Raw Data By P3 Adaptive
Tinkering is a Way of Life w/Darinee LouvauListen Now:
Today’s guest, Darinee Louvau, is twice retired from Microsoft and one of the few people who admit to an admiration of testing in production. Darinee is an amazingly gifted person. From hosting LAN parties in her Tween and Teen years to designing and building gaming crates for those parties, she is not afraid of getting her hands dirty. A self-proclaimed fan of lowering the cost of iterations, she moved up quickly through the ranks to become program manager for Microsoft on projects such as Answers, Bing, and even the streaming service, Mixer. She is a hands-on thinker, doer, and creator. We can guarantee you will never find anyone else with her can-do attitude.
Darinee shares how she personally prepares for inevitable mistakes in one of her hobbies by sewing with thrifted sheets to reduce costs. Using thrifted materials empowers her to make mistakes and continue to iterate. By using cheap materials to power each testing iteration she can lower the overall cost of iteration to get straight to the thing that matters, testing a solution against the actual live circumstance of what you really need to know. (Psst! Darinee . . . your inner developer is showing. 😊)
Also in today’s show, we learn the answer to the age-old question of “What happens to the code when Microsoft kills it” as well as the secret behind how search engines return answers. The one thing you won’t find in today’s episode is a binary answer about success. Darinee explains that it doesn’t exist and if you think it does, you are setting yourself up for disappointment. Sage advice for our listeners.
If you enjoy this episode, be sure to leave a review on your favorite platform as reviews help new listeners find us. And be sure you subscribe to receive new episodes in your inbox every week. You never know what you will hear when data gets raw!
Also in this episode:
Mixer – Microsoft streaming shuts down
Home School Valedictorian by Adelitas Way
Microsoft Bob – Killed by Microsoft
Darinee on P3 Adaptive Blog – A 48 hour case study from Bing – Zero to Sixty 🙂
Lost wax method in 3d printing
Rob Collie (00:00:00): Hello friends. Today's guest is Darinee Louvau. Over the years, we've definitely talked a lot on the show about imposter syndrome, but let me focus your attention for a moment on something related that I like to call friend envy. You know those friends that are just so high performing that if they weren't your friends, they'd just make you sick and you're glad that they're your friends? You can't help but comparing yourself to them at least small twinges of inadequacy in yourself. Well, I have been either blessed or cursed, mostly blessed with having lots of friends like that. I have many friends that I have to be careful to not compare myself against. We've all got our own talents, we've all got our own paths through life and measuring yourself against others is just really unhealthy in general, isn't it? Well, Darinee is a friend in a class, all her own.
Rob Collie (00:00:52): She is ridiculously high performing, graduated from college early. She's considerably younger than me, but she's already retired twice. She's been good at her jobs. She's been good at her hobbies. She has an amazing number of hobbies, but I think the biggest compliment I can pay her is that I've never once felt even the slightest bit bad about myself by comparison and her presence. And it occurs to me just now thinking about it, that hanging out with Darinee you never once get the sense that she's playing the comparison game. Why would we do that? Let's all just go be awesome and positive and happy together and what else would even make sense? She worked for a long time on Bing, which market share aside is a tremendous piece of technologists, an amazing platform. She's worked on Azure streaming services. She also got to work in the coveted Microsoft games division.
Rob Collie (00:01:47): She even wrote a guest post on our blog about power pivot many years ago, but it's the things she does in quote unquote, real life that I think are the most compelling and that same attitude that she brings to her hobbies. She brings it to work when she worked, show off. She's just one of the world's most infectiously positive and creative people while at the same time also being probably a locked in number one, overall draft pick of people you'd want on your island in case things went Mad Max. That's a lot of range folks. I'm really excited to be introducing her to all y'all out there. So let's get into it.
Announcer (00:02:24): Ladies and gentlemen, may I have your attention please?
Announcer (00:02:29): This is the Raw Data by P3 Adaptive Podcast with your host Rob Collie and your co-host Thomas LaRock. Find out what the experts at P3 Adaptive can do for your business. Just go to P3adaptive.com. Raw Data by P3 Adaptive is data with the human element.
Rob Collie (00:02:53): Welcome to the show. Darinee Louvau, how are you today?
Darinee Louvau (00:02:57): I'm good, thank you. I'm nervous.
Rob Collie (00:02:59): Well, we'll take care of that, don't worry. I'm really glad that we're going to do this because you are, take it from me, you're a fascinating person and it's my job to know. So you and I met at Microsoft, what is now a very long time ago.
Darinee Louvau (00:03:15): I think it was 20 years when I started and I met Jocelyn for the first time. She was my new hire mentor.
Rob Collie (00:03:19): Isn't that cool.
Darinee Louvau (00:03:19): It's crazy.
Rob Collie (00:03:21): You've had a lot of interesting jobs. You also have a lot of interesting hobbies. I definitely want to talk about both, but let's start with the software engineering. At what point in your life did you show like an inkling of interest in computers? How did you even come to even just deciding to interview at a place like Microsoft? There's a lot that leads up to that.
Darinee Louvau (00:03:40): Right. I started off playing with computers pretty young. I have an older brother who I did everything with. My parents were both in technology spaces. My dad worked at Intel for a long time. And so he always gave us opportunities to just learn a lot more about technology, computers. My brother and I were pretty young when we started building own PCs. And so I learned a lot-
Rob Collie (00:04:00): Oh really?
Darinee Louvau (00:04:01): From my brother and my dad primarily about that, doing a lot of gaming, playing Command and Conquer and then starting to program it like games are so awesome. I want to be a game programmer when I grew up. I want to work for Sierra online. That was the coolest company back then.
Rob Collie (00:04:15): Yeah. That's the one. Yeah.
Darinee Louvau (00:04:15): And I think that probably initially triggered my interest in programming, computers. I didn't do anything really often when I was young just to be clear, I wasn't like a support but I really enjoyed using computers and playing with them and then networking them. And one of my friends and I, we were reminiscing recently about some of our really old LAN parties we used to do on DLC cables with 10Base2 networks and setting static IPs for everybody who's there. My brother and I had a table on a whiteboard where we would write down, "Okay, John's going to have this IP-
Rob Collie (00:04:47): Oh my God.
Darinee Louvau (00:04:49): And Darren's going to have this IP and make sure all your subject masters are correct. And..."
Rob Collie (00:04:54): Wow.
Darinee Louvau (00:04:54): So yeah, gaming, I think was probably the main entry point into getting into computers and programming and software.
Rob Collie (00:05:01): Do you have any pictures of those old whiteboard diagrams of those... From those LAN parties? If you did, those are probably some really cherished artifacts.
Darinee Louvau (00:05:10): I have no idea.
Rob Collie (00:05:11): On the other side of the great film barrier, it comes along in the geologic record sometime around like what, 1999 or so-
Darinee Louvau (00:05:17): Exactly. I was thinking my... I think my digital picture history goes back to 1998 and that's pretty much it.
Rob Collie (00:05:23): What happened before then? Nothing.
Darinee Louvau (00:05:25): Right. Exactly. That's when life started.
Rob Collie (00:05:28): Luke and I were joking yesterday about how... I was born in 1974 and that's so much closer to World War II than it is to today. The stuff that was just even like 15, 20 years before we were born, the whole world was still in black and white. They didn't even have color yet. Those poor people had to walk around in a black and white world, when they finally got colored, it was that washed out color of the 50s of the film that got exposed to too much oxygen or sun or something. It was like, you just grow up thinking that those eras didn't have color, like even in their world. There's these epochs that are demarcated by the technology we had to record it.
Darinee Louvau (00:06:04): Yes.
Rob Collie (00:06:04): Now you got to go looking, right? You got to go looking through the film archives for the LAN party pictures.
Darinee Louvau (00:06:10): Yes.
Rob Collie (00:06:10): How old were you, roughly when you were assembling your own computers and diagramming network topologies for the LAN party? What age are we talking here?
Darinee Louvau (00:06:19): Looking back, I think I was probably like early teens, like tweens.
Rob Collie (00:06:25): All right.
Darinee Louvau (00:06:25): I think what was nice is just to be able to have been afforded those opportunities from my parents. They were always supported... my dad treated my brother and me as just the same. My brother is 15 months older than me. So pretty close.
Rob Collie (00:06:37): Yeah. That is cool. I remember when we were doing the earliest original Xbox LAN parties at Microsoft, we were working there and we would stick around on a Friday afternoon and come into one of the big conference rooms. Inevitably there'd be some struggle with the network, something wasn't working right, chaining these Xboxes together. And I developed a rule eventually, which was when we have those problems, find the youngest person in the room and put them in charge of fixing it. And every single time that rule worked. So-
Darinee Louvau (00:07:09): That's a good strategy.
Rob Collie (00:07:10): You weren't at Microsoft. What year did you get to Microsoft?
Darinee Louvau (00:07:14): 2002.
