Latest / Oxide and Friends / Engineering Incentives... and Misincentives

Transcript

Engineering Incentives... and Misincentives · Oxide and Friends

• 0:00
  Speaker 1: Adam, are you there?
• 0:02
  Speaker 2: Hello, Brian.
• 0:06
  Speaker 1: We are getting subtitles now.
• 0:09
  Speaker 2: I can't hear you if you can hear me.
• 0:11
  Speaker 1: And not only can I hear you, I can actually see you being subtitled? Is that do we have an option to turn that off, Twitter?
• 0:19
  Speaker 2: Yes. I'll just oh, now I can hear you.
• 0:21
  Speaker 1: Yeah. I it's it's it's the thing we do.
• 0:24
  Speaker 2: It's the thing.
• 0:25
  Speaker 1: It's the shtick. And and and the shtick was complete with, Twitter spaces once again crapping out of me. It's just sort of the space. It happens literally every time. I and I I made the the arguably, the mistake of checking for an app update, before like, 2 minutes beforehand, which is always a I don't think that's a mistake.
• 0:48
  Speaker 2: I think that's, like, I think we've been burned when we haven't done that. And I would say that
• 0:53
  Speaker 1: I haven't burned both ways, I feel.
• 0:54
  Speaker 2: Well, I I feel like, Twitter Spaces is becoming a less effective torturer in that its failure modes are all sort of known. You know what I mean? I feel like the when it would really get us was when we would get something new or different or sporadic each time.
• 1:11
  Speaker 1: It was a bit of a roulette wheel.
• 1:14
  Speaker 2: Yeah. And this time it's like, okay, it's doing the thing it does, like, your thing crashes, I can't hear you, and then, you know, it's it settles in.
• 1:22
  Speaker 1: It's a it's a little Abbott and Casella routine for Twitter spaces. Exactly. Does does every Twitter space have this problem? If so, I would like to just dial into new Twitter spaces just so I can listen to other hosts going through this because it's kind of a music. Yeah.
• 1:37
  Speaker 3: I don't know. No.
• 1:40
  Speaker 1: No. Good
• 1:40
  Speaker 2: for you. I don't know.
• 1:41
  Speaker 1: So Good for you. Yeah. That's that's what you wanna spend your time with.
• 1:44
  Speaker 2: Actually, you know what? I see fewer Twitter spaces these days. I don't know about you, but, like, fewer of them seem to pop up than than once upon a time.
• 1:52
  Speaker 1: Okay. And okay. How correlated do you think that is crypto collapse? Because I feel like so many of those are about crypto. Right?
• 1:59
  Speaker 2: Spot on. No. I I had I thought about it. But, yes, I feel like most of them were, like, get your n f NFTs here.
• 2:07
  Speaker 1: Alright. So I'm gonna invite a bunch of people to speak. Not all of them will be Googlers or ex Googlers, because otherwise, this can turn into a Zuubler therapy session. I I mean, god bless it. I I I I do I do love the Zuublers.
• 2:18
  Speaker 2: Some of my best friends. Yes.
• 2:20
  Speaker 4: Some of my best friends are Exactly.
• 2:23
  Speaker 1: Okay. So we are here because true to essentially everybody's predictions. Stadia. Stadia? Stadia.
• 2:32
  How do you pronounce that?
• 2:33
  Speaker 2: Stadia. Right? Shirley. Stadia? Stadia?
• 2:36
  Surely?
• 2:38
  Speaker 1: Nadir. Look. I I'm just oh, I'm sorry. So, yeah, Stadia. Yeah.
• 2:43
  I would I wanna punch myself in the mouth for saying Stadia. Who would not be what kid would not get beaten up for saying Stadia? Alright. Stadia. Clearly, Stadia.
• 2:51
  The writing was on the wall for this one. I feel when they launched it. I mean, this was
• 2:57
  Speaker 2: Absolutely. I don't know. I feel like it just the premise of it, seemed far fetched to me. I mean, what do I know? I mean, apparently plenty.
• 3:06
  Speaker 1: But you have gamers in your household. No?
• 3:09
  Speaker 2: Yes. Yes. Yes. So I guess I have that.
• 3:11
  Speaker 1: What do they think? What did they
• 3:15
  Speaker 2: Literally, what is Stadia? I've never heard of that. That that that
• 3:18
  Speaker 1: was Okay. Well
• 3:20
  Speaker 2: That was the reaction I got.
• 3:21
  Speaker 1: Alright. Yeah. Not not necessarily a good time.
• 3:23
  Speaker 2: You're making that up or pronouncing it wrong.
• 3:26
  Speaker 1: But, I mean, it has been observed by many that inside the confines of Google and I wish I'd never worked for Google. So, obviously, those that have, correct me where this is incorrect. But it it it feels that it is well known that Google incentivized the creation of new things and does not incentivize quite as much the maintenance of existing things, to say the least. So that I mean, that and that Reddit comment was very true and the other things I've seen as well. But the, it's very true to other things that we have certainly heard.
• 4:00
  And so this was true at Sun back in the day where, I mean, do you remember hearing what kind of the criteria were for staff engineer, Adam, when we were
• 4:09
  Speaker 2: I No. Oh, dear.
• 4:11
  Speaker 1: Tell me. So famously so Steve Kleinman, who went on to NetApp, I believe. Right? Yeah. Was an early engineer at Sun, and, famously, he was promoted to staff engineer for v nodes.
• 4:29
  Like, the idea of a v node, like, that is what the the metric became. The the this is like the bar for staff engineer. It's a good idea. It's a good it is a good idea. It's definitely a good idea.
• 4:40
  It feels like, it's a little I so I feel I feel like there's something almost cardiological about and as a company itself matures, I mean, a company will be in a greenfield where very large things are possible simply because they haven't been done yet. And then as the company matures, the things that are done are going to necessarily be they're gonna feel more incremental, but in fact, they're gonna represent the the there's a lot of work that needs to be done, a lot of hard work that needs to be done. It's just they feel less less in terms of, like, new abstraction and so on. Is that fair?
• 5:16
  Speaker 5: Yeah. No. I think that's true.
• 5:18
  Speaker 2: I think and and certainly, it's, you know, it was it was easier to see the shiny new thing rather than the people, you know, making it real and fixing all the bugs and satisfying satisfying customers for sure.
• 5:32
  Speaker 1: So you had this idea that, like, alright. Well, if Steve Climate invented v notes, like, what did you do? Which I think is kind of incentivized people to do and I think Sun was did did not have this as bad as most, but incentivizing people to do kind of big grand things. And if you didn't feel like doing something big and grand, you were kind of, like, told you, like, yeah, you're not gonna get you're kinda at the end of the line. And, famously, I believe in Tom I'm not sure if you knew Joe Eichel at at Sun, but I because I think Joe Eichel was also I think went to Ypsilanti, I believe.
• 6:05
  But Joe kinda famously left because he was told, like, yeah, you're not gonna you're kind of at the end of the line because you don't wanna do something that's kinda grand and and new, which is I remember Jeff in particular was always very indignant about that because he thought there was a loss of a very good engineer because he was content to do very important engineering without inventing new abstraction. What I feel like we should incentivize engineers to to do. Right? That's I mean Yeah.
• 6:35
  Speaker 2: Absolutely. I mean, if everyone whatever. Obviously, you can't everyone can't gather around the shiniest objects and you just do that only. At, you at, you know, I know at at Joyant and at Oxide, you didn't have engineering levels. I would say to just expose myself as as a management dork a bit.
• 6:53
  At Delphix, we did have management levels and we were very careful to make sure that things like staff and senior staff, like applied equally across like, QA and dev and different dev functions and so forth, and really tried to phrase it in terms of the breadth of impact and, like, how many, you know, people and customers and internal processes and and those kinds of things to to kind of steer people away from just the new shiny. Okay.
