands for Site Reliability Engineering. And it really was born inside of Google from an engineer whose name I wish I could remember off the top of my head. I'm not that smart. And he coined Site Reliability Engineering as what happens if you ask an engineer to build an operations function. And so some of the best SREs that you'll meet will talk about their job function and their reason for existence as being to automate themselves out of a job.

(Yotam Yemini at 00:01:49) And so that's a lot of what the ethos of the job function was coined as. How people actually adopt it or adapt it to fit their organization is taking a few different turns, some productive, some not, but that's the history.

(Joel Beasley at 00:02:07) So what is out there currently in the marketplace? Are these just really smart chatbots? Are they LLMs? What is it?

(Yotam Yemini at 00:02:14) So let me start with the good. So the good is some of these AI SREs are using language models for things that actually make sense. Like, for example, one of the jobs of an SRE is to write a postmortem of why an incident occurred. And so the idea that you could feed some data to a language model and have it summarize what has occurred makes a lot of sense. Other examples of AI SREs are things like translating an alert.

(Yotam Yemini at 00:02:46) Right? So there's a recent company that popped up, again, I don't know the name of it off the top of my head, that is building an AI SRE for Salesforce. Because if you think about Salesforce developers, people that build on the Salesforce platform, there's all sorts of innocuous errors and codes of things that if you're a Salesforce developer, you don't know what this alert means. And so having a chatbot that helps you figure out what does this alert mean, again, it makes a whole lot of sense. So those are some of the good applications around AI SREs. I think, depending how deep you want to dive into Reddit and certain conversations online, you'll see that a lot of people who are trying some of these other AI SREs on the market are trying and failing because what people are building is effectively, I'm going to take some observability data, so some telemetry signals, traces, logs, and metrics, I'm going to feed that to a language model, and that language model is just going to do magic.

(Yotam Yemini at 00:03:43) And as you probably know, there's no free lunch with a language model. There is no magic. They're really good at pattern matching. So I actually heard somebody yesterday say, we really shouldn't call them language models. We should call them pattern models because that's really what they're doing, is matching patterns.

(Yotam Yemini at 00:04:01) And that's all great, except for, again, if you get back to the essence of SRE and the concept of applying AI to SRE work, a lot of SRE work is about dealing with emergent behaviors, dealing with systems behaving in all sorts of novel ways that you haven't seen before. And so the idea that a language model is going to be the right solution for that type of problem, it just doesn't track. In fact, there's a lot of research on this point that language models really struggle with counterfactuals. When the information they've been trained on conflicts with a new observation, that's when they start to hallucinate and really don't know what to do with themselves.

(Joel Beasley at 00:04:45) I see that in my parents, you know?

(Yotam Yemini at 00:04:50) Yes. Yes. Yes. It's okay. It's funny you say that. I posted something this morning about how humans hallucinate too. You know, there's all this buzz now about people are railing against AI because it hallucinates. Like, hey. Let's not forget. Humans hallucinate all the time.

(Joel Beasley at 00:05:06) And if you have small children, you understand this as a daily occurrence. They'll just run off, and they literally will talk exactly like an LLM that just goes off the rails. They'll start answering you about who ate the goldfish, and then they're telling you about a snowman they want to build next summer. It's like, what are you doing?

(Yotam Yemini at 00:05:24) Yes. Yes. And so anyway, so that's sort of the landscape of what you're seeing out there. Some good uses for it when you're talking about summarization of the postmortem or translating some alert code, but the idea that I could just feed it a bunch of telemetry and it's going to on its own figure out what's going on assumes a lot in the delivery of that type of a solution.

(Joel Beasley at 00:05:47) And so you're building a tool in this space. Before we get deep into that specific tool, I want to know about your background. Did you work as a psychologist? Were you just studying psychology? How did you get from school all the way to starting Causely?

(Yotam Yemini at 00:06:03) Causely, by the way.

(Joel Beasley at 00:06:04) Causely. I will mess it up again. I promise you. I'm really good at that.

