OpenAI & Meta Distinguished Engineer (IC9) On Working With Zuck, Carmack & Career Growth | Philip Su

Philip Su grew to Distinguished Engineer (IC9) at Meta and OpenAI. He has a bunch of interesting stories about working with people like Zuck and John Carmack as well as a ton of advice for software engineers. I was really looking forward to chatting with him and enjoyed this conversation a lot. I hope you find it helpful! In this conversation, we discussed: • What Distinguished Eng (IC9) expectations look like • How he got promoted to IC9 • Working with impressive engineers like John Carmack • What made Zuck and Boz special as coworkers • Learnings from switching between IC and EM 6x • Why he joined OpenAI • Advice for his younger self Timestamps: (00:00) Intro (01:02) Growing to Senior Staff (IC7) at Microsoft (06:38) Management vs IC transitions (17:32) Demotion from IC9 to IC7 at Meta (20:28) IC7, IC8, and IC9 expectations (28:58) IC9 promo story (31:30) Building a strong eng team culture (36:16) Working with Zuck + Meta CTO (38:57) Working with John Carmack and other impressive ICs (41:44) Buying $23000 of coffee in a day (45:35) Why leave Facebook (49:25) Joining OpenAI (55:38) Writing well as a software engineer (1:03:00) Does software eng performance decline as you age? (1:07:00) Building credibility as a young manager (1:10:25) Should you be a generalist or a specialist? (1:12:43) Advice for his younger self Where to find Philip: • LinkedIn: https://www.linkedin.com/in/suphilip/ • Peak Salvation podcast he referenced: https://peaksalvation.com/ Where to find Ryan: • Newsletter: https://www.developing.dev/ • X: https://x.com/ryanlpeterman • LinkedIn: https://www.linkedin.com/in/ryanlpeterman/ • Threads: https://www.threads.net/@ryanlpeterman • Instagram: https://www.instagram.com/ryanlpeterman To hear more, visit www.developing.dev https://www.developing.dev?utm_medium=podcast&utm_campaign=show-notes-spotify-open-access

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Speaker 00

  One thing that impressed me most about Zuck, and this is true of Boz as well, is

• 0:05Ryan Peterman?

  This is Philip Su. He's one of the few who have been promoted to distinguished engineer or IC9 at Meta, which is three levels higher than staff. And I asked him about everything he learned along the way. Someone that stands out to you, that consistently impressed you?

• 0:21Speaker 00

  John Carmack, obviously legendary, right? Not only is he super prolific in coding, he had an ability to

• 0:29Ryan Peterman?

  Oh, wow. He also shared about his unique demotion.

• 0:33Speaker 00

  I was eventually re-levelled down from an E9 to an E7.

• 0:37Ryan Peterman?

  Later, he left Meta to join OpenAI before it got big, and shared an interesting perspective on why.

• 0:44Speaker 00

  The reason I joined OpenAI was what I'd learned from working at Facebook is I would much rather join the market leader or nobody at all. And here's why.

• 0:53Ryan Peterman?

  He's one of the guests I was most excited to have on, and hopefully you see why after listening to the conversation. Here is the full episode.

• 1:02Ryan Peterman?

  Thank you, Philip, for joining today. I'm so excited to ask you all the questions I have here. I think a lot of people are gonna give value out of it. So yeah, let's start digging into your career, starting with Microsoft. My understanding of your career is that you sprinted through Microsoft's ladder. What do you credit your fast promotions to the equivalent of Microsoft's E7 to?

• 1:29Speaker 00

  I think the growth comes from a few things. One is just I had several very good managers and teammates who I grew a lot from. So for instance, when I joined the Tablet PC team, the team was about 20 people, but there were three distinguished engineers on that team, which for Microsoft terms is like very rare to find a team like that stacked with talent. So I learned a ton of stuff very quickly from people on that team, for instance. I think another thing is that I just work very long hours is the truth of it. Like in my first year at Microsoft, I had a sleeping bag in my office. I regularly slept at the office. I would have an alarm that would wake me up at 3am because I was pursuing Da Vinci's

• 2:11Speaker 00

  theory that you can sleep four hours, wake up and then like sleep another four hours and wake up. So I would wake up at 3am, code a bit and then go back to sleep at 6am and then wake up at 10am and keep coding, right? So I do think, a thing that I tell people a lot is I think there are only three main things that contribute to fast career growth. One is of course luck, you know, being at the right place at the right time. You happen to be at Stanford when Sergey and Larry were working there. They happened to pick you to join the company, right? So some of it is luck. Some of it is talent, you know? Like people like to say things like, oh, everyone can be an astronaut and I don't think that that's true. Like the harsh truth is, you know, I am never going to be a great basketball player, right? Like they're just things that I cannot do.