Rob Collie (00:07:14): 2002, okay. So wasn't Xbox original, like a 2001 thing. It was right in that era.
Darinee Louvau (00:07:20): Yes.
Rob Collie (00:07:20): You could have come helped us get our private LAN working in building 17 if only we'd known you were there.
Darinee Louvau (00:07:27): Speaking of Xbox as the gaming, my brother and I and my husband, we made a couple of portable crates that had a screen, speakers, Xbox 360 and its power supply hooked into this custom made MDF crate that had metal brackets on the corner and the piano hinge and an IC connector on the side. And so we tote this big, old giant crate to LAN parties. My brother made one and then my husband and I made a separate one that was a little bit more portable and compact, but we would both took these gigantic heavy... Really heavy crates to LAN parties. And then everybody would be struggling with their pieces or the Xboxes taking him up all their cables. And I remember that very clearly from doing LAN parties back in the day. But so to be able to just roll up and dig all crate, open it up, stick in the cable and then be done. I mean, that was super fun.
Rob Collie (00:08:19): This is like a docking station set up type of thing, is that-
Darinee Louvau (00:08:21): No, it was... The crate itself had the screen in it on the top-
Rob Collie (00:08:26): Oh my God.
Darinee Louvau (00:08:27): It had an Xbox 360 on its side. It even had like a cutout in the wall so that the disc tray would open up and then it would have the power supply on one side with an IC quad on the side that you could plug in and then it had enough storage. You could put like two controllers in there plus a disc pack of discs. And so all you had to do was just open up the crate and then plug in network and plug in power and then you're just good to go.
Rob Collie (00:08:48): This could have been one of many side businesses that you had to have considered at one point. This is something about you that I know for sure. Which is that you and your husband and your whole family have zero fear trying and building new things, no intimidation at all about the tech and I'm not like that. I'm a nerd for sure, but I'm not really that into... You start talking about soldering programming, like a raspberry pi or something like that, my credentials fall off real fast. To you, that just seems like the good stuff, that's the fun. From a distance, I love watching that kind of energy and I love these stories about the things that you've cooked up. There are some of the best things. I'm insanely jealous of the personality that you possess to go and do those things. I really wish that were me, I know that it's not.
Darinee Louvau (00:09:38): I think there's probably I'd say two or three things that play into that. One of it is a little bit of my upbringing. My mom is super creative. She's a builder, she's a crafter, she did a lot of woodworking. And so to see my mom do this sort of thing was very influential to me to be brave enough to go and just do it, go cut something on the table. So go make this thing. We did a lot of fun stuff that we built. And then the other thing actually I think, came from work. And so I started working in 2002. I worked in Amazon Search then Live Search then Bing, and that was a really great experience because, well, I mean, I was new, early in career, new to the industry and all that but working on a service was a really interesting thing.
Darinee Louvau (00:10:22): I've worked on services all my career. So I can't really say much about Box products, but just the whole thing about something that you control and you can control the releases on so that you can release frequently meant that you could do a lot more experimentation and trying out stuff and if it didn't work, that was okay, right? It wasn't like you had to get perfectly right before you shipped before you put plastic over it. I think that was probably pretty influential. I learned that at work, but then it became very influential outside of work as well and I feel that way now. So for example, I've done a lot of sewing. I've sewed all my life and I do a lot of costumes and stuff. And so fabric is expensive.
Darinee Louvau (00:11:00): Fabric and materials are expensive and you're going to iterate. You're going to make mistakes. You're going to have to iterate, but I don't want to spend a lot of money. And so I buy fabrics at the thrift store. I buy sheets, I buy king size sheets for $6 because they're cheap so that I can experiment and try it out and make lots of mistakes, but not worry about the cost of that, the literal cost of it. It was an inexpensive investment first place and it encourages me to try things out as opposed to, if I had bought the most expensive fabric for it, then I would be stuck doing planning and planning and planning and testing like before I even get to like start cutting it. And so it's just funny just to see some of the stuff that I learned from work tie back into the hobbies that I do now. That's something that I think helps that fearlessness of trying new things.
Thomas LaRock (00:11:46): So it sounds to me that what you're describing is you are perhaps a fan of testing and production.
Darinee Louvau (00:11:55): Well, I don't... Okay. Let's see. I want to... No, that's not like a-
Rob Collie (00:11:58): He's trying to trap you. Careful, careful.
Darinee Louvau (00:11:59): I know.
Thomas LaRock (00:12:00): I'm not trying to trap.
Rob Collie (00:12:01): Careful of Socrates here. Socrates is laying-
Darinee Louvau (00:12:03): Exactly.
Rob Collie (00:12:04): A landmine.
Thomas LaRock (00:12:05): No, I'm actually... I want to lead us to a good place. So my background, I'm a data janitor. And so I was brought up with the understanding that you had a production environment, a test environment, a dev environment, maybe even a QA environment. And you had all these different environments that were replicants of each other, but what didn't dawn on me at first was the understanding of the necessity of protecting that data. The idea that all these environments you needed to protect the data as if it was production data, regardless of where it was. So that was right before, DevOps became a thing and there was this big push for the short iterations, test something in production, something small, make a change, monitor, see how it does. And that to me was the beauty of what they were building at the time with Azure, was this wonderful sandbox where they could actually do that little bit of deploy a change, see how it works for everybody and roll into the Box product at some point. So anyway, what you were describing was to me, essentially testing and production.
Darinee Louvau (00:13:14): Yeah. I just... I didn't want to... I'm hesitant to just say yes, because of all the naysayers of testing and production, people will be like, "Here are all the things that are wrong with it," but no, it's true specifically about it though. The aspect that I think is interesting about production is just how much more realistic it is. And this is something that we did in being that was really, really cool, where I think it's important to testing and production in a way that's safe for the customer. You can't put crap out there. You can't put really buggy stuff out there and just think, "Okay, I'm going to have them fight all my bugs." That's not what that means. What was really cool, what we did in Bing was we looked at traffic patterns and queries that people did and then we were able to replay that data in our own test environment or maybe it was a part that wasn't production that was not taking traffic yet and we would work over some of that traffic over there.
Darinee Louvau (00:13:59): I thought was the coolest thing when I was working there. Was just being able to get some very realistic customer usage patterns and validate that. And so there are aspects of testing and production that use that sort of pattern that I think are awesome. I definitely wouldn't condone having your users test your product for you.
Rob Collie (00:14:16): Darinee doesn't want become a meme. I don't always test my code but when I do, I do it in production.
Darinee Louvau (00:14:21): Right.
Rob Collie (00:14:21): She doesn't want that. That thing you just said about replaying the user activity, for something like Bing, a simulation like that is almost like rerunning the universe-
Darinee Louvau (00:14:31): It's goals.
Rob Collie (00:14:31): It's so grand.
Darinee Louvau (00:14:32): I mean, it's great.
Rob Collie (00:14:35): We just created a separate little universe over here and we're going to run all the traffic that came in over that period of time. We're just going to rerun it on a system that isn't expecting it.
Darinee Louvau (00:14:45): Yep.
Rob Collie (00:14:45): And what a cool idea.
Darinee Louvau (00:14:46): Yeah. There's so much stuff, we worked on answers and when you build an answer, you think about, "What are the query terms that you want to trigger on? What sorts of things do you want it to bring up this answer?" Obviously, if you search for your showtime, then you should see the Showtime's answer. But then there are all these other weird patterns.
Rob Collie (00:15:02): Let's clarify for the audience what we mean by an answer.
Darinee Louvau (00:15:05): Right. An answer is when you query for something in a search engine, and it's a very definitive intent that you're trying to find and if the platform knows that intent, then let's return specific data to answer that question such as a very structured UI element that shows you the movie theaters near you with the movies that are playing and all the times for it, as opposed to just results of webpages that have-
Rob Collie (00:15:28): That's right. Yeah.
Darinee Louvau (00:15:28): Use the Fandago or whatever.
Rob Collie (00:15:30): The classic example these days is what's 17 ounces in milliliters. Right? Don't take me to a website-
Darinee Louvau (00:15:38): Right.
Rob Collie (00:15:38): Just give me the damn answer. I've come to rely on that. That is a utility that I use now.
Darinee Louvau (00:15:44): Yes.
Rob Collie (00:15:44): Why would I do anything else?
Darinee Louvau (00:15:46): So I bring up showtimes because I have a funny anecdote for that. I'm thinking about the things that I learned working in Bing and again, it was my first project at Microsoft and it was huge and it was such a great opportunity. Well, working on Bing, I learned a lot about just availability and service reliability and what it means to run a service. One of those things was when your stuff isn't working, it's an important thing to go fix right now. It's not something that you're like, "Okay, I'm just going to wait until next week to go fix that or the next release train."
Rob Collie (00:16:15): Yeah.