(Yotam Yemini at 00:06:09) It's okay. We have a joke inside the company because we talk about the platform as a causal reasoning engine, and sometimes people call it casual reasoning. And so one of our—

(Joel Beasley at 00:06:19) That's my type of reasoning. I'm a casual reasonist. Yeah.

(Yotam Yemini at 00:06:24) Yes. Exactly. Casual reasoning. So anyway, the background. So yes, I did study psychology in college, probably in a different life. Maybe I would have gone and gotten advanced degrees in psychology. I was a little bit allergic to it, being in a family that already had—my sister is a doctor, my father was a doctor. I was just allergic to the idea of pursuing academia for that many years. I was really fortunate to get a job in coaching college basketball, which obviously has everything in the world to do with technology. Not really.

(Yotam Yemini at 00:06:58) The transition for me from psychology, where what I did research in was industrial organizational psychology. So that's effectively, how do people work in the workplace? So that was actually kind of cool because we built stuff like little helicopter simulators and had people come into the lab and take a little personality assessment. And then we tried to figure out, could we predict based on how they answered the questionnaire at the start, how they would actually perform in this helicopter simulator? And could we actually make predictions about who would be good people to have in a team setting environment.

(Yotam Yemini at 00:07:31) Long story short, I went from doing that undergrad to being very fortunate of getting a job in college basketball. I was an undergraduate assistant at my college, and I'll walk on for a cup of coffee on the team. And when I graduated, I was able to parlay some of the stuff that I had done to help the team out as a sort of practice scrub and office guy into a job. And long story short on that is that was right around the time that sports, coaching college basketball specifically, was going from analog to digital. Up until that time, the way that you would come up with your plan for how you were going to deal with an opponent was you would literally sit there, watch the game tape, and mark down on a piece of paper, okay, this is what I saw happen, this is how many times I saw it happen.

(Yotam Yemini at 00:08:23) And the idea of encoding that in a computer was not on the minds of any coaches, because I was a young guy at the time and I was coming out of having done psychology research and I knew how to build databases and I knew how to do things analytically on a computer. I kind of got slotted into that and it got me into coaching college basketball. And I think in another world, maybe I would have continued on that path. But that was my sort of entry into analytics, if you will. And then when I decided to hang up the whistle and the coach's clipboard and all that, and figure out what to do with the rest of my life, I transitioned into a tech company that happened to be applying analytics into infrastructure and cloud computing.

(Yotam Yemini at 00:09:08) This was 2010, so the big wave at that time was virtualization and people moving their servers from physical to virtual. Containers and Kubernetes really weren't a thing yet, and that was really sort of my entry into analytics and specifically into enterprise B2B tech.

(Joel Beasley at 00:09:27) Oh, that's so cool, man. I played basketball for one season. I played football for one season. I played baseball for ten years, and I, in my genetic encoding, I did not get the gene to watch sports. I have watched zero sports ever, but I loved when I played them. And I played them all the way through my grade school, all the way up through high school. I really enjoyed playing the sports. I just can't sit down and watch.

(Yotam Yemini at 00:09:53) Yeah. Yeah. It's, it is something about—I heard somebody say this the other day, and it stuck with me that there's three things that are absolutes if you want to be successful, at least in this country, in this culture. You have to be good at writing, you have to be good at reading, and you have to have played a sport of some sort because it helps you understand how to persevere, how to work with others. And if you can do those three things, learn how to write, learn how to read, learn how to work in a team environment, team sports being the closest proxy for that, you're setting yourself up for a successful life.

(Joel Beasley at 00:10:28) I would 100% agree. You know, when I looked back and I asked the—a fund invested in me early on in my career, and I looked back and I asked them, I say, hey, why did you invest in me? And they're like, well, as we got to know you, we realized that you were very coachable. And I said, oh, okay. Cool. It's like, well, what else are you going to do? If people give you better information, you have to apply the better information to move forward. You know?

(Yotam Yemini at 00:10:50) Yeah. I mean, look. Again, I feel like I could rail on this forever, so I'll try not to get on my soapbox for too long. But even nowadays, information used to be everything, but now anyone can have access to information. So now it's all about aptitude. Right? It's like, how fast can you learn and how hungry are you to grow? And again, back to coachability. That's so much of the foundation of what you need to succeed.