• 2:56Speaker 00

  Similarly, someone who is not naturally good at public speaking, they can always improve to where they're passable at public speaking, but they're not going to become a Cicero through like a lot of practice, right? So I think luck, talent, and the last thing I do think is hard work, you know? If you are willing to outwork everybody, I think who is equally talented and equally lucky, you're just going to get further. I am not saying that I would recommend that to everyone and in fact, I do think that my work life was way out of whack for many years there. So I'm not at all saying that this is what I recommend people do, but I do think that outworking a lot of people goes a long way.

• 3:37Ryan Peterman?

  When you first got into the industry, it sounds like you were really hungry and I was also, you know, super motivated too. And one thing that I wonder though is, you know, when you work so hard, it trades off with your health and then can eat into your, I guess, productivity throughput. After you lived that experience, do you think if your goal was growth at all costs, that that was actually optimal?

• 4:03Speaker 00

  Yeah, this is a great question and I think there are a lot of people that will defend the 40 hour work week with things like, you know, when Henry Ford studied his people, every additional hour was like lower in productivity. I think one thing people don't like to discuss often is that I do believe your productivity does go down on a per hour basis as you stay up for way too long. But I think for many people, that curve keeps going a long time before it gets negative productivity. Meaning like, an additional hour worked is actually subtracting from your productivity. So I feel like, the truth is, working fifty, sixty hours, you will probably get a sum total of more done, but each incremental hour is probably less effective, right?

• 4:46Speaker 00

  But I think there's one other thing that people don't take into account, which is, you know, you could say like, for instance, why do surgeons still have residencies that have thirty-six hour shifts? Right? Part of it is hazing, I think. Part of it is just because they did it when they were young, so they're gonna make you do it when you're young, right? But I think another part of it that I heard from a surgeon is that the biggest, quickest way to gain experience in surgery is to have a lot of surgery hours. So the more you're there, the more you experience. And if you can pack in six years' worth of experience into four years, then you're just going to be a better surgeon coming out, right? And so I think although nobody wants to be operated on by a surgeon who's in his thirty-fourth hour, let's say, right?

• 5:29Speaker 00

  At the same time, you would like a surgeon who has experienced a very wide variety of things. So I think although I wouldn't recommend my crazy work hours from when I started work, I would say it exposed me to a lot more experience than the average junior person starting at Microsoft.

• 5:46Ryan Peterman?

  So there's diminishing returns, but there are still returns and... So if you're a growth-at-all-costs career person, you should, I mean, by that logic, you should work more. Of course, there's other parts of life.

• 5:59Speaker 00

  Totally. Like, if, if you were diehard convinced that career was the most important thing for you, and in fact, that you were willing to sacrifice things like relationships and other things like that, I do think the truth is, working longer hours is going to get you there faster.

• 6:15Ryan Peterman?

  So you grew to the equivalent of, you know, E7. My understanding, that's senior staff. And you said there were three distinguished engineers. Is that the equivalent of E9? Or would have been at Microsoft?

• 6:28Speaker 00

  Yeah, probably something like that, E9. It's the equivalent of a vice president level in compensation within the company.

• 6:37Ryan Peterman?

  E6 to E7, I'm familiar with that at Meta. When you got that promo at Microsoft, was that as a manager or as an IC? And what's the story behind that?

• 6:48Speaker 00

  Great question. That was as a manager. But I think I also then became an individual contributor at the new level as well.

• 6:57Ryan Peterman?

  I see. What made you decide to switch into management in the first place?

• 7:05Speaker 00

  I think when I was younger, I was very ambitious for myself and I came from a culture, you know, I'm ethnically Chinese. I came from a culture where people really felt like managers were more respected and would get further in career. And so I think I had the somewhat misguided notion that in software engineering, becoming a manager was the next growth step for me. So that's largely why I became a manager. But in my career, I fluctuated between managing people and being an individual contributor at least six times. And I think the main reason for that is I really love being an individual contributor. But I'm often asked to manage if the team needs a temporary manager for a while, for instance. Or I would much rather be the manager of a team if it's poorly managed.

• 7:50Speaker 00

  You know what I mean? So my order of preference would be be an individual contributor, be a manager on a team because the team doesn't have a better manager, right? Like, that's my order of preference. But in general, I think left to my own devices, I'd rather not manage people.

• 8:06Ryan Peterman?

  I see. And in that case where you said if it's poorly managed, are you saying that there's cases in your career where the existing management was lacking and you volunteered yourself? Or you're saying someone in your management chain said, Philip, can you please help us?

• 8:25Speaker 00

  Yeah, both of those things have happened before. So for instance, in my career, I've had four separate times when a previous manager of mine asked to report to me. So I became their manager, basically.

• 8:38Ryan Peterman?

  Really?

• 8:38Speaker 00

  And I think that, yeah, and so that has happened four times. And I think part of that is because I've learned from great managers, you know, although I don't prefer to be a manager. Manager, I do think there are parts of what I've learned that have been useful. Right. And so I have been asked to be a manager before. After pursuing, trying to become a manager early in my career, I sort of stopped looking out for those opportunities as actively as I did an individual contributor opportunity.

• 9:09Ryan Peterman?