Darinee Louvau (00:16:16): That's something that's so... It's so important and so critical and that's something that working on services for the rest of my career at Microsoft, it's like, just... It's so important for that. But I remember once one weekend, it was during the weekend and my husband and I wanted to go watch a movie. And so I searched for movies at the local theater and the answer wasn't showing. And for me, I'm like, "Oh my God, my service isn't running." The service is down. I can't debug it right there on my phone. I can't go home because my husband wants to see the movie. I thought... And I'm ready to go email the other people that needed to know like, "Hey, this isn't working." We pull up to the movie theater and the movie theater is closed. It had shut down. And so for me, I was panicking that my service wasn't working because I was expecting the movie theater to be up 100% availability and my answer was-
Rob Collie (00:17:02): Yeah.
Darinee Louvau (00:17:02): But no, it turned out that the actual theater itself had closed.
Rob Collie (00:17:07): Yeah. The movie theater was 404-
Thomas LaRock (00:17:08): Wow.
Darinee Louvau (00:17:09): Exactly. It was unavailable.
Rob Collie (00:17:15): That is awesome. I love it sometimes when... If you work in software long enough of any sort, you eventually encounter these situations where you think it's a bug, but it's actually telling you the truth. And that is one of the patterns. One of the most delicious patterns in software that every now and then you get to experience. We've told the story a few times on this show, how we had a client who was building a power BI model with us to replace their existing reporting system. The first thing you do is you need to make sure the power BI model can match, can tie out with the existing reporting system for accuracy's sake, right? And no matter what we did, we couldn't get close. We were off by like 10% in this really important dollar figure and we couldn't get it to match and we were debugging validating every last little piece of logic. And in the end it turned out that their reporting system they'd been using for years had been wrong forever.
Thomas LaRock (00:18:08): That never happens. Come on.
Rob Collie (00:18:12): It's like, oh my God.
Thomas LaRock (00:18:13): That's so weird for me to hear.
Darinee Louvau (00:18:16): That's an interesting scenario too, because then it's a question of, do you go and fix that or not? Because it's important to have baselines you can compare against later, but if you've been running with something for so long and suddenly change it, then it makes you start to wonder, well, when you start comparing it, against the changed thing then what does that mean and its tricky.
Rob Collie (00:18:33): Yes. Yes. And there's also almost like socially or politically dangerous moment there when that's discovered, that's when you really find out what kind of culture you're working in, is it a CYA shoot the messenger culture-
Darinee Louvau (00:18:47): Right.
Rob Collie (00:18:50): I mean like how punitive versus we learn from the mistakes, that kind of thing. There's a lot of tests of character that happen at those moments and many anecdotes of consultants who were involved in those situations being fired for having spoken the truth. I mean, even Steven Levitt, that keynote at the past conference at one year talked about how he was hired to do something and he asked him like, "Do you want me to sugarcoat it or tell you the truth?" And they said, "Tell us the truth." And he told him the truth and they fired him.
Thomas LaRock (00:19:21): Yes.
Darinee Louvau (00:19:23): That's interesting. So one thing that I think as I've gotten older, I mean, I've been fortunate to be able to take breaks from work and be financially dependent but so we quit in 2014 originally.
Rob Collie (00:19:35): Let's slow cook that for a moment. We don't want to blast past that. So you started work at Microsoft in 2002.
Darinee Louvau (00:19:42): Yes.
Rob Collie (00:19:42): Did you graduate from college young?
Darinee Louvau (00:19:45): I did-
Rob Collie (00:19:45): Younger than most?
Darinee Louvau (00:19:48): I graduated when I was 20. I remember celebrating my 21st birthday at Microsoft.
Rob Collie (00:19:50): Okay. You got to Microsoft and weren't allowed to drink yet. And 12 years later, you and your husband retired.
Darinee Louvau (00:19:58): Yes.
Rob Collie (00:19:58): And at the time that you did that, I had been at Microsoft for most of that time. I wasn't anywhere close to retiring. So there was a lot of discipline. Y'all had a plan from the beginning and you stuck to it and it worked. I know that you were saying this to set up another story, but I did not want to go Russian past something so significant.
Darinee Louvau (00:20:20): My husband has a couple hobbies. He has fewer hobbies and one of them is around finances and thinking about money and being safe there. I mean, I think some of it has to do with his background. He grew up on a vineyard. His father was a grape grower and that job entailed a lot of risk. I mean, a lot of yearly risk. Would there be enough rain for water? Would there be enough sun for growth? Will the well dry up or not? And that was really scary for him growing up. I mean, I think his parents tried hard not to expose that to the kids too much, but he could see it. He could see when there were seasons that were really risky and scary, having all your eggs in one basket come October when the harvest happens. And that was really, I think, influential into his upbringing.
Darinee Louvau (00:21:03): You could either follow your parents or you could decide to do something called poly opposite and he decided to have a more stable job, number one, but also just think about how he managed our money a little bit better. And so for me, it was very easy. I made money, I gave it to him. I trusted him to do the right thing with it. But I think, it's not with one aspect of it, like my husband's value, like he just... That was one of his hobbies. We just got comfortable with a level of living that we were happy with. I mean, Microsoft was back then and is now still very generous with their new hire packages and I don't think it's anything to complain about. And so you can live very comfortably, I think with that income.
Darinee Louvau (00:21:41): And so we got used to a level of living that with our entry level salaries that we were very happy with. I mean, you can find happiness and lots of things that aren't so material. And so as we just got comfortable with the level of living that we are, then any reasons that we got, they all just went directly into 401k, into retirement, into investment, safe investments. And that was nice because we never felt like we needed to increase our spending, increase our level of living. Fortunately, we didn't have friends that compelled us to either, which is good. And anyway, just having that sort of idea of have a level that you're comfortable with and then just anything on top of that, just save it away. And so that's what we did. And then, I mean, also I have to... We're very lucky to have two Microsoft income. That's extremely... I don't know if you guys have mentioned that the acronym, the dual income, no kids, the dink acronym, that was something that was popular back there or our sitcom, which was single income, two kids, oppressive mortgage. Fortunately we weren't there.
Rob Collie (00:22:43): I've never heard that one, but that's a good one. Y'all even like at times anyway, rent out rooms in your house.
Darinee Louvau (00:22:50): Actually. Yeah. We had a couple opportunities to do that, which was again, very lucky opportunities that we had a renter who worked at Microsoft, but she lived primarily a couple hours away. And so she would work, I think three or four days a week and she would just come down, stay in the room. She didn't even use the kitchen. She had a bathroom in her room. This is before we had kids. And so that was a very fortunate situation. For her, it was probably dirt cheap rent. For us, it was additional income that we weren't-
Rob Collie (00:23:17): Yeah.
Darinee Louvau (00:23:17): We didn't have to spend into, but it was nice to just funnel that as well into saving. There was a program students from the UK who would come over for the Summer and would sell books in the United States. And one time they had knocked on our door, they go door to door. They sell books and then they also try to find lodging. So it's a pretty brutal experience, but-
Rob Collie (00:23:39): Yeah.
Darinee Louvau (00:23:39): But my parents actually, they encountered the same program years ago when I was a little younger. So I knew about that program and when they came by, I was like, "Yeah, sure. I can help you out and you guys could stay in the basement and pay us, I think like $25 a week in rent." And anyway, yeah. So we had a couple opportunities to rent.
Rob Collie (00:23:55): All right. So thank you for that side light. I think that was important, but you were originally driving towards a point that I completely deflected you from. Do you have any idea what it was you were about to say?
Darinee Louvau (00:24:06): Yes, it was going to work after first retirement. And I think you can feel this way a little bit after you take a break too, I think. I felt so much more refreshed taking two years off. Coming back in 2016, I worked on Mixer, which is a platform for live streaming. And I came back in and I just felt like refreshed and motivated to just do the right thing. You talked about sugarcoating things or not. And for me, I just, I felt like I'm going to do what I think is the right thing for the customer, regardless of the consequences. And it was just actually very liberating and very freeing. And maybe I just also came off with higher confidence in general, but it was a great time.
Darinee Louvau (00:24:45): I had a great time and I just... It was nice to come back to work with a different perspective. I wasn't there to earn an income. I wasn't there to play a part and do my financial engine. It was really just, I was there because I enjoyed the work. I didn't want to be not working. I wanted to build really cool stuff for my customers. And if there any moments where it's like, "Should we do this or should we do that?" I could always choose the right thing. I never had to worry about the consequences from that.
Rob Collie (00:25:12): I have to throw this joke in. I'm sorry. So I can set it up, you're retired, again?
Darinee Louvau (00:25:18): Yes.
Rob Collie (00:25:18): I thought you retired years ago. Yes, but what about second retirement? You retired again, so 2016 is when you went back-
Darinee Louvau (00:25:28): Yes.
Rob Collie (00:25:29): To Microsoft for a while. How long did you stay that time?
Darinee Louvau (00:25:32): Four years.
Rob Collie (00:25:32): Were you still there at Microsoft when COVID came along and sent everybody home?