• 7:18
  Speaker 1: So this is an interesting point because it's you're right. Like, we don't have leveling. I didn't have leveling at Joyant as well. Is leveling is that the original sin? Do we create these incentives because we have the levels?
• 7:33
  Speaker 5: Well, I I think that
• 7:34
  Speaker 2: I mean, let let me try this out on you. This might sound too Ryan from Preslow, but I think that people I think people do whatever the culture that they're in.
• 7:44
  Speaker 1: I think that's right. Yeah. Right. I agree. Right.
• 7:46
  Speaker 2: Like, like, just, you know, squishy humans kind of react to other squishy humans and that's what we do. And so, you know, I was thinking about it oxide and one of the things that is, is kind of a strong incentive is like demo day people like having a thing to demo at demo day. And we also do a good job of, you know, applauding, not just the newest shiniest, wizziest thing, but someone's test infrastructure that lets us robustly test some really annoying to test part of the system.
• 8:17
  Speaker 1: Totally. And I and so we we do demos every Friday. It is totally unstructured. It's basically, like, it is a forum for people to demonstrate their work. Like, there's it's no more there's no more structure to it than that.
• 8:30
  It's not, we don't grade people on it. People are not required to do it. Some people do it frequently. Others do it infrequently, but there's no it's just but I do agree with you, Adam, that, like, it's I I and I we've had, you know, folks join the company and they get kind of excited for when am I gonna demo something. And they begin to kinda think of it in terms of that.
• 8:53
  And I think that that's broadly so far pretty healthy. Right? You tell me if it's not. Yeah.
• 8:58
  Speaker 2: Yeah. I mean no. For sure. Seems very healthy. I think like any incentive and reward structure, it could be, you know, we could be looking back in 5 years and thinking like, oh, geez, how did we get here?
• 9:09
  Speaker 3: Because I
• 9:10
  Speaker 2: think that there are lots of incentive structures. I'm sure, like, like the one at Google that didn't intend to create these kinds of effects where people scurry away from the thing after release after they got promoted. But you can imagine, like
• 9:22
  Speaker 1: Oh, yeah.
• 9:23
  Speaker 2: Managers kinda taking their taking their reports aside and saying, okay. Like, you haven't demoed in a while. I really wanna get you, you know, more money or more kudos or or more
• 9:32
  Speaker 1: God, are you expecting the future? This is so terrible. A firmer pat on the
• 9:36
  Speaker 2: head or whatever it is, you know, and say, like, let's let's focus on demo, which is not to say, like, that that that it necessarily will be perverted
• 9:45
  Speaker 3: that way.
• 9:45
  Speaker 2: We can see it happening. Right? And and actually, that's most clear. I think in my experience in like sales organizations where, you know, setting up a comp plan is a very complicated thing to do. And I think that often sales engineers and, or pardon me, sales folks are motivated by additional money and they'll kind of optimize like protein folding.
• 10:07
  They're they're so efficient in finding the best way to like extract cash from, from a comp. Yeah. And, I, you know, I think engineers are different in that, like, I think the motivations are are often more complex, but, but still kinda find their own efficient ways to navigate it.
• 10:26
  Speaker 1: Yeah. And, also, just think one thing I wonder is, you know, as this domain has gone from, it has become more and more lucrative, have we perverted our incentives more and more? It feels like that way more people I I know people who got plenty of money that have been really taught to care about even more money in a way that I don't understand.
• 10:51
  Speaker 2: It is sort of like the simplest I mean, it's sort of the premise of capitalism. I mean, I don't know if you've heard about
• 10:57
  Speaker 1: it, but Tell me about it. Yeah. Yeah. I
• 11:00
  Speaker 6: mean, like An RD event?
• 11:02
  Speaker 2: Like, yeah. My my my wife works in an organization that one of the things they do is focus on CEO pay. CEO pay to motivate, like, you know, outcomes, but also to motivate like ESG and environmental stuff and, and diversity and inclusion. So, you know, they, they feel like CEOs kind of do what they're paid to do. And I think that's variously true for other folks.
• 11:27
  I don't, I don't know how ubiquitous that is in in engineering or in Silicon Valley. I think there are a lot of other incentives that people that motivate people for this.
• 11:36
  Speaker 1: I think there are incentives that motivate people way more strongly. I mean, I think it's like it is a certainly speaking for me personally, it it it I mean, yeah, it's like currency is great, money is great, but it there are other I mean, it is way more important for me to do something that is useful to other people, that is useful to that is something that is useful to my peers, that is useful to customers, me like that. That I just get way more motivation about out of that than
• 12:04
  Speaker 2: In your career, have you ever had someone, like, put an incentive structure in place that that had you not, like, fundamentally change behavior, but maybe kind of change priorities in somewhat?
• 12:16
  Speaker 1: Sure. Because after the acquisition when Samsung acquired Joynt, we they put in place a retention package that was a 3 year long retention package that was all, it it was backloaded, which I've used an active like, I view backloading any kind of incentive package as a very interesting kind of confession, which is like, look. This place kinda sucks. And the longer you're here, the more you're gonna realize that. So I've actually gotta pay you way more in these out years when you're gonna be, like, in excruciating pain in order to compensate you for that.
• 12:51
  I mean, I just feel like they're back loading that kind of retention. I think, like, in in retention itself is
• 12:57
  Speaker 2: kind of peculiar. You know? Well, in what it sounds like the the behavior it motivated was not
• 13:04
  Speaker 1: That's right. Okay. Yeah. For sure.
• 13:07
  Speaker 2: No. No. I mean, it is important behavior, I guess, in some
• 13:10
  Speaker 1: Yeah. I mean, and it's actually in many ways was clarifying because, like, the most miserable year of my career was also the most lucrative. So it's, like, it definitely and and where it's, like, okay. I feel like our responsibility to, like, my family to stay, but I'm, like, I'm definitely not happy. And so, yeah, it did motivate me to do something that I wouldn't otherwise.
• 13:31
  And I was, like, disgusted by myself about it. Yeah. You know, it didn't feel it did not feel good.
• 13:38
  Speaker 2: We don't feel that way about you.
• 13:39
  Speaker 1: Why would you say that unless you did feel that way about me? Why would you volunteer that? I mean, obviously, you feel that way about me, or you would have volunteered.
• 13:47
  Speaker 2: Well, I mean, the best part of selling out is the money. My friend loves to say.
• 13:53
  Speaker 1: I I I feel bad.
• 13:54
  Speaker 2: And I I'd love to hear from other folks, like, if there were if there were things, like, incentives that that changed behavior in in other ways. You know, I'm now, I'm thinking back to, like, my early days at SunBrian when Mike would Mike and you would, like, take the count of bugs that I had fixed and use this as, like, weapons to other to, like, compare to other groups of folks. And, actually, that was usually motivating for me to think, like, I I would sit there, know, watching Law and Order endlessly on the weekends, like, fixing random NDB bugs. Because I was, like, fired up to, like, pad that count. And, like, bugs are not, like, a great metric or whatever.
• 14:32
  Speaker 1: You know, I it is funny. I thought of exactly that today. Actually, because as I was thinking about, like, other ways in which engineers have been incentivized. Because you could go in and say, look. We're gonna pay you by by the bug fix, or we're gonna pay you by the line of code.
• 14:46
  It's just very easy to see how this becomes immediately perverse. Or actually, another example from Sun, Sun, do you remember the patent rewards?
• 14:56
  Speaker 2: Oh. Oh, yeah. No. That totally that
• 15:00
  Speaker 1: did change behavior. And in fact, I still, do I still have I think I still have at least one piece of furniture that was paid for with a with a with a patent downy. I don't know that I So
• 15:12
  Speaker 5: so some would pay us
• 15:13
  Speaker 2: Yeah. When we initially kinda filed a disclosure. There were different compensation structures, but it was something fairly lucrative, like $500 for basically filling in a web form.
• 15:24
  Speaker 1: That's right.