(Joel Beasley at 00:11:15) Let's talk about Causely and your background. Why did you start Causely? Why did you look in the marketplace? You're like, hey. There's a gap. It's missing. I need to fill this. What was the drive for that?

(Yotam Yemini at 00:11:26) Yeah. So the team here, one of my favorite things, we talked about team before, this team is aces. And the founder of Causely was also the founder of the first startup that I ended up working at. Right? So I transitioned from coaching college basketball into this tech startup I told you about that was in this sort of virtualized and cloud infrastructure world, in around 2010.

(Yotam Yemini at 00:11:49) And for Shmuel, who's the founder of the company, it's really a continuation of something he's been obsessed with ever since he was in the army as a systems programmer working on mainframes in, I don't know, the seventies or eighties, which was when something would break, people would call him, and it would annoy the hell out of him, and he would say, why do we have humans in the loop of troubleshooting stuff? Why can't we figure out how to get systems to manage themselves? And I think that's something a lot of engineers can relate to. And each of the companies that Shmuel has been a part of starting, Causely being the third, has had a sort of a chapter in the book of the story of self-healing autonomous systems. He would be the first to tell you, and I would be the second, that's a journey. No one company can solve it all, but Causely is sort of the next chapter in a story that started with a company called Smarts in the nineties. Smarts was the leading provider for network management software. Every large telco used them. This was back when the problem in the nineties was packets getting dropped on the network and these telco providers having to figure out why and fix it fast.

(Yotam Yemini at 00:13:05) And so Smarts was a very successful company in the nineties, acquired by EMC in 2005. And then he spent a couple years inside of EMC as a CTO in their advanced research office and started Turbonomic, which at the time was called VMTurbo, and it was something for VMware environments for managing them more efficiently and more performantly. Flash forward, Turbonomic was acquired by IBM for $2 billion in 2021, and it's now part of their AIOps portfolio. And then about a year later, after Turbonomic was acquired by IBM, he started Causely with a few of the former early engineers from Turbonomic. Actually, also one of the engineers from Smarts, and also a new engineer that was added to the team as well as one of the cofounders. And so that's how it all sort of came together, sort of this next chapter in the journey of trying to build towards a future where systems can manage themselves.

(Joel Beasley at 00:13:59) That is so cool. And so you have a dream all-star team to coach now.

(Yotam Yemini at 00:14:03) It's amazing. Yes. So almost every single person that was in the company and invested in the company when I joined, in terms of the board, the founders, the employees, were people that I previously knew and trusted. I think that makes such a difference when you're trying to start something new and build something special.

(Joel Beasley at 00:14:22) Are you using that in your recruiting strategy going forward as you grow? Like, are you teaching these people how to recruit their friends and other people they trust?

(Yotam Yemini at 00:14:30) That itself is an interesting topic, especially because you're pulling out my heartstrings in psychology and industrial organizational psychology and how do you build teams. So I've tried to be split-brained about that. One is that, certainly when you're building a team that you're going to ask to go do heroic things and special and exceptional things, finding people you've worked with before where there's that inherent trust has immeasurable value. And that's especially true if you're a remote company. So for us, just the reality of the team is we're distributed.

(Yotam Yemini at 00:15:05) We've got one of our cofounders in Germany. One is in New Jersey, and one is in—actually, two are in New Jersey, and then another one's in New York City, and then I'm in the DC area. So we're just, by definition, remote. And so having that trust just matters so much. I said I was split-brained about it because on the other hand, you don't want to end up in an echo chamber, and you don't want to end up in the whole, well, this is how we did it before, so that's how we'll do it now type of thinking.

(Yotam Yemini at 00:15:33) And so I have been somewhat intentional in trying to make sure that we also intentionally try to bring in people with different perspectives. And I think those are some of the ingredients you need if you want to create a team that's stronger than the sum of its parts. You've got to bring people with diverse perspectives. I always say diversity isn't just about your gender or your race. It's also your experience and what are you bringing to the table. And so that's how I sort of think about team building.