Darinee Louvau (00:25:37): Yeah, I think that was actually probably one of the impetus for retiring again, actually. So it went back 2016, I worked for Mixer, which by the way, we could talk about that too because it was actually another awesome, great place to work in terms of scale, like in terms of thinking about scale and concurrency and all of that. So then the pandemic hit, I think in early 2020, and then around the same time our product got canceled. And so it was a... You can say Microsoft and join a new team, which is great, but it's in the context of COVID and it's context of working from home. That was hard for me. I think the hardest part is the initial meeting new people, building a relationship, getting to know folks. And for me, looking at that, seeing people at 2d, like on a screen is just a lot harder to work with. And for me, I'm so used to... People would walk by my office and I would go like-
Rob Collie (00:26:33): Yeah.
Darinee Louvau (00:26:34): I would go into something, I'd run the secret net, I'd do rounds around my floor because I get antsy just sitting around something would trigger my mind because I would see something. I would see somebody talking like, "Yeah, I got to go talk to this person about this thing." No longer have that was just really... Was really hard. And to further meet a new manager, meet a new team, figure all that out. It was hard. Maybe I took the easy route by quitting again but-
Rob Collie (00:27:01): Yeah, we've lost all respect for you after your second retirement.
Darinee Louvau (00:27:04): Well, I think with purpose a little bit at the beginning, I mean the kids... That first year, the kids had to... They switched to remote school. They then had to do a whole year of remote school and so it felt good to help out with that. I felt very fortunate, especially talking with my Microsoft friends. Now they're friends who have to work. That's really hard, like to be able to balance working with supporting your children in these new learning environments. I felt very fortunate to be able to do that to say, "You know what, I can do this. I can help out the kids and they can help them learn." The teachers were awesome. They put in so much energy to try to make it a great experience for them. So I felt like as a parent, I should try to do a little bit as well to help them out. So it was nice to not have work to burden that energy.
Rob Collie (00:27:53): The old saying or whatever it is, the revolution will not be televised. As you're talking about that, there's just so much suffering and difficulty and struggle that went on during the COVID years that almost by definition happened below the radar, no one sees it, all the things that were happening that were all happening like these little tiny silos, individual homes and everything, even the people who were dying from it, unless you knew them personally, no one was publishing videos of what ERs look like. Wasn't a lot of video coverage of that and so it was like, you hear all of this and you know it's going on, but you can't viscerally perceive it. It's just a very strange thing. Plus it also hints at how unsupported everyone was, right? The social network that would even just like emotionally buoy people during times like that was also basically disabled and just... I think I've just now reached the point where I can watch TV shows and I see gatherings in a scene, people are all out at dinner or something like that.
Rob Collie (00:28:51): And I don't get weirded out by it. For a couple of years, there's like, "They get to do that?" By the way I saw, just in passing, I saw an album the other day. It was just... I was just fast forwarding through Apple music and an album title went by that was... I thought it was really clever, Homeschool Valedictorian.
Thomas LaRock (00:29:17): So they don't have any siblings?
Rob Collie (00:29:19): Well, no twins because no-
Darinee Louvau (00:29:22): Right, exactly.
Rob Collie (00:29:22): There's different grades. Yeah. So, as long as you're not twins, you're okay. You're going to have covaledictorians. I've seen it.
Darinee Louvau (00:29:31): That's true, Mike.
Thomas LaRock (00:29:32): Not in that household.
Darinee Louvau (00:29:34): My high school had five salutatorians, two valedictorians.
Rob Collie (00:29:38): How many of those valedictorians and salutatorians were you?
Darinee Louvau (00:29:40): I was one of the salutatorians.
Rob Collie (00:29:43): Yeah. I'm so surprised that you were one of the seven. So she wasn't setting that up. She wasn't hoping that reveal that but I... This is why I get paid the big number of zero for hosting this podcast. Something you said a long time ago in this conversation now, using the cheap materials to power the iteration so that you don't get paralyzed in planning. And by the way, you and I both know that even if you got paralyzed in planning, when you actually started to cut fabric, you would discover that your planning was inadequate.
Darinee Louvau (00:30:18): Absolutely.
Rob Collie (00:30:20): Yeah. So you pay the paralysis cost and then you'd still screw up. Right?
Darinee Louvau (00:30:24): Yes.
Rob Collie (00:30:25): Whereas if you lower the cost of iteration, in this case using cheap material, right? Like buying sheets, bed sheets to use, that's genius, right? There was a very strong parallel for me to what's been happening like in my world, in BI, the cost of iteration used to be, in BI, used to be insanely high. You had no choice but to plan, plan, plan, plan to death-
Thomas LaRock (00:30:49): Right.
Rob Collie (00:30:49): And then go cut the fabric and discover that your plan was garbage even though you spent forever on it. I mean, the metaphor does start to break down a little bit, because we're not chewing up fabric. By lowering the cost of iteration, you can just get straight to the thing that matters, which is testing it against the real actual live circumstances of what you actually need to do. That has changed everything, which is really the foundation of our whole company is developing a business model that takes advantage of that capability of that, let's just get down to real immediately. It's proven difficult. Operationally there's been a lot of challenges, but the micro of our business is easy, solving an individual problem in the data space with these tools that are very, very iteration and friendly and very, very friendly to just getting started. That's the part of the business that's easy. As an overall business, there's been lots of challenges that we've discovered over the years and been addressing, but somewhere, there's a metaphorical blog post we could write about the cheap fabric.
Thomas LaRock (00:31:50): I like that.
Rob Collie (00:31:51): Yeah. And there is something now that I don't even know what it means, it's one of those new buzz phrases, data fabric. I don't know what it is, but maybe you should make your data fabric out of bedsheets, something.
Thomas LaRock (00:32:03): Didn't we talk about the Dataverse?
Rob Collie (00:32:05): There was Dataverse. We had Dataverse on the show, yeah.
Thomas LaRock (00:32:08): The Dataverse is the hipster data fabric.
Rob Collie (00:32:10): It is. Dataverse is a flavor of fabric.
Thomas LaRock (00:32:12): I don't know.
Rob Collie (00:32:13): Yeah. A side project of mine that eventually is going to see the light of day. At one point's going to Lampoon the phrase data fabric with, but what about data cashmere?
Thomas LaRock (00:32:22): I thought data corduroy.
Rob Collie (00:32:23): Data corduroy.
Thomas LaRock (00:32:24): Corduroy was the... What's the data corduroy here?
Rob Collie (00:32:26): Yeah, I don't know. I think cashmere really skewers it perfectly.
Thomas LaRock (00:32:29): Velvet.
Rob Collie (00:32:31): Now velvet sounds too good. So Mixer was a streaming platform, I was going to ask you why I've never heard of it.
Thomas LaRock (00:32:37): Yeah. Yeah. No, I heard of it and I'd forgotten that it was killed.
Darinee Louvau (00:32:41): Yes.
Rob Collie (00:32:42): Oh boy.
Darinee Louvau (00:32:43): That was such a great project to work on though. The project was great, but I think the thing that I took away the most from it were the people, the people that I worked with. That was the first project that I worked on at Microsoft that the real opportunity there for me was being able to work with other partner groups, besides engineering. It was working with the social team support, business planning, marketing. I mean, so often you work at an engineering company and you're just working on your service and you do a really good job with it, with monitoring, you gather requirements, you spec it out, you build it, you monitor it, you run it, all of that. But to be able to be exposed to and really take in the inputs of all of these other non-engineering groups was something that was such a cool experience.
Darinee Louvau (00:33:28): I mean, we were all on the same floor. So we were physically in close proximity, which was great for both organization and physically close. It was neat to see the social team was so positive. They were just such a great positive group of people because that's just their culture and their personalities. And it was neat to see that and to incorporate some elements of those different groups back into engineering as well.
Rob Collie (00:33:51): Why was Mixer on the face of it, naively, I could say? Well, why didn't you have that same experience in Bing? And what was it about the Mixer team that made it so cross discipline at its core?
Darinee Louvau (00:34:02): Well, some of it might just be me where I was in my career. I mean, I was pretty new working on Bing, my world was pretty small because I just didn't know about all the other stuff going on. I didn't understand how these things worked together because I was pretty new in my career. Many years at Microsoft, I started to understand better work that I had the opportunity to really work with them rather than just be only focus on an engineering problem. And so I don't think that it was any different because of the product per se. There are still those people involved. Some of it was unique though. Some of it was my own perspective, but some of it was unique, the cultures and the different divisions at Microsoft are different. I had a lot of fun in Bing-
Rob Collie (00:34:40): Big time.
Darinee Louvau (00:34:41): But when I switched to gaming and before Mixer, I worked on cloud streaming, cloud gaming and switching over to Xbox was very different culturally in terms of the fun came bottom up, I felt like. It was bottom up and it was supported by management as opposed to like, it was injected top down a little bit. We did some fun stuff. We had some really fun stuff in Bing, I remember this really awesome scavenger hunt in the city of Seattle. We were taking these cars around and this was really fun team event, but that was a very organized activity. Whereas the culture was a little more grassroots and more bottom up when you are working and gaming. Individually, we had a lot of fun doing stuff, creating our own fun instead of waiting for it to come down from above. And so I think just that tight cross individual and cross discipline interaction, it was just a different culture.