• 15:24
  Speaker 2: And then and then, like, a1000 bucks or $2,000 for actually getting the patent filed. Right. And then a piece of wood when you got it granted.
• 15:34
  Speaker 1: Yeah. And they changed that. So because the the the granting process could take so long. So do you it actually and I think this might have even happened before you joined. It used to be that they had this philosophy of, like, Taco First incentives.
• 15:47
  Well, you know, the patent process takes so long. You really wanna it it you you wanna front load the incentive, and you really want to incentivize people to file the invention disclosure, which is the internal disclosure inside of Sun, before you get lawyers involved. And so that was so that was, like, $2,000 for the invention for, like, the web form. And then it was, like, $500 for the patent granted. So you can imagine what might have happened.
• 16:17
  Oh, yeah. Absolutely. People I mean and it's like and I actually didn't I mean, I didn't I I didn't do this, certainly. But I discovered this when I we had an issue on Jurassic. And this guy had this world readable directory of all of the disclosures he filed.
• 16:36
  And he he filed, like, 300 disclosures. And it was disclosures for, like, things like, you know, lavatory in hovercar. You know what I mean? Where he's got I mean, he would have these events, and there's a question in the disclosure of, like, have you reduced this to practice? Which is, like, such a terrible phrase.
• 16:58
  Like, the it's, like, this is reduced to practice. And, of course, the answer for all of these things, like, no. None of the none of the hovercar laboratories have been reduced to practice. All of these things have just been, like, shit he dreamed up and was getting, like, you know, a decent amount of money for. So Yeah.
• 17:14
  Yeah. They that's why they changed it.
• 17:16
  Speaker 2: I mean, it was a different era too. Right? As we talked about 2 weeks ago when we're talking about, RIM and BlackBerry, where, like, these patent trolls would pop up and sue you over something that kinda never really worked for them, but, you know, good enough.
• 17:30
  Speaker 1: Yeah. I mean, it that's true. Like, Sun wanted a big port pat pat portfolio, and they kinda got it as a result, but they definitely had to fine tune the incentives. And I was definitely thinking about what you're referring to is we had a an office that was not very productive. And when when everyone was working at a direct physical site, and what we showed was that you you personally, as a new engineer at Sun, had fixed more bugs in the previous, whatever it was, 6 months than the entire office, like, by far.
• 18:03
  I mean, it was not it's not at all close. It's I didn't realize that you were actually changing your behavior as a result of that. That's kind of interesting.
• 18:10
  Speaker 2: I don't think I mean, I don't think I was, you know, radically these were things I wanted to do anyway by and large. Like, I was fired up to do it. But, like, yeah, it's like, I mean, it's like any kind of pat on the head. Right? Like, when when you give people attaboys or give people encouragement, like, usually, that turns into doing more of the same thing.
• 18:29
  I mean, which is also why encouragement and praise need to be carefully thought about. I I had a funny one. This is a a little bit, I I hope this is not too far out of bounds, but one time Uh-oh. With these 2 engineers no. No.
• 18:42
  No. No. Nothing to
• 18:42
  Speaker 4: do with you. We we had 2
• 18:43
  Speaker 2: I I I think this is. So we had 2 engineers who, you know, over Christmas, like, stayed around, fixed some problem. This one, I'm CTO of Delphix, and, The CEO, added all hands, recognized these 2, gave them some award. I think it was a stock grant, and really called them out. I had an engineer approached me a week later, and she was like, you know, I just worked really hard beforehand and did everything I could to make sure we didn't have these kinds of problems.
• 19:16
  And so are you saying, like, you want me to skip Christmas? And so and or Are you saying that
• 19:23
  Speaker 1: you, like, want me to, like, not actually, like, why would I fix the foundation when I can be a hero because the the building is Yeah. Yeah.
• 19:31
  Speaker 2: And so it it was it was an interesting one. Right? Because it it made me think about and I'm not sure, like, there was a clearly right answer in that circumstance. But it did make me think a lot about, like, what you praise in public. Like, what you praise in public to get other people to see those kinds of effects and what you praise in private because, you know, you appreciate the work but don't necessarily want that to be set as the standard of the pattern.
• 19:55
  Like, like the, like the V nodes as the metric for, promoting someone to staff.
• 20:02
  Speaker 1: Yeah.
• 20:03
  Speaker 2: Right? Like, once you do that publicly, you raise the bar and make it hard for everyone else. Whereas if you tell them privately, this was terrific work and we really appreciate it, and the messaging publicly might be different. You know, it it sets different kinds of incentives and expectations.
• 20:16
  Speaker 1: Okay. That's interesting. Private versus public. Also, what about if would she have felt the same way if the praise has been public, but there'd been no compensation associated with it?
• 20:27
  Speaker 2: Yeah. My I think that might have changed things as well. Right? Because I'm appreciating someone's sacrifice without necessarily saying that there was, like, a cash prize or what or stock or whatever was attached to it.
• 20:37
  Speaker 1: Yeah. It'd be interesting to know how because I feel I I do one of the things I definitely wonder, certainly, and those of you ex Googlers can kinda kinda chime in on this, but I feel like part of the big problem is that all of this stuff is tied to comp. So it's not just, like, you know, something that the kind of the culture values, but it's all, like, there there's a direct pass through the compensation. And that then you you start to compensation feels 0 sum. If, you know, if you're praised in an all hands and I'm not, I just feel like I'm less likely to be embittered by that than if, you know, you're compensated and I'm not.
• 21:17
  I feel it's, like, easier to be like, wait a minute. Like, there's a pie. It's finite. You just got a size of pie. I you just got a chicken sandwich that I can't get to put in big
• 21:27
  Speaker 2: Yes. 0 some sandwiches. Yes.
• 21:30
  Speaker 1: But I wonder, like That's
• 21:31
  Speaker 2: a good point.
• 21:31
  Speaker 1: Like, like, how much of that is is and, like, how much of it also I feel like this is this peril of I mean, I the the bug example had occurred to me as well because we did kinda open Pandora's box there a little bit in that we or we we cracked it open. I'd like to believe that we shut it quickly after that. But the I mean, in many ways, it was confirming something that we are confirming is the wrong word. It was it was showing quantitatively something that felt very true qualitatively.
• 22:00
  Speaker 2: And for those listening, yes, this made me a lot of friends at that office.
• 22:07
  Speaker 1: Well, part of the thing that was was just saying, but it was came in through an kind of an acquisition. We had a we had a VP that liked to we had a VP that really liked to hire people. That was like the oh, actually, talk about merchant incentives. There, that's a good one.
• 22:22
  Speaker 2: Yeah. Totally. Right. Size of your
• 22:24
  Speaker 1: Size of your org.
• 22:25
  Speaker 2: Hire. Yeah.
• 22:26
  Speaker 1: And he was all about the size of his org. Like, that was how he and I think this is the way the way a lot of people measure themselves. And so he went and like, there was a company who's basically gonna downsize a bunch of people, and he went and hired them all with I mean, in a total, like, no Huddl kind of fashion. And it was kind of we had to take them all or none of them, and this is in the late nineties or 2000 when it's like, you know, we can't find people anywhere. And so we ended up with a bunch of people that were not people we should have hired, is what I pulled down to.
• 23:02
  And in part because of this of this perverse incentive that that he was incentivized to build his work instead of actually doing the right thing. Yeah. I think it also gets to, like, why people do things, though. I mean, this is just with like, wait. What makes people feel good about the work they're doing, and where do you drive that satisfaction from?
• 23:22
  And I think if you make that too extrinsically motivated too much of the time, people get strung out on it.
• 23:28
  Speaker 2: Yeah. But I think a lot of what we're talking about is that I still extrinsic motivation. Right? Like, even when you're saying that you like and and I like I I share this feeling that I like doing things that my colleagues appreciate, that other people appreciate. Right?
• 23:41
  Like, I like building some random crate and seeing someone consume it that I've never seen before. Right. I mean, I think those are extrinsic. Right?
• 23:50
  Speaker 1: Yeah.