(Joel Beasley at 00:16:06) Yeah. Well, 100%. I mean, diversity of thought is what everybody really wants, you know, because that's what gets you the benefits and the results. Now one of the strategies that I've used that was pretty good that's worked out well for me at least, was finding and encouraging people that are on my team to find people they admire. Like, who do they learn from?

(Joel Beasley at 00:16:28) Who are they reading? Whose conference talks are they going to? And then trying to pull those people in because

(Yotam Yemini at 00:16:34) I love that.

(Joel Beasley at 00:16:34) They've already got the respect. They've already got the market validation of expertise. They've got a view. You can find people that have opposing views of them fairly easily, and you can pull those people in.

(Yotam Yemini at 00:16:45) Yeah. I love that. I think they always say that, right?

(Yotam Yemini at 00:16:50) Like, study the person whose job you want. Yeah. And those are sometimes the best places to go.

(Joel Beasley at 00:16:57) So, Causely, what's the problems that you solve? For, like, so people that are listening right now, how do you make their life better? Why would they buy Causely?

(Yotam Yemini at 00:17:07) Yeah. So I'll talk about what we're solving today, and then I'll just talk about where we're going. So if you think about a developer's life in a medium to large enterprise, you could really split it up into three things. Feature work, reactive maintenance, troubleshooting stuff, and then proactive maintenance, trying to prevent stuff from breaking or making things reliable. The reality is that usually you don't get enough new stuff delivered because you end up getting too distracted and bogged down by the reactive work. And while you know the proactive work would help reduce the amount of reactive work you have to do, you almost never get to the proactive stuff because it almost feels like a pie in the sky thing.

(Yotam Yemini at 00:17:50) Right? It's like, for some people, it's like eating their vegetables or getting to the gym. It's like, I know I should do it, yet I don't do it as much as I should do it. And so today, the problem that we're solving is more in automating the reactive part of the job, so to speak. And then where we're going and we're starting to accumulate proof points in this with our customers is to more of the proactive work.

(Yotam Yemini at 00:18:13) Again, if you get back to the essence of an AI SRE or what an AI SRE should be in my vision, it's something not that's just like an on call engineer that sits there in Slack with you to respond when something goes wrong, but really a companion that's proactively helping you improve reliability and helping you figure out, hey. How do I actually keep things in a healthy state in the first place? And so we could talk more about what's required to get there, but that's a bit about what we do today. And it's a bit of a developer productivity gain is a bit about how to think about it, but it's also a bit of an operational gain too because, again, companies are gonna be different in terms of where they place the operational burden of keeping systems reliable, and they're gonna be a little bit different about how do they respond to incidents when they do occur. And so those are sort of the categories of places where we're delivering value today.

(Joel Beasley at 00:19:08) Do you have any case studies or anybody that's using it today that's had a meaningful impact within their organization? Anything you can share?

(Yotam Yemini at 00:19:18) Yeah. So the best companies that we're working with so far in terms of what they're achieving with it, they're set up in a little bit like, I hope this doesn't hit the wrong way, the Spotify model where you sort of have like squads and tribes and really nobody owns the full picture. Like, sure, maybe somebody owns the P and L and maybe you have a CTO or a head of engineering, but no one really at the level of people who are doing the work, no one owns the full picture. I have different service owners and different teams that are responsible for different services. And so the outcome that we're delivering for those types of companies is that without us, let's say you have something like a database problem.

(Yotam Yemini at 00:19:57) So if somebody is a service owner of a database, that database is slow. I'm just oversimplifying this. That database being slow inside of one particular service team's service may actually manifest as a set of different anomalies that are observed by other service owners. Those service owners without Causely might end up wasting a lot of time chasing or triaging what you might categorize as spurious alerts. They're not wrong alerts, like, yes, something's wrong.

(Yotam Yemini at 00:20:32) Yes, your service has more latency or yes, your service has an increased error rate, but it's actually not because of you. It's because of this other service. And so the benefit that Causely brings is being able to infer in real time and continuously, and before you even have to write a single query, it will tell you, hey, I've inferred that the cause of all those anomalies is this service. It's with this one service. And being able to be that sort of arbiter of truth, that independent sort of third party arbiter of truth in Causely, drives a massive benefit because now you're avoiding war rooms.