Rob Collie (00:35:38): It probably wasn't as well funded as something like Bing is, right? I was talking to someone the other day about the contrast between working at Microsoft and then starting a startup with my wife, right?
Darinee Louvau (00:35:50): Yeah.
Rob Collie (00:35:51): With no investors, like to really bring it into sharp relief. I said, even though they were a contractor and it was like a part-time job, there were people at Microsoft whose only job was to go and collect the dirty towels from the locker rooms and replace them with clean ones. That was a role. There was a team to do everything. Now it's just me and Jocelyn sitting in a room. Looking at each other going like, "Okay, so I suppose we need to do some accounting."
Thomas LaRock (00:36:23): I thought you were going to say, we need to do some laundry.
Rob Collie (00:36:25): Yeah.
Thomas LaRock (00:36:25): So it looks like one of us has to pick up these towels.
Rob Collie (00:36:28): That's right. Yeah. That's all we had was towels. When you're on the lean startup culture, you don't even have clothes. All you have is towels.
Thomas LaRock (00:36:33): We're paid in towels.
Darinee Louvau (00:36:36): I envy that so much about you and Jocelyn. I mean, I think that... I've had a very cushy experience working for Microsoft. There's a lot of comfort working for a large company. And so to have and start your own company with all the risk involved is very admirable. It intimidates me, like now that I'm retired again, I'm thinking, "Okay..." And the first year I spent helping my kids out now I'm thinking, "Okay, well what's my purpose now." Now they're going to go back into in-person school and doing great. So I can't really lean on that anymore to say that's what I'm doing, but I have to... I want to find something that feels purposeful and meaningful. The idea of a business is intimidating. It's interesting to you, but I'm not... The space that I've... I see you, I see Jocelyn and I think, "Wow, that's so cool."
Rob Collie (00:37:23): There was necessity involved. I mean, there was an inspiration for it. I got advanced notice of how a massive industry could be disrupted. That was a gift. It was a gift that Microsoft gave me. There's so many things in life that's like, "I'm so much more capable and competent and strong having gone through it." But now that I've gone through it, if you ask me, "Would you do it again?" There's some hesitation because it was so hard. Things are relatively smooth now. There are a number of phases of going through that development, that evolution that we're inherently dangerous. There's always a little bit of luck to go with the skill. I feel the same way about you wiring your camper wagon to keep the air conditioning running all night.
Rob Collie (00:38:10): I think I'd rather start a company than break out the raspberry pi or whatever it was that y'all were doing, which we definitely want to talk about, because we have had on this show, the bread pacitor, Chris Rae, another Microsoft that I knew. I don't think you would've ever crossed paths with him, had a submersible drone that stopped working when he was in Greece with his wife, he figured out on the base station that was on land. If he squished a piece of bread down on this capacitor, it would work again until the bread dried out. But he tried all kinds of different foods and everything and none of them worked except for bread.
Darinee Louvau (00:38:46): That's awesome.
Rob Collie (00:38:50): At that moment, I'm like, "He should call Darinee."
Darinee Louvau (00:38:53): That's awesome. It's interesting to think about hobbies and tune what hobbies I enjoy. The hobbies that are most interesting to me are the ones that I relate into projects that I could do for other people. I'll give you an example. So I spent a good amount of time modifying my car. I have a 2009 Mazdaspeed 3. I always enjoyed cars. I appreciate them. I enjoy driving them a lot. And I really wanted to modify my car and upgrade it. And so I redid the turbo inlet pie, the intake, I got an access support so I could tune it. I redid my high pressure flip up and internals remote amount. I did all this stuff and it's awesome. I also did those suspension, but watch it's cool because... It feels good because I did it. I used my hands to do something. Especially if you work on software, your whole career.
Rob Collie (00:39:43): Yeah. Physical.
Darinee Louvau (00:39:44): Being able to touch it and get cut by it, that's something that feels really good. You see it with your eyes more than just on a screen with a bunch of numbers go by on a graph. That's something that I think is really awesome, but I personally get enjoyment out of driving my car aside from how noisy and high NBH it has. But it feels really good because I did the project, but it's just me. Whereas recently I've gotten into 3d printing that is just a whole rabbit hole of fun. And it's just so fun to build stuff for other people to do a bunch of stuff that I either helps other people out or gifts or props for costumes. I've done a full size store breaker for a friend, I've done a full size Mandalorian rifle for a friend. It's just fun to build these things.
Thomas LaRock (00:40:32): I noticed the sword on her wall.
Darinee Louvau (00:40:34): Yeah.
Rob Collie (00:40:34): We need to talk about swords and lollipops. This was written down.
Thomas LaRock (00:40:38): Are those 3d printed as well?
Darinee Louvau (00:40:40): No, this is before I got into 3d printing and now of course, I'm thinking, "Holy cow, now that I knew more about 3d printing, how does this change my whole costume and prop building process with this new power?" Oh my God, there's just so much that I could be doing for it. But some of that, that's foam, some of that, that's PBC and installation foam.
Rob Collie (00:41:01): I love the juxtaposition of these brutal weapons of war. These swords and then the giant lollipop. Because there's some barbarian that goes out into battle armed only with the lollipop mace.
Darinee Louvau (00:41:16): That's from Castle Crashers.
Rob Collie (00:41:17): It is?
Darinee Louvau (00:41:19): Yeah, that's just the... Yeah. It's from Castle Crashers. Yes, that's just the hell of it. My daughter was that for Halloween a couple years ago.
Rob Collie (00:41:26): And Castle Crashers is the lollipop used as a mace?
Darinee Louvau (00:41:29): Yeah.
Rob Collie (00:41:31): Okay. So there you have it. It is a fairsome weapon. Not only does it knock you senseless, but it leaves a really sticky residue on the wound that's tough to clean up. So do you still have the camper wagon thing that you modified or have you moved on from that?
Darinee Louvau (00:41:50): We actually sold it last year, used car prices are pretty good.
Rob Collie (00:41:54): Yes.
Darinee Louvau (00:41:55): And so it seemed opportune. We had three cars at the time, so we weren't really using the van. It seemed a little excessive to have three cars, that's a lot. So we sold it, but we did keep all of the internals. We built a bed platform for the back that had out of wire frame, cut it down to size so that we'd fit the contours at the bottom of the van. It had the perfect height for pullout drawers underneath, had a custom cut. We cut it, custom cut sounds really fancy.
Rob Collie (00:42:21): By you. Yeah.
Darinee Louvau (00:42:22): Yeah. We cut an Ikea mattress down so that it would fit perfectly in there. And then I built these boards that would go across the middle row and the front row where the kids could see, they could each sleep across the front. So all four of us could sleep pretty comfortably in the van. And then we even built custom. But these reflect panels over the windows to help insulate the walls too when we're really cold. There was one place we stayed, I think it was in the spring in Eastern Washington and we woke up once and there was snow around. We couldn't open our sliding door because it had frozen shut, but it turns out that the side that was facing the sun had just melted enough. We could open up that side.
Darinee Louvau (00:43:02): Now that fan was really... It was great. It was really fun. It was a hybrid Pacifica. We sold it, but we would... I would love to sell the internals to some other interested group who also has a specific guy who might use that as well. I think that is one thing that's challenging with making a bunch of bespoke stuff is what do you do with it all. At some point you're going to run out of wall real estate, like you proud of it, but I don't know... I don't know yet what to do with all that stuff.
Rob Collie (00:43:28): The internals also included the special circuitry that you devised-
Darinee Louvau (00:43:33): Right.
Rob Collie (00:43:33): So I've now tead it up like three times.
Darinee Louvau (00:43:35): Sorry.
Rob Collie (00:43:36): Let's tell the story.
Darinee Louvau (00:43:37): It's a hybrid Pacifica-
Rob Collie (00:43:38): It's a hybrid car.
Darinee Louvau (00:43:39): Yes. Which is cool because you could plug it in to charge. And that also means that you can get air conditioning. So you don't want to run your gas guzzler air conditioning all night long while the engine on. But if you have hybrid, you can and because you could plug it in, then you could charge it off of no power from a campsite or whatnot. That was awesome. Except that we found out that it turns off after 30 minutes. It's probably some safety feature that we obviously wanted to bypass because we wanted to sleep comfortably all night. And so we were looking into different ways to keep it on for longer. We were looking at the canvas system, communicating with the car and all that, but couldn't figure that out. But eventually we prototyped this, actually. We decided, you know what, all we need to do is push the button, start to turn it back on.