• 23:50
  Speaker 2: I mean, there's there's also this this you know, I like the the satisfaction of a job well done, you know, in, like, in doing something beautiful, but I like I'll I'll readily confess that I'm, like, much more extrinsically motivated. I mean, this will come as no surprise to you as I show, like, every scrap of code I've ever written when I was in my twenties to you or whenever, like, requiring a firm pat on
• 24:12
  Speaker 1: the head. That's that you're being too hard on yourself. But, no, I you're right. I mean, that even where we say there's intrinsic motivation that there is, it is like, we have a colleague. I mean, I hold all of our colleagues in very high regard, but we hold a colleague whose highest praise, in my experience, is Slick.
• 24:32
  Like, that's Slick. And I and, yeah, like, I find, like, those are, like, serious Scooby Snacks. Like, you say, you know, we call something you did just like it. It feel yeah. It feels really good.
• 24:41
  But it feels good in part because it it feels like that it feels deep. You know what I mean? It feels earned. It doesn't feel like it's not just given out for anything. It's not cheapened.
• 24:53
  And it's not cheapened in part because it has no there's nothing monetary associated with it. You know, he's he he is under no obligation to issue a certain number of slicks per quarter. You know, we he's not stack ragging the organization, figuring out how he can also give out an infinite number of clicks. You know? I mean, there's no nothing
• 25:13
  Speaker 2: Well, except for they do kind of devalue. Right? You do get sort of flex
• 25:17
  Speaker 1: it's Slick inflation. Slick inflation, now that I say that, would actually be pretty though I would have move to something else. It's true. It would it would be devalued. That
• 25:25
  Speaker 2: is true.
• 25:27
  Speaker 3: It
• 25:27
  Speaker 2: didn't mean I you know, I don't know if you have this feeling, Brian, because I I know when you left Sun, you were VP of Eng and then CTO. And, you know, I I had been a, like a manager and then a CTO and then later CEO and then now nothing.
• 25:42
  Speaker 1: But it made me think
• 25:43
  Speaker 2: a lot about, about, like, about praise. Right? And it changes. It's weird. It's weird how the things you say, both positive and negative, cast a much longer shadow.
• 25:54
  And it and it was, you know, I was working for a CEO once who was very stingy with praise.
• 25:59
  Speaker 1: Yeah.
• 25:59
  Speaker 2: And, and and, I remember just thinking and and saying to him at one point, you know, this is the the cheapest thing that you have to give out and it and it actually matters so much.
• 26:11
  Speaker 1: It matters so much. I agree with that. Yeah. I I I so, you know, did I tell you when I got religion on this? When we, were so you and I were together at an Oregon side of Spain, and we had a bunch of, younger engineers, couple younger engineers.
• 26:28
  And sun is collapsing.
• 26:30
  Speaker 4: It's the fall of sci con.
• 26:31
  Speaker 1: You you and I are running to the roof with our with all of our things waiting to get helicoptered out. And, I remember I was with 3 of the most junior engineers, one of whom still works with us today. And I was like, look. You know? Hey.
• 26:47
  I'm leaving. I just wanted to tell you, you know, you 3 did a, like, a great job. And I'm they they were like, oh, good. Okay. Okay.
• 26:57
  That's really good to know. I'm like, wait. What? And like, well, we just like I mean, we kind of assumed we were doing a good job because, you know, no one told us we were doing a bad job, but okay. That's good to know.
• 27:07
  Speaker 5: We're doing a good job.
• 27:08
  Speaker 1: They were kinda like, oh my god. Did we screw? I screwed up. I felt like I screwed up. I did not give you nearly enough.
• 27:15
  And so my big belief is, like, if you are and and, you you know, folks that have worked at Oxide have probably heard me say this. If you tell me something positive about someone else, I'll be like, tell them that. Make sure they know that. And, I mean, praise shouldn't be disingenuous. But, like, don't be stingy with it.
• 27:31
  If you've got if you're if you're feeling something positive about someone else, say it for sure.
• 27:37
  Speaker 2: I I mean, I I have become much more a fan of praise, and not once has somewhat have I said, you know, you did a great job on this or I really appreciate this and have the person say, well, you know, shove it. Like, I don't care.
• 27:48
  Speaker 1: Like, that's a little bit. That means nothing to me.
• 27:50
  Speaker 3: Yeah. Exactly.
• 27:51
  Speaker 2: And it's and it's weird how these, you know, these these kind words can mean so much to our colleagues or, you know, to open source maintainers or whomever because, like, people often don't go out of the way to to give that praise. And it's a lot easier to give negative feedback.
• 28:04
  Speaker 1: But you also don't know. You're like, oh, okay. Good. Thank god. That is good.
• 28:07
  I wasn't sure. Like, I don't know. Are people using this? Was this a good idea? Is this working?
• 28:12
  You know, you just know. I said no. Satnam, you had your your hand up, and then, Ken, you were joining us as well.
• 28:18
  Speaker 4: Hi, thanks. I mean, I've been thinking what you've been saying about incentives and money. And I think of all my academic colleagues who clearly are not motivated by money. But, yeah, they do great work. So what is it there going on there?
• 28:33
  And I think part of that is you want the respect of your peers, right? And if you can create systems where that becomes a currency other than just your title and how much money you're paid, I think that would be like a kind of a better world. Anyway, a little bit what I was thinking about, giving people praise because, like, so the academic community has got all these prizes and awards for academics and students and postdocs and so on. So but but I think building systems where we have you you could somehow create respect rather than, dollars. Maybe perhaps something to be learned from the academic community for that because that's something I think that works very well where they Don't care about their level.
• 29:15
  They no one cares if you're an assistant professor or a full professor and you and your salary there is is very low compared to what you could be earning in industry. So it's not all about money. I think how how you are accepted and how you feel think your peers accept you and what they think of you. For a lot of us, I know for me, that's a very important and powerful thing.
• 29:35
  Speaker 1: Super powerful. I know that's a very good point. And I feel like I feel one of the challenges that that, you know, we have in and I think academia also suffers a bit is, can I do I cheapen myself if I praise your work? If I if I praise your work, am I, you know, am I saying that you that that you are my intellectual superior, you know, in in academia? And you you really wanna not get kinda caught up in that in that hierarchy and know that you can safely praise someone else and not worry about it cheapening yourself.
• 30:07
  But, yeah, you're right about that. It's only academics are not motivated by money, clearly, because they could make a lot more elsewhere. Are academics, how content are academics in your experience?
• 30:22
  Speaker 4: Yeah. So I mean, I've just beat off some threads. There's this one bad thing in Twitter where someone said he's an academic and he said how much money would a company have to pay you to make you leave academia? Which kind of, like, ticked me off. I mean, I I I wrote a kind of a snarky reply because it was all it was written on the surface that I am doing a good thing as an academic and I could sell my soul and go to the industry and therefore do bad things but at least have been paid money to compensate for it.
• 30:51
  So I think I do come across a few academics that feel very unhappy about what they're earning, but only when they've just had to teach a giant class
• 31:02
  Speaker 1: of Teaching is really hard.
• 31:06
  Speaker 4: Well, I mean, I know a ton I mean, I used to be an academic. You know, clearly, I wasn't doing it from I had mode I was very motivated, very driven. I worked the hardest I worked in my life was as a globally paid academic. So clearly, if somebody got in there other than the the money
• 31:19
  Speaker 1: I think a lot of them are are architect, but
• 31:22
  Speaker 4: they won't complain, as is, as they want about all of the things that are, that are not good. And, you know, in in computer, I mean, most people's academics don't have a choice. Right? We've chosen like, my sister's an academic in English literature. She's a professor at university that she she can't go and work for Google, right, qualifications.
• 31:40
  Most of our friends who are academics can, if they wanted to, right, they have that option to convoke with the fee and they don't. Right? They stay as a academic.