(Yotam Yemini at 00:21:10) You're avoiding four different teams troubleshooting stuff that all just sits with one specific service owner. When you think about people trying to reduce their MTTR or their Mean Time to Resolve an issue, most engineers will tell you that once they actually know where the issue is, fixing it's not like, that's not the hard part. The hard part is like knowing definitively like where is the actual issue, what's the cause of the issue. And that's the type of positive business outcome that our customers are getting is being able to resolve those issues faster, and then, like I said, also prevent them in the first place by applying our analytics for more sort of like shift left type of use cases.

(Joel Beasley at 00:21:49) So when people are onboarding, like, when they're becoming customers or evaluating using this tool, what's the top question that they have? What's their main concern?

(Yotam Yemini at 00:22:00) I think their biggest thing is the skepticism of, am I really gonna see value quickly? And look, I get it. Right? Like, people reach out to me all the time trying to sell me something these days. You pick a category, there's like 50 new startups tomorrow that claim that they have some AI approach to that thing.

(Yotam Yemini at 00:22:18) Look, the fact is, we are talking about solving a complex problem, and we are talking about solving it at scale. And so in order for me to demonstrate the value of that, you have to be in enough pain or you need to want enough the positive business outcome of what I'm talking about to take that first step, to say, okay, I'm actually gonna try this out. And then when you do try it out, you're pleasantly surprised that it's actually as advertised. But, you know, that's the biggest hurdle somebody has to take.

(Joel Beasley at 00:22:51) So somebody that has this big pain point, how does that express itself? Like, if you were to talk to anybody that's listening to the show, you're saying, like, if you have this massive pain point that is a recurring problem, you need to contact us now. We can help with this. What is that thing?

(Yotam Yemini at 00:23:07) It's one of a couple things, and I have to give you two answers. So one is, we're not getting enough done from a feature velocity standpoint because we're taking too much time troubleshooting. So our teams are just spending way too much time troubleshooting, we're not shipping fast enough, we're not shipping enough of consequence to our business because of the toil and operational burden of keeping things reliable. So that's one. The second, there is a lot of tops down board pressure in a lot of these companies to prove back to the executive team tangible use cases of deriving value using AI.

(Yotam Yemini at 00:23:46) You like it or not, that is reality. And in this category of developer productivity and managing the performance of your systems and your applications, like, in this category, we are that. We are that really easy, fast way to demonstrate value in applying AI to solve a real business problem, which is the productivity of your engineering organization. So it's one of those two drivers, really the two.

(Joel Beasley at 00:24:20) Let's go a little high level on this. I'm curious to try to wrap my mind around it. So is this new in the sense of the board, the higher level people kind of mandating a specific technology? Because I've seen them do it before with methodologies of working. Right?

(Joel Beasley at 00:24:41) What is, you know, like, they want them to be agile. There you go. Agile is a big I've seen them push. They've read the book. They've seen the thing, and then they start pushing it.

(Joel Beasley at 00:24:50) And that's kinda how you're working. But I haven't seen them. I guess, this is one of my first time seeing them push like a specific technology. Like, we wanna see AI results. Is this a new thing?

(Joel Beasley at 00:25:03) Or am I just newer in my career and I haven't seen it?

(Yotam Yemini at 00:25:07) I think it's, um, I think it's on par with the noise around cloud back in whatever, o five, o seven, ten, like around then. Right? I mean, I think boards and people, you know, publicly traded companies were always talking about the investments they were making and shifting to the cloud. And maybe, like, agility, your comment about agility might have also been tied, like, it might be correlated in time. We'll have to go back and look on when these things spiked.

(Yotam Yemini at 00:25:35) But, like, board pressure around agility probably is correlated to like board pressure around the movement to cloud. Um, and, look, I think a lot of people will say that, like, AI is sort of like the next wave after cloud in terms of something that is getting a ton of top down pressure because I think people see it as the massive however many trillion dollar opportunity that you wanna see it as at a high level. And the productivity gains that certain businesses are driving are really attractive to executives and to boards and to investors. The nuance there is that, and actually the first person I ever heard say this was John Chambers. You know, John Chambers used to be the CEO of Cisco through like all of its formative growth years.