Darinee Louvau (00:44:26): You have to press it three times because you have to get from ACC to on off to back to ACC again. And so we actually prototyped this first using Lego Mindstorms. We had a little Lego Mindstorms robot with a little motor that would rotate and push this button on and off, which was noisy but prototyping it, it was good experience because it showed that it was a viable thing. We also discovered that if you plug in the driver's side seatbelt, then it would... It wouldn't chime five times, it would only chime twice. So that also helped with that cycle, that one of the safety chimes would be reduced. So anyway, we figured out, we could just, basically... If we could just initiate the push button start three times on a 28 minute schedule, then we'd be able to keep the car running all night long. So it was... Is a simple raspberry pi and then later, or do we know circuit that we tapped into the back leads of the ignition, the push button ignition and then we just trigger that button.
Rob Collie (00:45:27): I love the mechanical representation of it. You know those boxes where you push the button.
Darinee Louvau (00:45:33): Right. And it turns itself off.
Rob Collie (00:45:34): And the top of the box flips up and a hand sticks out and pushes the button, it closes again.
Darinee Louvau (00:45:38): Yes.
Rob Collie (00:45:39): Sort of reminds me of-
Darinee Louvau (00:45:41): Exactly.
Rob Collie (00:45:42): So when you sold the car, you pulled the custom circuitry, like as far as the buyer's concerned, they were just getting a stock Pacifica.
Darinee Louvau (00:45:50): Yeah. It was totally stock by the time we returned it, but yeah, we have all this stuff for it. It's awesome. Any interested parties, whole bed platform, a whole push button, AC-
Rob Collie (00:46:02): What year and model though? Because the Pacifica is going to change slightly in dimensions from model year to model year and the bed won't fit anymore. And-
Darinee Louvau (00:46:10): That's true.
Rob Collie (00:46:11): Talk about bespoke. This is-
Darinee Louvau (00:46:11): I actually don't know what year it is. I guess it's 2014 or so.
Rob Collie (00:46:20): All right. So if you... Anyone out there listening, it's like a 2014 or so.
Darinee Louvau (00:46:20): Sorry, 2018. I just looked it up. It was 2018.
Rob Collie (00:46:23): 2018 Pacifica. Turn it into a camping mobile camping platform overnight.
Darinee Louvau (00:46:28): That's right.
Rob Collie (00:46:28): Hit us up. We'll connect you.
Darinee Louvau (00:46:29): There you go.
Rob Collie (00:46:31): Yeah. We won't even take a finder's fee.
Thomas LaRock (00:46:33): I was going to ask about Mixer, I had to look it up here. So that was supposed to be something similar to Twitch.
Darinee Louvau (00:46:41): Yes. We computed directly in that space.
Thomas LaRock (00:46:44): I thought it was weird that when it got killed that basically was sold to Facebook. There was something I saw briefly, it was about Mixer and then users were transitioned to Facebook gaming and I'm like, "Well, that's an odd partnership right there."
Darinee Louvau (00:46:59): Yeah. Well, Facebook gaming, I mean, they're getting into that space also of live streaming. It's good to have competition in the face, having just Twitch alone and that the creators are stuck. It's a great platform, but they're stuck if there's something that they're not happy with and there was really no alternative. So to have YouTube live streaming and Facebook gaming involved in... That's good. And I think Mixer helped to impact that market as well in the end, in terms of some of the cool things that we did that too.
Thomas LaRock (00:47:27): So I think ultimately my question was going to be more about for not just you and Rob, like when Microsoft kills something, what really happens to that code? Does it just always disappear? In this case, did it end up... Just the users went to Facebook gaming, but all that code that was worked on for Mixer does that live in some other product inside of Microsoft?
Rob Collie (00:47:50): It goes to the warehouse at the end of Raiders of the Lost Ark.
Thomas LaRock (00:47:53): Well, there's that graveyard outside building 16 or wherever where I found where Microsoft Bob is buried. They have that on campus. I did that... Maybe that's where is it like Mixer, there's a little tombstone for Mixer there.
Darinee Louvau (00:48:07): That's a good question. I got a couple projects canceled. I don't know if it's me or what, but I worked on game streaming prior to Mixer. Ultimately xCloud was launched. And so I think some of that got to be reused there.
Thomas LaRock (00:48:20): Yeah.
Rob Collie (00:48:21): Darinee and I met partially as a result of a canceled project. I was on windows live fantasy sports.
Thomas LaRock (00:48:28): I didn't know that existed.
Rob Collie (00:48:29): Yeah. It doesn't.
Thomas LaRock (00:48:30): It did.
Rob Collie (00:48:33): No, it never did. So we'll eventually get Mike Gratick on the show to talk about... Because he was involved in that longer than I was. Microsoft had partnered with Fox Sports to do a fantasy football site. Microsoft did the software and Fox did the content and the front end and all that. It was a disaster. Microsoft's code didn't... The service didn't hold up. It had all kinds of outages. It was not a good thing. And so Fox said, "No, we're going to break this partnership. We're going to go code our own. Thank you, Microsoft." And Microsoft said, "Okay, well, we're going to go learn from our mistakes and launch our own service. So they both said, "Well, we're going to take our ball and go home." Right? So Microsoft is going to be at the front end of their own thing and blah, blah, blah.
Rob Collie (00:49:15): Anyway, so I think the code was going great. There wasn't any problem with it, but shifting business priorities within the division, which eventually became Bing led to the cancellation of that whole project and my team woke up one morning and we were part of Bing. Well, it wasn't called Bing yet it was Windows Live Search, but might as well call it Bing. I went over and joined the Bing team, the answers team that Darinee was working on and program management in Bing at the time was so weird to me, it was much more of an engineering thing. I was used to being much more involved in the creation of something, and there wasn't really any room to create there and it wouldn't have been appropriate.
Darinee Louvau (00:49:56): You know what's funny is that as an exient like on the other side, I totally was wondering... I'm early in my career-ish, early-ish, I was also wondering like, "What do program managers do here?"
Rob Collie (00:50:08): Nothing. Nothing is what we did. I found the program management culture there to be incredibly toxic as well, because when there's not enough work to do you try to differentiate from each other in other ways, you try to... And mostly it was schmoozing. I was bitterly angry that that whole culture existed. Anyway so I eventually find my way out back into power BI, which was a great bounce. You were right to wonder what we were good for. And we did see some good examples, right? When I saw Janine take over a lot of the things that I had nominally been responsible for, I saw what I was supposed to be doing. I think she did a really good job there, but I was never that person. That was not my personality to be that kind of organizer.
Darinee Louvau (00:50:53): Yeah.
Rob Collie (00:50:54): I was more of a traditional product manager. It was a bad fit for me, but I stuck it out for a year.
Darinee Louvau (00:50:59): I switched to program management when I left Bing and joined game streaming and Xbox, and I definitely understood better what that role should have been back when I was in exient and so I appreciate it now. I mean, obviously, I was a program manager for the remaining years that I was there.
Rob Collie (00:51:18): There's certain things though that you could really only be exposed to working on something like Bing. I remember Ed having... Ed was our development lead. So for a little while there, it was me, Darinee and Ed that was this trio. Darinee and Ed were valuable and I was not, but there was this service, it was part of the answer service. Right? And so when someone types a query into Bing and presses enter, all these backend services get lit up and get a chance to examine that query and decide whether or not they have something to contribute. So for example, like the answer service, do we have an answer that's worth potentially displaying above the search results, I mean the same place? Like the ad service, do we have ads that are prepared to be displayed here, right? So it gets farmed out in parallel to all these parallel services, but then they all need to come back together and get back on this train to send the results back to the user and there's a time limit.
Rob Collie (00:52:22): The page has to be sent back to the user within a certain number of milliseconds. The train's going to leave if the service hasn't responded in time, right? And there was this one answer that we were running, I forget what it was. I think it might have been weather, that was periodically missing the train, it wasn't getting back in time. Ultimately the reason, and it was only happening, I think like on the East coast, only East coast users were seeing this outage. The weather service was running exclusively out of either the East or the West coast physically and the extra speed of light latency just getting across the country was making the difference between making the bus and missing it to be exposed, to something like that was insane. Where else are you going to run into an engineering challenge like that? And that's basically the only thing I remember from Bing.
Darinee Louvau (00:53:12): That's a good example of how... Where you place your data centers, where they go and how it impacts the overall experience just because of that. I will give another example where that... You do get exposed to that, which was game streaming. So search was great, awesome from an experience of learning about scale, learning about all of these services and globally distributed and requests per second. But for the most part, they were all stateless. It was just like a query comes in, you answer it, you respond to it and then you go back and then you're good to go. Later when I worked in game streaming and at Mixer, things were just a lot more... Were more stateful. If you're playing a game... When I say game streaming, I mean, you're playing a game on a client under computer, but the game itself is running in the cloud and it's just serving back frames.