• 31:49
  Speaker 2: So I
• 31:49
  Speaker 4: think a
• 31:49
  Speaker 1: lot of them are very content very content. Yeah. Interesting. Well, I feel like teaching is one of these things that you also have to be I mean, teaching, obviously, a big part of academia, and it's talk about, like, a very distant reward. And I know I mean, you know, we should all go back and thank the teachers that had the most impact on us because I don't think they hear that frequently enough.
• 32:12
  I I let's go to Ken and then to Ian. Because, Ken, I think you joined
• 32:15
  Speaker 3: earlier. Yeah. Hi. Can you hear?
• 32:19
  Speaker 1: Yep. We can hear you.
• 32:20
  Speaker 3: K. Great. Yeah. I think one of the things so it is around visibility. So, like, a lot of bug work in general and fixing and making systems more resilient.
• 32:37
  Largely, if you do it by yourself, tends to go unnoticed or be invisible unless you actively try to make it more visible. And one of the things I've noticed, I've been developing a chaos engineering program, from the ground up. And it's it's, a you can have it be a team or a person that enables kind of the visibility in an organization to go to to sort of pan out to make that work a lot better.
• 33:09
  Speaker 2: Hey, Ken. We're kinda losing you, I think.
• 33:14
  Speaker 3: This isn't better.
• 33:17
  Speaker 1: That's we get I mean, we blame Twitter spaces, of course. And, of course, also, Adam, this could be the quirk where the recording is fine. Oh, yeah. I'll
• 33:24
  Speaker 2: check that. I'll I'll check
• 33:25
  Speaker 1: that one. Yeah. Just continue again. Okay.
• 33:46
  Speaker 2: Ken, sorry. I gotta cut you off. The the I I think you're unintelligible both to us and that I'm recording. Sorry.
• 33:52
  Speaker 1: Although it makes that is better because we have had this failure mode where the we are unintelligible to one another, but then totally intelligent with the recording. And then you go back and listen to the recording. We sound like we're having a stroke on this. So, Ken, I think what you're saying I'm I'm getting kind of, like, every third or fourth syllable about how you make bug fixing more visible in an organization. And I think there are some actually interesting ways to go do that and and to make, you know, how do you kind of uplift this?
• 34:22
  I mean, I do feel, Adam, like, one of the things that we do try to do is you gotta make a big deal out of things that are really hard, but don't look like a big deal. And bug bug fixing is one of those things. It's really hard, but it it you know, great. Like, the system now works as well as I thought it worked before you did any of this. Like, you and I think it's incumbent upon an organization to really encourage that.
• 34:44
  And put it it's not easy. I the Ian. Yeah.
• 34:51
  Speaker 7: I would say to the, academic, line of argument. In in terms of computer science academics, I think a lot of people churn out of that as a career path, significantly earlier in their journey now than they did, say, even, like, 20 years ago or something. Right? So a lot of people year 1, year 2 of their PhDs, or if they do complete a PhD, will turn out into industry at this stage because, there's pretty significant financial incentives to do so as well as, the the difference in the amount of freedom they get to pursue the work that they want to do has slimmed over the years, whereby industry is offering some pretty free form positions whereby you can pursue the kind of research that you want to do and publish it in public. So, a lot of people I I I don't yeah.
• 35:48
  I think that if you are looking at people who are currently lecturers, they've already filtered out a lot of people at that point in the, you know, 4 plus years of training to go up to that point.
• 36:00
  Speaker 1: Yeah. For sure. And I think I mean, actually, for whatever it's worth for me personally and, Adam, I feel like you also, like, thought you're gonna go in academically.
• 36:08
  Speaker 2: Yeah. I was pretty confused. Yeah.
• 36:09
  Speaker 1: Do you are you still in the mode that I felt you were in for several decades where you still believe that?
• 36:16
  Speaker 2: No. I mean, I know that I'm too old if that's if that's Okay.
• 36:19
  Speaker 1: If that's
• 36:19
  Speaker 2: the Oh, no.
• 36:20
  Speaker 1: No. No. But the okay. This is like a new development. I don't mean to make too big of a deal.
• 36:23
  But this is, like, I I feel this is like one of you like, you you had, like, a, there was an alternate timeline where Yes. Right.
• 36:32
  Speaker 5: I feel I feel
• 36:33
  Speaker 2: like there have been, I mean, like, even as recently as, like, maybe 2015 or 16, I thought about, you know, seeing if I could join Phil on his talk adventures, down at Stanford and things like that. No. For sure.
• 36:44
  Speaker 1: And you had not yet used up your NCAA athletic eligibility ultimate.
• 36:50
  Speaker 2: That was that was a factor. I will Right?
• 36:53
  Speaker 1: I mean, that was part of the that was part of, like, that particular ultimate timeline if I recall correctly.
• 36:57
  Speaker 2: Absolutely. That I would be, like, the old man on a on a Frisbee team of 21 year olds thinking that I could school them, but actually just being run around.
• 37:07
  Speaker 1: Wait. But but playing dirtier than they'd be going to play and throwing your visor much more frequent. Absolutely. Calling way more travels for sure. But so you but to that, at some point, you realize, like, okay.
• 37:19
  I don't think I'm gonna come. Well, I'm gonna take it. How about
• 37:21
  Speaker 4: this You
• 37:21
  Speaker 1: know, I think I say that. I'm, like, I'm actually not totally convinced the dream is set for me either.
• 37:24
  Speaker 2: You know, I I was I was at my, my older son's back to school night, and his physics and computer science teacher, same guy, went to MIT, you know, in the nineties, worked in industry for a while, hated it, and has been teaching at, at the high school for 15, 20 years. And he made it sound so romantic, that that it sort of, like, breathed on those embers a little bit because he clearly loves it and the kids love him. And I think
• 37:55
  Speaker 3: there is there
• 37:55
  Speaker 2: is something that, you know, I I love teaching. And, like, I get a lot of it, in, you know, a lot of those same, benefits from like working with colleagues and mentoring colleagues and over the years. But there, there is something, that continues to be very romantic about, you know, teaching, probably not in high school, but teaching in college?
• 38:16
  Speaker 1: It's very it's potentially very high impact. I mean, you can do you can really affect people's lives. You can you can change the direction of the lives.
• 38:21
  Speaker 2: Look. I mean, conversely, like, pull pulling those, you know, academics out of academia into industrial research, conversely, feels also like it has a big impact. Like, you're you're underserving you know, you're you're kinda eating the seed corn. Like, you're underserving those youth. So I think it's I think it's kinda complicated.
• 38:41
  Speaker 1: Who's who are we eating? Am I are we seed corn eaters
• 38:44
  Speaker 2: in this That's a good question. No. I would say Are we
• 38:48
  Speaker 1: just buying the food stocks for future
• 38:49
  Speaker 3: generations in this metaphor? I'm trying
• 38:51
  Speaker 1: to
• 38:51
  Speaker 2: keep that I think, like, the folks who who tempt, great, great teachers and professors into industrial research, so that they don't have to teach anymore. Yeah. A little
• 39:03
  Speaker 1: bit. Interesting. Yeah. It's a good point. It's a good point.
• 39:04
  Well, so but I'll tell you the thing that turned me off about academia, that is still a bit of a turn off about it's still a turn off about academia for sure, is the the the the novel phobia that that everything only the things that are new have merit, and it felt like the there were it was few and far between the the academic papers that I could see myself writing with the ones that were actually going into implementation details didn't matter, didn't felt like them. It didn't it didn't feel like they mattered anyway in the, and this is part of the reason I love the the the late Yohan Lidki has got terrific papers because the implementation really mattered in those papers. I don't know if you ever read any. Like, do you know if you read any? It's really good stuff.
• 39:47
  But then I and and I kinda fell in love with those, and then you realize, like, wait a minute. This is a sliver of the and this is, like, when we went to a Deepak Academy. Yeah. We we went to oh, we're such a bunch of jokers. Well, we went to this tech conference on debugging, and I'm like, we are going to the the cathedral of academia.