(Yotam Yemini at 00:26:26) And he once said this, and it always stuck with me that, you know, digital transformation is not a technology problem. Like it's a little bit about technology, but it's also about are you able to conveniently use this technology wave as a way to catalyze change in your organization itself. In terms of change of your processes, change in how your people think, change in how they operate, change in what they do every day. You know, that's where it becomes a big opportunity. And I think you did see that by the way with cloud.

(Yotam Yemini at 00:26:59) Right? Like, if you think about before cloud, you had very specific people, I'll just use an example in IT, like, you had your network people and your storage people and your server people and your, you know, Java admins, and your whatever. And, some of those people still exist, you'll tell me, even in the world of cloud, but I think you saw some of that collapse and a lot of those people went away and they were reborn as either cloud architects or cloud engineers or platform engineers or things of that nature. And so I think you did kinda see that shift happen with cloud, and I think you're gonna see that shift, you're already seeing that shift happen with AI, in every function of a business, not just in engineering.

(Joel Beasley at 00:27:47) That is interesting. I hadn't heard that before the way John Chambers laid that out. Utilizing the technology as a way, like, the wave of the new technology as a way to change behavior in your organization. That's fascinating.

(Yotam Yemini at 00:28:02) The article, I think he, I think it was published in somewhere. I gotta try to find them online. Maybe we could share it with the viewers after the show.

(Joel Beasley at 00:28:09) Maybe if they're lucky. We'll see how we feel about it after

(Yotam Yemini at 00:28:12) this. I don't know if anybody reads anymore, to be honest. I know. I'm kinda sick of people taking stuff I share with them, and then they just type in it, they put it in the ChatGPT, and then they ask ChatGPT to summarize it. And then they send me back a reply that I can tell was just written by ChatGPT.

(Yotam Yemini at 00:28:26) And I'm like, did you even take time to think about this thing?

(Joel Beasley at 00:28:30) You know what I get a lot? I get a, hey. What do you do for work? I'm a podcaster. Oh, there's this podcast episode I listened to last week about this other.

(Joel Beasley at 00:28:38) You've gotta listen to it. You have gotta listen to it. And I'm like, okay. Send it to me. I don't think I have ever, the, like, the number of times, even in my personal life, when there's like movie recommendations, my hit ratio is so low on people recommending me stuff and me actually consuming it.

(Yotam Yemini at 00:28:56) Oh, is that? Yeah.

(Joel Beasley at 00:28:57) Yeah. What about you?

(Yotam Yemini at 00:28:58) Well, I was about to, sorry. I was about to go on a tangent. The, um, I was gonna say that, you're probably saving your time. So have you heard of Sturgeon's Law?

(Joel Beasley at 00:29:06) No. No. No.

(Yotam Yemini at 00:29:07) Okay. So you may like this. 90% of everything is crap. And so the reality is if people recommend you 10 things, like, you could probably throw out nine of them, like, none of them are probably crap. People tell me 10 things to go listen to, none of them, not that great.

(Yotam Yemini at 00:29:26) But look, you asked me, do I actually listen or read the things? It's like, I do because like when you get to that one out of 10 that's actually worth listening to or actually worth taking something from, you're like, okay, cool. I gained something from that. Like, that's, you know, that was worth it. And so anyway.

(Joel Beasley at 00:29:42) No. Alright. So, uh, I mean, I'm not filtering the crap. Alright.

(Yotam Yemini at 00:29:46) I am a two x er. Look, I will say I'm a two x er on podcast. I do not listen to a single podcast on one x. I'm a two x by default. That's actually, that's an interesting interview question you should ask people.

(Yotam Yemini at 00:29:59) Like, are you a one x er or a 1.25, a two, a 0.75? I think it tells you a lot about a person.

(Joel Beasley at 00:30:05) You know what? For me, I'll tell you right now. I don't listen to a lot of podcasts.

(Yotam Yemini at 00:30:12) I'm a zero. I'm a 0.75.