Darinee Louvau (00:53:57): It's taking in controls from you and then serving back frames. And so that was actually now a lot more interesting in terms of latency and latency sensitivity than answers where 200 milliseconds was something that was important to get back to from a web request perspective. But here, we're game streaming, it was a new question. 200 milliseconds is not going to cut it. You're going to have a really crappy experience playing a game when you jump, then it takes 200 millisecond to see that action happen.
Rob Collie (00:54:24): Yeah.
Darinee Louvau (00:54:24): And so one of the things I did on game streaming was figuring out like, what level of latency is one detectable? Can you even notice that there's additional latency here? And two, at what point does it actually impact your game play? Which I thought was just one of the coolest project. I mean, I'm working at games for one that was so cool, but to also be able to design the test, the user research study around this, and I worked with the user researchers on this who were awesome. To conduct the study, we had a whole bunch of participants come in. We were playing different games of all sorts of genres. So we were curious for those two questions, didn't matter for the genre of games. It didn't matter for fighting games, didn't matter for racing games for platformers, that sort of thing. Which we had hypotheses, right? We would expect a fighting game where frame counting was so important that latency would be super important, maybe an adventure game, like a more casual open world adventure game that didn't matter as much or for racing, that would matter a lot too.
Darinee Louvau (00:55:22): And so we found some really interesting things there. We were able to identify a level latency of 45 milliseconds that would be a good place where either they couldn't see it or it didn't impact their game play and I'll tell you how we tested that too. Then we also saw some interesting trends for the different genres. So for finding games sure, we expected that to batter and it did, but it was... One thing that was interesting was for racing games, the detectability wasn't as high, because it was attributed more to the dynamics of the car, the wheels feel a little more splashy, not that "Okay, I'm used to this and then it's not happening." So I thought that was interesting, but as far as testing, does it impact your game?
Darinee Louvau (00:56:02): That was interesting because I was hoping to coordinate these user research studies, but I also decided to participate in it because I want to know what my users, who are these participants, what are they going through as they go through my study? So I decided to participate in it. And it was a... It was basically a session where we had people play Halo Reach in firefight mode, what we would do behind the scenes is slowly crank up the latency for some, we would not crank it up at all, for some we would and so on. And I sat there and I played, and it was the most grueling two hours of just playing, playing, playing over and over again.
Darinee Louvau (00:56:39): I was exhausted by the end of that session. And I don't actually know if my station was one that we adjusted, I see on maybe it was or it wasn't. I was just so tired at the end of that, of just playing over and over and over again. So that was just helpful for me as a program manager to participate as a user and see what my users are going through and maybe make a few changes.
Rob Collie (00:57:04): Would two hours of Halo Reach firefight exhaust you under normal circumstances?
Darinee Louvau (00:57:08): Probably. Yeah, I think so. I mean, I think it was something that I didn't really predict as somebody running this program, like, "Okay, well, let's just have people play and then we'll make these changes then." I mean, I didn't think about the emotional impact the exhaustion for it, which then later I had a good experience, better understanding in that. We talked about having fun on teams and all that. Well, one of the things that we did on that team was we did a 24 hours of Lemans, like race in for say, "You know what, let's do this." I had a boss who was really interested in this and has always wanted to do it. And I'm always a... Yeah, I'm always in to participate. So I think it was sports of four at the time started up, we started on Saturday at noon and we went all the way until Sunday at noon.
Darinee Louvau (00:57:54): And we had multiple teams of three drivers each, and we had rules like air damage was on, no stability control, we had to use the official car class as well. You can only switch drivers in the pit. So you drive around and then you'd have to go pit and then you'd like hop out of your chair and then the next person would hop in, get all set up, ready to go and drive off. That was also grueling. I drive a lot of recent games and I'm not very good at it, but when you spend 24 hours doing it, you get better, you really do get better. And that-
Rob Collie (00:58:27): I bet.
Darinee Louvau (00:58:28): I thought that was really cool. We should have taken track... Kept track more closely of our lap times and that, but you could definitely see by the end, like wow, we're getting a lot better. And it became more of a... Like a morale event. It was like a... It wasn't about who won. It was pretty clear who was going to win after the first six hours or so, miles and miles ahead. But it was cool because then it became more of an endurance race and everybody was participating and you wanted everybody to get through it. So middle of the night and checking in like, "How are you guys doing? You're holding up okay?" "Yeah. John's taking a nap and whatever." We'd have these shifts and we'd be checking in with each other. And it was just such a fun event to do, to build up our comradery and get better at the game.
Darinee Louvau (00:59:12): Unfortunately, I think the maximum race time was six hours. So we'd have to do four, we had to do four, six hour races. And so it was always nice when we switched, because and we tried to do as fast as we can. So we wouldn't break the momentum, but it was really nice to get your arrow damage reset, after six hours, which would immediately get blown away in the first couple lasts when inevitably you'd hit the wall or something.
Rob Collie (00:59:32): That's six hour limit. You just clearly needed like a circuit board that pressed the on off button-
Darinee Louvau (00:59:37): Exactly.
Rob Collie (00:59:39): Every five hours and 58 minutes.
Darinee Louvau (00:59:41): Yes.
Rob Collie (00:59:42): I really think you might enjoy meeting Chris Rae-
Darinee Louvau (00:59:45): Sure.
Rob Collie (00:59:46): Because I mean, he is a car racing person through and through. He participates very frequently in something called the 24 Hours of Lemons.
Darinee Louvau (00:59:56): Yeah.
Rob Collie (00:59:56): Have you heard of it?
Darinee Louvau (00:59:57): Yes. Yes.
Rob Collie (00:59:58): I mean, he's always off racing some completely inappropriate car around some track for long periods of time, he's constantly modifying things and is constantly buying cheap, exotic cars.
Darinee Louvau (01:00:10): Right.
Rob Collie (01:00:10): He has a Toyota Century on its way from Japan. Do you know what those are-
Darinee Louvau (01:00:15): No, I don't.
Rob Collie (01:00:16): The Toyota Century? You should look it up. It is the craziest thing. It is this handcrafted artisanal, the pinnacle of luxury in Japan.
Darinee Louvau (01:00:28): Wow. I see it now. Wow.
Rob Collie (01:00:29): It's designed to be driven by a chauffeur and you ride in the back. There's a panel that folds down in the seat in front of you and you put your legs out like an Ottoman. And they also are available ironically almost like dirt cheap for export after they've been used for a while.
Darinee Louvau (01:00:48): That's awesome. That's really cool too. I can imagine seeing that on a track for 24 hours. The endurance aspect is so... I mean, it's so different from just from speed, right? To be able to hold up to the battering.
Rob Collie (01:01:03): The car that Chris was driving this past weekend in one of these crazy things, it only has fourth gear now. It doesn't have any other gear. Do you remember writing the Power Pivot blog post-
Darinee Louvau (01:01:16): I do.
Rob Collie (01:01:16): For our site?
Darinee Louvau (01:01:16): Yeah. I just reread it actually this morning and it's so funny.
Rob Collie (01:01:22): I haven't reread it recently. I ran out of time, but you had a problem in Bing and you just kicked the tires of Power Pivot with it for like a day and got some really, really compelling results. I'm sure you've been... That's all you've been doing since then. Right? Is DAX, have you used it at all since then?
Darinee Louvau (01:01:41): So I love Excel, just generally. I haven't had awesome opportunities to do some deeper analysis on data that would require the sophistication of DAX and Power Pivot but I will tell you, I have a very awesome Excel spreadsheet that I enumerate all of my 3d printing costs and I have some really cool graphs of like, at what point did I stop doing resin printing and FDM printing like took off just from a investment, materials cost perspective. I've got slicers for my material brands and how much I've invested, how much I spent on each one, how I broken it down by materials versus accessories versus printer costs. I mean, I just love looking at that data and seeing where it's going to and just reflecting a little bit. It wasn't worth going down that path. Next time I know I can take those insights and do a different direction next time. It's just really nice to have it quantified and Excel. I mean, I just love it.
Rob Collie (01:02:39): Do you own your own 3d printers or do you use rent time on time shared ones?
Darinee Louvau (01:02:45): No. I have my own couple.
Rob Collie (01:02:48): David Osorio on our team is also a big, big 3d printing enthusiast. He goes by Oz. I was actually talking to him earlier this morning. I should show you the 3d printed stick figure Heisman that we use for our fantasy football league gets printed in steel.
Darinee Louvau (01:03:04): That's awesome.
Rob Collie (01:03:05): With a bronze, a steel bronze mix. It's pretty big, it's this tall, it's heavy.
Darinee Louvau (01:03:12): Pretty great. I would love to do like a lost wax sort of thing with... You 3d print something in a material that can be used in a lost wax technique where you can pour in molted metal and turn it into metal. I think that'd be cool.
Rob Collie (01:03:26): So this is printed by actual normal 3d printing. It's extruding the metal in a powder form, but with a wax or something binder?
Darinee Louvau (01:03:35): And then do-
Rob Collie (01:03:36): And then when it's done, they kill it. Right? And it melts and burns away all of the binder.