• 40:06
  Like, this is the the this is the the mecca for academia because there can be no higher calling in academic computer science than working on debugging. And then we arrived there and realized that this whole community feels like they are of like like, they're boat people. Like, they are adrift. Like, no one will take them. I mean, they're just like
• 40:25
  Speaker 2: Yeah. Pariahs. And I I don't think the conference happened again.
• 40:29
  Speaker 1: Yeah. Right. And then what here, we thought the conference happened infrequently because it was, like, the Olympics. It was like, you can't have the Olympics every year or you cheapen it. And it's like, no.
• 40:37
  We can't, like, get a venue most years. Like, we can't get Yeah. So, Dan, you had your your hand up.
• 40:45
  Speaker 5: Sure. First a little side comment about industrial research. I'm gonna take, some exception to what Adam said. I mean, look look at Bell Labs. They did UNIX, c, the transistor, the laser.
• 40:56
  I mean, all of that stuff was side effects of industrial research, and it really has changed the world. Well, hold on, Dan. Just to
• 41:04
  Speaker 2: be clear, like, I I'm not anti industrial research. I'm just saying they're they're kinda two sides of the coin that, that if you're, if you're taking someone out of academic research where they're, you know, inspiring young minds potentially, and bring them into a different context, like there, it's not purely you know, there there's there's a there's cost benefit needs. That's that's my only point. Not No. I I my Bell Labs denier.
• 41:25
  Yeah.
• 41:26
  Speaker 1: That's cool. Why would he say he's a Bell Labs denier unless he's a Bell Labs denier? I mean, obviously, the
• 41:30
  Speaker 4: Civic No. No. I
• 41:31
  Speaker 1: I I I think I think it's just a little
• 41:33
  Speaker 5: more of a back and forth there, Right? I mean, you know, people will jump between industry and academia and, you know, like the Bell Labs folks would talk to people in academia all the time, and they would go down to Princeton and teach classes and blah blah blah blah blah. But I wanted to touch on the, engineering incentive thing. So this all kind of kicked off as near as I could tell with this, tweet about a Reddit comment. And I don't know who wrote that, the original Reddit comment or what the context there was, but it was definitely a little bit misleading about the specifics of Google.
• 42:05
  So Google's big thing wasn't so much like you have to launch. It was, you have to demonstrate impact. And the easiest way to do that was to launch something,
• 42:15
  Speaker 1: for sure.
• 42:15
  Speaker 5: And this I think is actually really and this speaks to this idea of like, well, how do you incentivize maintenance work? It is very difficult to demonstrate the impact of doing some maintenance. Right? So there's this great novel by Bulgakov called The Master and Margarita and long story short, very, very interesting novel. I highly recommend reading it.
• 42:37
  The sort of synopsis of the story is the devil comes to Soviet Moscow and has this grand ball in the 19 twenties. And the first order of business is to acquire an apartment, and he does that by basically killing somebody. And the
• 42:49
  Speaker 6: way that this guy dies is this
• 42:50
  Speaker 5: woman spills some oil on some trolley tracks. This guy is walking out of a park after almost convincing Satan that he doesn't exist. He slips on the oil and he gets decapitated by the trolley. Alright. Now, like, why do I think this is relevant?
• 43:04
  Because if somebody could have just like bumped into this person and prevented the oil from spilling, then the guy wouldn't have died and then Satan couldn't have had his grand ball in the novel, and then the world would have been deprived of this literary masterwork. The point being that you might do some you might make some very small changes during the course of one's day to day work, get that committed, and then that's all that prevents a major outage 6 months or a year or 5 years down the road. You really have no way of knowing that. Yeah.
• 43:31
  Speaker 1: Let alone quantifying it. And this is the master and market arena. Is that right? Yes. Yeah.
• 43:38
  Interesting. Yeah. Right.
• 43:39
  Speaker 4: I mean quite with
• 43:40
  Speaker 5: the cat. Highly recommended.
• 43:42
  Speaker 1: Well and I mean, just like actually, I mean, the the value of, like, having an impact on a student. It's also you don't know kind of years from now. And so yeah. So, Dan, how do you how do you square that? Because I mean, we obviously we know that so much of that work has an enormous impact, and we want to encourage it.
• 44:02
  I mean, how do we, how do you do it?
• 44:06
  Speaker 5: I think that you have to make maintenance work an explicit thing that people like in in organizations where they're the perf style evaluation, you have to make maintenance work an explicit thing that is rewarded.
• 44:19
  Speaker 1: Yeah.
• 44:20
  Speaker 5: You know, and and you have to be willing to say, like, we will promote somebody for doing maintenance. And if you can demonstrate that you have somehow improved the code quality of, you know, a library or whatever, you're gonna get a promotion out of that.
• 44:33
  Speaker 1: Which can be problematic. I mean, I the certainly, it was a really so at Sun, there was this distinguished engineer rank that was this basically country club that voted on its own membership. It was kinda outside of the normal promotion process, and it was a huge and contentious deal when they the first field engineer became a DE, and it brought out the absolute positive worst in people. And for someone who is, like, indisputably technically qualified, but spent their life customer facing, which a lot of us would view as, like, very, very important. But it was really there are a lot of people that don't believe that that's important.
• 45:15
  Speaker 5: And that's a shame.
• 45:17
  Speaker 1: Yeah. It is.
• 45:18
  Speaker 2: Well, I think that comes down to the culture that you're in and what you reward. So, like, I I love your example, Dan. I think it's even more important that you do it and show people and talk about it even that you write it down a priori. Right? Like, you can write down all you want, but if people don't get promoted for these things, then, you know, that it sort of doesn't matter.
• 45:34
  But if you're Yep. If you're promoting folks or rewarding folks or recognizing folks and those are the accomplished, then then people drift to them. I mean, there there's another side of it where, you know, I I've had, you know, at at a previous role, I had someone working on our team, come to me and say, well, what kinds of things can I need to do for the next round? Because it says that, you know, I, you know, I need to basically what we were looking for in this organization is folks who could help us see around corners and help us like, you know, be broader. Like we needed just more people being more thoughtful rather than just doing what they were told.
• 46:14
  And, you know, this guy came and asked, how could I do
• 46:17
  Speaker 1: that thing? So, listen, I'm just coming to you to understand. I'm asking you to tell me how I should see around corners.
• 46:24
  Speaker 2: So sort of a tough thing, and and he was doing great work, but it did sort of top him out in the organization. There was sort of a point at which we need we needed the independent thought and in independent operation and leadership.
• 46:36
  Speaker 1: Yeah. Interesting. But it was, like, you also, like, you don't necessarily want everybody seeing around corners.
• 46:40
  Speaker 8: I mean,
• 46:41
  Speaker 1: you don't wanna, like you know what I mean? Sometimes you can overly enshrine some that kind of strategic thinking.
• 46:48
  Speaker 2: Absolutely.
• 46:49
  Speaker 1: And then discourage someone who's like, well, actually, no. I'm not gonna go, you know, help this customer out or go to bug this problem.
• 46:54
  Speaker 5: That's true.
• 46:55
  Speaker 1: I wanna see around corners.
• 46:56
  Speaker 2: That that's right. I mean, this sort of had to do with where did we feel short. You know, that we tried to align our incentives and
• 47:02
  Speaker 5: Yeah.
• 47:02
  Speaker 2: Based on where we felt shorthanded.
• 47:04
  Speaker 5: So here here's a concrete example of what I was referring to. And I'm gonna speak to this in in generalities because I don't feel super comfortable talking about the specifics, but there was a team at Google that had a big maintenance problem. They were working with an upstream open source component and, you know, with a high rate of change and pulling in those changes and merging Google specific changes back, well, not specific, but just Google author changes back upstream was a big problem. And one night I was happened to be eating dinner with a couple of the senior engineers from that team, and I said, look, why don't you take like 4 people and say you have a year to figure this out, get as many of our internal patches as possible upstreamed and figure out a more streamlined process. And at the end of the year, you'll all you will all get promoted.