(Joel Beasley at 00:30:14) It's like if you work at a pizza restaurant, like, oh, we like to go for pizza. It's like, I've done a lot of pizza today. But I do listen to Audible quite a bit. Okay?

(Yotam Yemini at 00:30:26) Oh, okay. Yeah.

(Joel Beasley at 00:30:27) And my multiplication on Audible is directly related to the cadence in which they speak. So some people are naturally speaking faster. I'm fine with it. Other people, I'm doing the 1.25 to two x. You know, I don't want them chipmunk.

(Yotam Yemini at 00:30:41) I don't

(Joel Beasley at 00:30:42) want them to go chipmunk. But, like, you know, just a little bit faster is good.

(Yotam Yemini at 00:30:46) Yes. Yeah. Yeah.

(Joel Beasley at 00:30:47) Alright. So you're two x podcast guy, Sturgeon's Law. We're learning a lot. I feel like this is a little one-sided. Like, I'm learning a ton here today.

(Yotam Yemini at 00:30:56) Well, this is good. I, uh, you know, I'm a wealth of knowledge. So kidding.

(Joel Beasley at 00:31:01) No, dude. You really are. And I interview a lot of people. You were great. You know a lot.

(Joel Beasley at 00:31:06) You know a lot.

(Yotam Yemini at 00:31:07) Well, you gotta be a student of the game, man.

(Joel Beasley at 00:31:10) Do you coach any of your kids sports now?

(Yotam Yemini at 00:31:12) I coach my third grader's basketball team, and then my first grader now has told me that he wants me to coach his team. So now I'll have some of my first year of two teams that I'll be coaching, and I'm sure my daughter is somewhere behind them in a couple years.

(Joel Beasley at 00:31:26) That is exciting.

(Yotam Yemini at 00:31:28) Yeah. I'm expanding from basketball to football this year. I was sick of my son being on bad flag football teams where the coaches didn't know what they were doing. So I got with one of his buddies' dads, and we're gonna coach a flag football team we're putting together. So now I'm becoming a multi sport coach.

(Joel Beasley at 00:31:45) Multi sport coach. That's the progression. Do you know the position he's playing? Or

(Jotam Yamini at 00:31:51) Well, with flag, there are no positions. It's sort of like position-less, and they move all over the place. And, okay, again, I just—I try not to be that sideline parent, the helicopter parent in sports because I was actually involved with sports at a pretty competitive level. I never take that stuff too seriously. I just cannot stand when coaches don't get all the kids engaged. That's just what drives me nuts. Like, get the kids engaged. And that's why I said, enough with it. I'm gonna pick up how to coach flag football this year. We'll see how I do.

(Joel Beasley at 00:32:26) You're not gonna mess them up, really.

(Jotam Yamini at 00:32:28) No, no, no. It's all about having fun. That's it. Like, have fun. That's the whole point. Like, hey, why didn't little Tommy over there get a single pass thrown to him all game? Like, this ain't the NFL. Pass the ball to different people. Get people involved.

(Joel Beasley at 00:32:42) Yeah. My neighbors have kids that are my kids' ages, and they're like really into sports. And I just—I don't think I could get that into it.

(Jotam Yamini at 00:32:52) What about—are your kids doing sports at all or no?

(Joel Beasley at 00:32:55) So my son asked to play baseball. He played T-ball last year, and then we moved, and then we actually missed the sign-up by like a week this summer to play for this season. So he wants to play T-ball or baseball. And my daughter does gymnastics, and she's progressing. She hasn't gotten into competitions yet. She's eight, but she goes every week, and she makes a lot of progress. And then she used to do dance, which had recitals, but she preferred gymnastics.

(Jotam Yamini at 00:33:30) I know I'm supposed to be the one getting interviewed here, but I gotta ask you a question. Have any of those kids tried to quit something? And if so, how did you deal with it?