Darinee Louvau (01:03:43): That's super cool.
Rob Collie (01:03:44): And then it melts into a single crystal. So when you flick it actually sounds like metal. It's turned into one solid. I can't believe that it works.
Darinee Louvau (01:03:52): That's awesome. That's awesome.
Thomas LaRock (01:03:53): I need to drop... I have a hard stop. Wonderful to meet you and I do hope you come back sometime, because I'd like to hear more about how-
Darinee Louvau (01:03:59): Sure.
Rob Collie (01:04:00): You and Rob and Jocelyn all work together.
Darinee Louvau (01:04:09): I think one thing that's super helpful for any field that's helped me a lot is having a mentor in a space. I've done a lot of different things, but it's not like I pursued them myself. It's not like I was like, "I'm going to go learn this myself." I think the ones that I was most successful at was ones that had... I had a mentor I could talk to or somebody who brought me along and I'm always up for learning something new from that. But I think that that's one of the things that I realize now as an adult that I appreciate a lot more than when I was younger, took advantage of that in the past but I wish I had more access to mentors in new spaces that I want to pursue because I think that's one of the best ways to really get comfortable and familiar with it in the space. It's very challenging to do it by oneself.
Rob Collie (01:04:53): Well, if you decide you want to start a business, I'm here to give you all of the lessons that I learned the hard way. That doesn't mean that I'm really like some guru, but I can tell you all the things that I wish I'd known at the beginning and I could give you some advice because that's probably the thing that... That's not probably, it is the thing that I've learned the most about... Since we left Seattle is what is business like. In a word, it's a little brutal but it's also super rewarding in the same way like the way you talk about having replaced the turbo on your car or all those other things that I would never ever in a million years attempt. There's a real satisfaction in seeing this living, breathing organism that honestly, in some ways doesn't even need me anymore.
Darinee Louvau (01:05:38): Yeah.
Rob Collie (01:05:39): It benefits from me being around. But if I weren't here, it would keep going and because it's got a mind of its own now. It's self-sustaining and that's freaking awesome.
Darinee Louvau (01:05:48): How quickly did you feel those positive vibes from it? The reward from it because it's been years, right? And we;re talking about fast iteration cycles, right? And maybe it's because I'm greedy now that I think I want instant, I don't want to say it's instant gratification, but you want to be able to try something out, fail or succeed and keep going, iterate and keep going. How long does it take for running a business to feel that and be motivated to keep going?
Rob Collie (01:06:17): That's a really fascinating question. I've got a few nuanced, different answers. One of them is there's a lot of false positives on that cycle. There are moments where you're like, "Success." And then right around the corner is a comeuppance that maybe you could have seen coming. And so the Microsoft that I joined in '96, it was a very, very, very paranoid culture, always on the lookout. Like "Who's sneaking up on us?" Even when there wasn't anyone, there was a lot of looking over the shoulder and I get that a little bit now because there are threats that you don't expect. Some of them are self-inflicted. A couple of times I started to feel like I'd done really well and hubris seeped in,
Rob Collie (01:07:02): In that pride were the seeds of the next crisis. Even when things were like... We gotten through a lot of those problems, COVID came along. Microsoft was going to be okay, everybody knew that. Right? But there was no guarantee whatsoever that this company we had built was going to survive that sort of disruption, especially in the early stages of it and turned out we were fine. We even grew a little bit during COVID. And again, that worked out, it was just luck to draw, certain people's business models were destined to not survive that disruption. "You're a restaurant? Well, bye-bye." It just turned out that our business model being 100% remote, we still grew a little bit. The other thing I was going to say though, is that anything in this space... I guess there's plenty of success stories where it's like, "It just worked. It just happened and all of a sudden tada, success," right?
Rob Collie (01:07:53): But I have a friend right now who's working on a big creative deal with Disney. And three months ago they were on the cusp of signing this deal and it's going to change his life. And I was telling him, "Man, I hope I'm wrong, but what's going to happen here is this process of getting across that finish line is going to be so spread out that you're going to be deprived of the opportunity to celebrate it." That's exactly what's happened. Three months later, there's no sign that the deal's falling apart or anything, right? The starting gun hasn't fired. They haven't signed the deal. They haven't finalized it. He's just been in limbo forever. So by the time they'd eventually sigh and collapse over the starting line, all of his will to celebrate this amazing milestone will have evaporated.
Rob Collie (01:08:40): And I think there's a lot of that. There's a lot of that too. I think that your personality would be very well suited to it because you celebrate small wins, every little thing-
Darinee Louvau (01:08:49): Yeah.
Rob Collie (01:08:50): That you achieve, you feel it. And that's very sustaining. I'm the same way too and I think that's kind of what kept me afloat through all of those difficult times, was the incremental progress, as opposed to some feeling of pride and durable accomplishment. That takes a long time, depending on how you account for it. I'm now potentially 13 years into this experiment. You can call it nine years, you can call it 13 years, you call it any number in between. So there's a joke which is, let me tell you about my 10 year overnight success. There's a lot of truth to that. At the same time I meet people, they're like, "Yeah, I started a company and 18 months later, I sold it for $20 million." I'm like, "What the hell? What university do you come from?"
Darinee Louvau (01:09:34): If you go into it thinking that at the en, there's a binary question of, "Were you successful or not?" Then I think you're already setting yourself up for disappointment or failure that-
Rob Collie (01:09:45): Yeah. That's right.
Darinee Louvau (01:09:47): That really is about journey and the incremental wins and iterations that you make throughout and that it can't be simplified down to just yes or no.
Rob Collie (01:09:55): Let's just say financially. So where we're at right now as a company is amazing. We're like 50 plus people now and we'll probably be more than double this size a year from now. We're growing super fast. Everything is going at the moment until the next asteroid like event, like COVID comes along, things are going really well. Right? But still financially, I've been better off at Microsoft all these years.
Darinee Louvau (01:10:17): Right.
Rob Collie (01:10:17): So if you want to measure things, you can always find a way to look at it, but I've enjoyed this more than I ever than I ever enjoyed working at Microsoft. It's definitely a lot more personally satisfying even than having put features into Excel. This is much more personally gratifying for sure. You just need to find something, you've got a million ideas, I'm sure.
Darinee Louvau (01:10:39): Yeah. I have a few thoughts that I'm thinking about. I definitely want to retain that sense of personal... Everything I've done so far is bespoke and I don't ever want to mass produce something, but if I did create something that was interesting, that would be useful for more people then how can I still retain that feeling of gratitude that I designed that while still being able to reach a lot of people as opposed to just one at a time. So there's some interesting things out there that I'll think about.
Rob Collie (01:11:06): And like you said, it comes back also to what's your definition of success? I think we covered a lot of ground. Thank you so much for doing this. And I'm interested to see what you do next. No hurry but Darinee everything that I think you've ever been associated with, I've always been interested in what you've been up to. You and Scott and your broader family, y'all just do the most interesting things and I would love to still be living near y'all. I'm not sure that I would participate because I probably couldn't keep up, but I'd love to spectate.
Darinee Louvau (01:11:35): No, it was so fun. I'm so sad that you guys lived so close for just a brief moment and it was just so fun. Every time I pull out GuloGulo or Rattlesnake, I'm like-
Rob Collie (01:11:45): Those are great board games. Yeah.
Darinee Louvau (01:11:47): I'm going to pitch one thing, which is my husband's blog. It's relentlessoptimizer.com.
Rob Collie (01:11:55): Okay. Relentlessoptimizer.com. Tell us a little bit about that.
Darinee Louvau (01:11:57): I talked about his couple hobbies, one of them being finances but one of them is also around performance, optimizations and programming. And I think the blog title covers the mindset for both of those, which is, he gets into a space and then optimizes it. And so his blog is about his explorations into that sort of thing. Some of it's finances, some of it's video games, some of it is data processing. But anyway, I just wanted to throw a quick pitch out for it.
Rob Collie (01:12:21): Well, I appreciate that. And I even thought about asking the two of you to do it together, but I'm like, "No, no, no, no. Individual shows, this is a couple worth two shows." The fact that the two of you have signed on so readily for each other's personal missions and things, right? It doesn't sound like it was hard for you guys to do that. Still it's like, "Scott, you want to be on a financial journey," and okay. "Darinee, you want to be modding your car, like a 1950s character from Greece?" Okay, fine. All right. Well listen, thank you so much for taking time out of your second retirement, second breakfast to talk to us. As is so often the case, this podcast is an excuse to catch up with people that I haven't talked to in a while and it's great. It's a professional reason to reestablish personal contact.
Darinee Louvau (01:13:13): This was super fun. Thanks so much, Rob.
Speaker 3 (01:13:15): Thanks for listening to the Raw Data by P3 Adaptive Podcast. Let the experts at P3 Adaptive help your business. Just go to P3adaptive.com. Have a data day.
Sign up to receive email updates
Enter your name and email address below and I'll send you periodic updates about the podcast.