• 47:52
  And it was like, I had just suggested, you know, kicking a puppy or something. I mean, you know, like the other folks around the table just kinda looked at me like like I had said, you know, something very offensive. It was and and it was clear that wasn't gonna fly, but to me, it seemed like a pretty decent idea.
• 48:08
  Speaker 1: Interesting. And you got like, hey. Look. Everyone knows you got this prop this problem is, like, well known. Everyone knows this is an important problem.
• 48:14
  I'm just proposing a way for you to use the extent incentive incentivization structures to solve the problem. Like, it feels very reasonable.
• 48:23
  Speaker 8: I I mean, as another former Googler. Right? Like, there there's periodically these tweets that go around that are like, hey, senior folks in in engineering. Like, talk about your times where you did something that that caused a major problem so that the new folks, when they inevitably do something that takes down production. It's not, you know, completely up foreign.
• 48:46
  And I usually describe a case where I made a typo in a regex, and it got deployed to all of Google and ended up taking out 2 whole clusters for multiple days.
• 48:55
  Speaker 1: And clusters in the sense are not small.
• 48:59
  Speaker 8: No. No. We're talking order, like, tens of thousands of machines.
• 49:03
  Speaker 2: Good job, man.
• 49:05
  Speaker 8: And but the thing is is, like, coming back to the impact side, while I while I offer that as an example of, like, where I took down production, and it wasn't a big deal. The other side was it actually served as a way as, like, this way of demonstrating how much impact there was behind the software I worked on such that the postmortem of why did this happen and how could we prevent it in the future actually pro finally provided a path for folks on my team to be promoted.
• 49:31
  Speaker 1: Yeah. Interesting. Yeah. Interesting. Because you were able to provide some visibility into, like, hey.
• 49:35
  This is why this stuff is really important. And presumably, there were a a bunch of the things, and I'm Rick, I imagine this would came out in the postmortem. Some of the I would presume some of the infrastructure that you could deploy to this kind of thing happening is probably something that people wanted to build and felt like they couldn't because they didn't have enough visibility.
• 49:55
  Speaker 8: Yeah. It I mean, ultimately, for that specific one, it gave enough air cover to go out and figure out what a proper plan would be to make this more sustainable. And it turned out to be a a 7 year long plan to eliminate a core piece of the the Google, server software stack and and infrastructure stack and replace it with a a an alternative. And, you know, like, there are people who got their promotions off of a project that I started that I never met.
• 50:24
  Speaker 1: Okay. So, Ewig, the plan okay. 7 year long plan. I've got some follow-up questions. Was that a 7 year long plan at the outset?
• 50:32
  No.
• 50:33
  Speaker 8: No. It's just the realities of the complexity of it. It was a core piece of the infrastructure that nobody quite understood how it was interacted with from everywhere, and so everyone was afraid of changing it. So you you end up with this in, like, you know, all sorts of organizations. There's there's also that xkcd comic of, like, here's the giant tech stack, and here's the one component that's written by this person in Nebraska.
• 50:55
  Right? Like, it's the same thing, except that you don't necessarily know where that component is or who else relying on it. So making those changes, like, it's it's hard to understand what the impact is, but you also run into fear around making changes to things that you suspect have a lot of impact. So you you you know, there there becomes this we don't wanna disrupt things too much. So working on a potentially impactful project can also be held back from fear of causing too much breakage.
• 51:26
  Speaker 1: Totally. Well, it's so but, Rich, there's another interesting thing there that is you know, not to bag on Google too much because it's all the fact that Google had a culture, if not, like, the the metrics around doing the post mortem, a la I mean, it because there there will be other companies that would not do a deep understanding of why that failure happened, and they would just punish those involved and move on. But to Google's credit, like, it sounds like really wanna understand why this failure happened. Dig deep. Get a get a full rigorous postmortem.
• 51:59
  Because I do feel like that's another way to kind of enshrine the importance of some of this unseen infrastructure is to make sure that failures themselves are really important to fully understand.
• 52:10
  Speaker 8: You make it sound glamorous. The the reality of the situation was was definitely much more having a culture of writing postmortems meant that I I could actually stand up and say, we should write a postmortem for this. And that Yes. I still had to take on that like, recognize that that was important and take on the role of actually writing it. And writing down things that I knew were gonna be inflammatory to a lot of folks and sending it out to fairly senior folks so that they saw the problems that were actually there.
• 52:41
  So Yeah. It it was partly that the culture gave enough ways to to that you could do it, but you were still taking somewhat of a risk of of actually going through with the process. And, I mean, that's not like, we're getting a little off topic, but it was, like, there are a lot of incentives and and disincentives throughout the company culture, and you have to be able to pick and choose, like, which incentives you're gonna follow, and which disincentives you're gonna ignore to be able to actually accomplish something and be able to use that as demonstration for, you know, longer term incentives. It it just it becomes, its own, like, muddled mess of trying to navigate the incentives.
• 53:33
  Speaker 1: Yeah. That's interesting. And then what's just like, also, like, what are the incentives to change the way things are incentivized? Because I do feel that you also have this very much the the this kind of, this confirmation bias for those that do well in a particular system. Reward that so that they wanna that they they want to keep that I mean, Satnam, to go to your academic point, those that came up through the system of tenure are gonna be loathed to give that kind of job security out without people going to the same things that they went through.
• 54:07
  And so how do you gotta incentivize people to improve the system? I think it's another one that's that's pretty thorny. Mhmm. I'm not sure if you could answer that one.
• 54:19
  Speaker 2: You've probably seen this, Brian, in, like, executive compensation where, you know, boards of directors just wanna kind of all these executives pay go up and up and up. And this is why one of the we had the ratio between CEO and the average employee has has increased dramatically. But some of it is just believe believing in the system of incentive and going around the table and and rewarding each other.
• 54:43
  Speaker 1: It does seem to be that I mean and and you have these things that have got nothing to do with an executive's performance that that form a huge amount of their compensation. And, yeah, it it does seem to me that you are, the some of these comp committees, I do not understand why you got the the kind of the way they operate. And I guess it's the it's it's the belief that the executives are creating heaven and earth underneath them, which is not always the case. Absolutely.
• 55:12
  Speaker 2: But it's the it's other executives making those kinds of decisions. So it's kind of buying into that cargo call.
• 55:19
  Speaker 1: Totally. Totally. Christopher, you're trying
• 55:22
  Speaker 2: to get in here.
• 55:23
  Speaker 1: Hey there.
• 55:24
  Speaker 6: So, I I've had the, academic experience for about 20 years, into my late forties, and just about incentivizing young people. We took on a lot of summer students and summer interns and co op interns. The key there was getting the team of experienced people together, writing out on the whiteboard all the nagging little things we wanted to do and sorting them into sort of short 3 week easy to do topics and then, longer 3 month ish, little more challenging topics and and divvying them up so that incoming youngsters would have what we call the quick win. That was the 3 week project that could be done, after a week of intensive training and and, some a bit of handholding and then putting them on that 3 month longer project. And, the the point of that is incentivization is that the win itself is the incentive, and to give them that early win was key.
• 56:43
  Now I'm, in my late fifties now, and, I left academia. I I put, I I put 12 PhDs out on the market, and that's way more than my, optimal reproduction number. So and and half women, half men. So That's great. Hey.
• 57:03
  That's there you go. So now I'm working, incidentally with a bunch of ex Bell Labs guys at Ericsson, and, I'm no longer I no longer have to play the smartest guy in the room
• 57:18
  Speaker 1: because I'm not. It's very liberating to not be the smartest person in the room, I found. Speaking of the person who is who is emphatically not is very yeah. It's like and it would that's okay. So that's another interesting I mean, you you raised a bunch of interesting points in there.