(Joel Beasley at 00:33:38) Oh, that's actually a great question. Yes. So we have a Beasley family rule. And the family rule is that you make the decision to do something and you have to complete the season whether you like it or not. And sometimes that expresses itself different ways. Sometimes it's, I just don't wanna go, and we make them go. And we've had moments where they did not engage or play, but they had to sit there and watch their team do it. Because they're just like, I'm not doing it today. I'm not doing it today. Specifically, my daughter with dance—she didn't wanna do dance that day. I'm like, that's fine. You don't have to, but we're driving there. You're sitting there and you're watching your team do the dance stuff.

(Jotam Yamini at 00:34:14) I like that.

(Joel Beasley at 00:34:15) You can quit and stop outside of the season. And then if it's things like music, we just do a time-based commitment because there's not a natural season to it. So for music, my son wanted a drum set, and I said, okay, you have to do drum lessons for a year. You have to commit to weekly drum lessons for a year, and he did. And then at the end of that year, you could choose if you wanna continue or not continue.

(Jotam Yamini at 00:34:37) See, now this is a reciprocal interview because I'm learning things too. This is great. I like those.

(Joel Beasley at 00:34:42) What are you doing?

(Jotam Yamini at 00:34:43) Well, I'm not doing a great job of it. That's why I was asking. I think the part that I took away from what you said is just the expectation setting. You know? Like, the very clear expectation setting at the start of what we're signing up for. And it's hard with kids because, you know, they don't know what they want yet, and they want the path of least resistance on certain things. But it's important to get them to understand the value of sticking with something and of persevering through hard things and of not quitting. It's just, you know, kids give you so many opportunities like that to coach them up like that. And, you know, I think—I haven't met a perfect parent. I'm not one either. And so I'm always looking to pick something up from somebody in a particular part of, hey, how do I be a better dad?

(Joel Beasley at 00:35:34) Yeah. Well, when it first came up for us, you know, they didn't wanna go, and then we were faced with this idea. And I could tell when I was telling them, like, you can't quit, and I could tell that their brains just weren't understanding it in the way I was saying it. So I started saying, you can quit. Here are the parameters in which you can quit.

(Jotam Yamini at 00:35:54) I like that.

(Joel Beasley at 00:35:55) I'm like, yeah, you can quit. You just can't quit until next season. Like, you just have to show up. I don't wanna dance. You gotta show up. You gotta watch your dance.

(Jotam Yamini at 00:36:02) I like that. That's like giving my daughter the option of you want a bath or a shower. It's not, you know, it's like, you do it this way or this way. But like, here are the guardrails. I like the guardrails.

(Joel Beasley at 00:36:11) Yes. And that is exactly all the good parenting advice I have. Thank you for tuning in, folks.

(Jotam Yamini at 00:36:19) That is Joel's good parenting advice at the end of the day. I love it. I love it.

(Joel Beasley at 00:36:24) All right. As we start to wrap up, let's tell people to go buy things from you. Where do they go? What do they do? We already talked a lot about the problem that they may be experiencing with getting to, you know, the new feature work or being bogged down in the research type of work of debugging. Tell me where people can go to reach out to you to learn more about SRE stuff.

(Jotam Yamini at 00:36:46) So two things. First, the website's causely.ai. So causely.ai. Second thing is there's nothing that gives me more joy in my day than talking to other humans about whether it's about Causely or whether it's about just their problems in the realm of applying AI to SRE work or just, like, you know, I geek out on this thing all day long every day. And so shoot me a message. I'm on LinkedIn. I'm fairly active on there. I'd love to talk to anybody who listens to this who just wants to talk. No sales pressure. No hard sell from our team. Promise you that. I promise you you will learn something. I promise you you'll smile and have fun, and you'll meet good people by getting to know us. And so, yeah, just go to the website or shoot me a note.

(Joel Beasley at 00:37:36) Excellent. So LinkedIn, the website, we'll make sure both links are in the description for the podcast, and that's it. Jotam, we made a podcast. How do you feel?

(Jotam Yamini at 00:37:46) I feel great.

(Joel Beasley at 00:37:47) Thank you so much for listening. And if you found this episode useful, please share it with a friend or colleague who you think would get value from it. And if you have topics that you'd like to hear discussed on the podcast, either add me on LinkedIn or send me an email [email protected]. Every time I get an email or LinkedIn message, it absolutely makes my day and inspires me to keep going.