• 57:34
  I I do feel that there's a danger when people are vying to be the smartest person in the room. That can be a that can that particular incentive can get very toxic very quickly, where you are now, like, I am incentivized to now actually, like, shit on your idea so I can make sure that you're not the smartest person in this particular room.
• 57:57
  Speaker 6: And and for the one of the clues of why I left academia was, I was working at the National University of Singapore for 6 years. They capped my salary payable to software guys, to a point where I couldn't recruit. And so I could I could get all this grant money, no problem, but they wouldn't let me spend it, at a level to recruit the the kind of people I needed. So I had to set up my operations such that I had to work exclusively with trainees, and I had to adapt to that. That was that was not, as I said, I got tired of being the smartest guy in the room after 6 years of that.
• 58:42
  Speaker 1: Well and so the other thing you mentioned that I thought was interesting was the I I mean, certainly, the the finding the the quick win, which I definitely think is important. I think that you guys gotta find that, like, that quick win, and then you need, like, what's the medium sized win behind it? So you're not going from the quick win to, like, okay, great. Now jump in the deep end and and struggle. You gotta kind of find that.
• 59:04
  And but I always believe that you should give interns the the projects that are really exciting that no one has the time to really go into. I I always felt that was a lot more interesting than giving them which sounds like you're kind of those 3 month projects that are more speculative in nature. Yeah. Yeah. Exactly.
• 59:25
  Which is great because I think it it it it generates a much more I don't know, Adam. I mean, you did
• 59:29
  Speaker 2: For sure. I mean, and I I did the same thing, after
• 59:32
  Speaker 4: it had been done
• 59:33
  Speaker 2: to me. And and I agree. I think it's more interesting time, and you have everyone excited about it. Right?
• 59:39
  Speaker 1: You sit down to the
• 59:40
  Speaker 2: lunch table. Table, you tell people what you're working on, and a lot of people have spent their spare time thinking about it so they can offer useful mentorship as opposed to slogging your way through some some early bug. It is interesting thinking about the psychology both of ramping people up and and incentives. There's got to be thinking, Brian. I don't claim to know many psychological experiments, but my favorite is they give all these, all these monkeys, cucumbers, totally happy with the cucumbers, and then they give a bunch of other monkeys, bananas.
• 1:00:14
  And the monkeys with cucumbers take the cucumbers and throw them at the to the researcher. And I feel like we get a a fair bit of this with with, like, incentives and raises. Right? The the raise Totally. Like, the the job I had was fine.
• 1:00:27
  And now you gave a banana to this numbskull? Like Absolutely. I think you're a cucumber.
• 1:00:34
  Speaker 4: You guys say you can take
• 1:00:35
  Speaker 1: your cucumber and shove it. That's right. Incoming cucumber. No. I definitely feel that's a yeah.
• 1:00:40
  What a great first of all, I was studying most of your retelling from Am I the monkey in this metaphor? Am I the cucumber? Who am I? Where am I? I know that I know that I'm here.
• 1:00:47
  I just don't know. But, no, I totally agree with you. And you can you can get people to care about things that they actually wouldn't otherwise care about because you've you've you've introduced scarcity. And, you know, it's like, I you know, the no. This group over here has this and you don't.
• 1:01:05
  You kinda create all these things that all of a sudden people start behaving in ways that you don't necessarily want them to behave. It's not actually conducive way. It's not actually what you're trying to do.
• 1:01:15
  Speaker 2: I'm sorry. And same same thing with baselines. Right? Like, we how many, you know, prospective colleagues have we talked to who were too hooked on their mega corp salary to get above? And and a lot of cases, they don't necessarily need that.
• 1:01:27
  And and and, you know, need is a weird word to use with regard to money, but it it means, like, it's not necessarily built into their lifestyle. They're not you don't need to pay to support their family or whatever. But it feels like taking something away, and I guess it is taking something away. But, you know, that same job offer before they had had that experience, would have been much more compelling. So, like, the the way we sort of bank what we've already got as the new baseline, both when it comes to, you know, compensation and title and all of these things.
• 1:01:59
  It becomes very hard to to step backwards.
• 1:02:02
  Speaker 1: Yeah. I mean and I, you know, I think other people should do this as well. One of the things that we ask people when they applied Oxide is when were you happiest and why. And, it's never based on, you know, a promotion or compensation. When we ask people when you were happiest and why, it is always based on, you know, I I solved this problem that's other group needed or that I did this thing that was that had intrinsic satisfaction or brought satisfaction to others.
• 1:02:27
  It's it's gotten nothing to do with with the kind of the currency of the realm or oh, I mean, I should say I mean, every once in a while, you do get people who are surprisingly reviewing in that regard that you guys, like, probably not a fit. But it is actually an important that we should all kinda figure out when was I happiest and why? When was I unhappiest and why? Where am I where are my true incentives? And I think figuring that out and then figuring that out for other people and kind of trying to align all that with what we wanna actually go do while not creating, you know, bugs that get created spuriously so we can close the or products that we create spuriously so we can kill over what have you.
• 1:03:08
  Right. That's the trick. Adam, I know we wanna keep we're we're trying to to end on time these days. How are we doing? We're doing great.
• 1:03:17
  Great. Okay. Good. It's been a great discussion. It's been interesting.
• 1:03:22
  Yeah. And I think that, you know, we can, agree that there are are things that we wanna avoid. Are you we didn't mention the the the COBRA's in India. I would love to know if that's a you've heard of that. The the it did the British in India wanting to eradicate COBRA's offering a bounty for cobras, and then people started breeding cobras.
• 1:03:48
  Speaker 3: Of course. I didn't
• 1:03:49
  Speaker 2: know that. That's such a great one.
• 1:03:51
  Speaker 1: It's a great one. I just need to I would I I would love to get, like, act because I I it's kind of this, like, Malcolm Gladwell esque knowledge. You know what I mean? We're just, like, god, that feels so great. Like, is that true?
• 1:04:03
  I hope that
• 1:04:04
  Speaker 2: That's what I was gonna say. I was say gonna say I choose to believe that it's true. I think
• 1:04:07
  Speaker 1: it is true.
• 1:04:08
  Speaker 2: Because I definitely wanna drop that, in lots of occasions.
• 1:04:11
  Speaker 1: Yeah. I mean, it because it's and I think there's a lot of COBRA breeding that happens, and we wanna avoid the COBRA breeding when we incentivize engineers.
• 1:04:19
  Speaker 2: Yeah. Man, it gets hard. It gets hard, especially as teams grow and, you know, you lose touch with what is motivating people and, and who's not being praised sufficiently and who's being praised more than the rest of the company thinks whatever it is. Just I don't know if there's a silver bullet.
• 1:04:38
  Speaker 1: God. Let's just hope that future generations of Oxide engineers are coming back to this recording and be like, when did Demo Day get so fucked up
• 1:04:44
  Speaker 2: around here?
• 1:04:45
  Speaker 5: Like, I was like
• 1:04:46
  Speaker 2: I think these cucumbers, buddy.
• 1:04:49
  Speaker 1: Exactly. That's right. Alright. Alright. I think so we do what I know I said this last week, but I think, we we are gonna come up to when when they actually schedule on European time.
• 1:04:59
  So I think we may we'll get our actual act together this coming week, Adam, to get that done.
• 1:05:03
  Speaker 2: Is that like a commitment? Like, a week from today, European hours, go.
• 1:05:08
  Speaker 1: Okay. I can see where this is going, and I agree. I have dropped the ball here. Okay. I get it.
• 1:05:14
  Yes. I, let's see. Let's huddle. Let's figure it out. I wonder Okay.
• 1:05:18
  Because you I would I want the open source the firmware folks to be able to join. So I need to be checked with them. So
• 1:05:24
  Speaker 2: Stay tuned. Watch this space.
• 1:05:25
  Speaker 4: Stay tuned. Watch this space.
• 1:05:27
  Speaker 1: Alright. Thanks everybody. Talk to you next time.
• 1:05:30
  Speaker 2: Thank you.