  That I've never heard of that before, where your manager asks to report to you. How does that happen?

• 9:16Speaker 00

  Well, I think each situation is a little different. So on some of the situations, they reported to me on the same team. Right. On some situations, several years later, they asked to join my team. You know, and when we had both moved on. So I think that in the cases where it happened on the same team, one time it was because the manager wanted to be an individual contributor on the team. You know, and I was a strong individual contributor that wanted to be a manager. And so it was almost like a swap in place. Right. I've had other times, several times where I joined a different team somewhere else. And then my manager who liked working with me was very happy to work for me as well on the new team. And so that's been the majority of cases.

• 9:59Ryan Peterman?

  At some point you get in your career, there's that decision point of, do you want to continue down the IC track? Or do you want to become a manager? What would be your, your framework of thinking through that for, for someone who's considering that for themselves?

• 10:14Speaker 00

  Yeah, great question. I, personality wise, I'm very open to new experience. So in general, I'm going to advocate for trying new things. You know, so I feel like in general, it can't hurt to try it if you think that you might like it, or you think you might be good at it. But I would put a few asterisks on. One is like, be very sensitive to whether or not it locks you into a career you don't want. So for instance, I've told you that I've switched back to an IC probably six times in my career. Right. Many people cannot make that transition. Meaning they have become a manager, now they're a manager of managers, they're far away from the code. They either are unwilling to take a lower level job in order to be an IC,

• 10:58Speaker 00

  which I'm very willing to do, right? Or they are unable to meet the expectations of their level as a manager, as an IC. And so I think people have to be very careful not to accidentally walk through a one-way door on that. For me, I've kept that door open two ways because I do the switch often. And when I'm managing teams, I also love the actual coding. So I still try to dive into the code whenever I can. And so I've kept both sides sort of evergreen, but I don't think that's true of many people. So like one is be careful that you don't walk through a one-way door that you're not positive, that you want to walk through, right? I think another thing about that is people often mostly think about themselves when they think about wanting to be a manager.

• 11:40Speaker 00

  But what I classically tell people is you are ready to lead a team when your team would have elected you to lead them, right? Like, think of it this way, would you want to work for you? That's the key question. Would you want to work for you? If the answer is not a solid yes, I would be thrilled to work for me, right? You probably have a few things you can improve in yourself to get to that point. And people cannot make you the lead of anything. Like, they can nominally make you the lead, but as you know, in software development, people can slow roll things, people can like, can sandbag things. Like, there are all sorts of ways your team can work against you, actually. So you need people to want to work for you.

• 12:22Speaker 00

  So I think before becoming a manager, one key thing is take a critical look at, like, would I work for myself? And like, if not, what can I improve in myself to make it so that people want to work for me? I think when you have those improvements, it'll be clear. And then people will ask you to manage at that point.

• 12:40Ryan Peterman?

  That was interesting you talk about the two-way doors. Because when I think about it, I made a similar decision for myself of, you know, pursuing management versus IC.

• 12:51Ryan Peterman?

  Presumed past a certain point, you've become more specialized. Like if I was a manager of managers and then I switched to my equivalent. Let's say I'm a director and now I got this big org underneath me. And I switch into the equivalent, which would be principal engineer. Those jobs are so different that I would have thought almost no one would be able to keep the door open. And I was thinking if I did do switches too, I would be, I guess, resetting some progress potentially. So yeah, I'm curious how you think about that. Like, did you ever feel like you lost progress by switching between the two tracks?

• 13:33Speaker 00

  Yeah, well, I definitely lost progress and intentionally and willingly so. So for instance, when I switched from being the site director of Facebook London to being an individual contributor on Oculus, I was eventually re-levelled down from an E9 to an E7. And that was something that I very much sort of wanted and I fully acknowledged was necessary because the skills I could demonstrate as a, as a site director were very different from the skills that were required to be a good coder on Oculus. And so that was a re-level where I very much bought into it. Now, many people cannot handle that, like either for ego reasons, like they feel like they can't handle it

• 14:15Speaker 00

  or for compensation reasons, like there are some people whose lifestyles just grow to fill whatever compensation they're making, right? And if they refuse to change their lifestyle, you know, they will paint themselves into a corner where there are increasingly less jobs they can take because they are unwilling to move. So I do think sometimes that switch entails that movement. And I also think there are certain times in career where that switch is easier and certain times where that is harder. So like, here's an interesting dynamic. I think early in career, like one or two years in, let's say you are a software developer. You want to be a product manager. I feel like one or two years in, it's easy to make a lateral move. You know, you're like a junior software developer. You want to become a junior product manager.

• 14:58Speaker 00

  Many people are very willing to coach you on that. I think in the senior, you know, E sort of five-ish range, I feel like the transition is very difficult. If you want to go from an E5 software developer to an E5 product manager, just like that, I think it will be very difficult to be good relative to your peers, unless your software development work involved a lot of product management-y sort of things, right? But funny enough, these things are a diamond shape. Once you're senior enough, the skills once again converge. Like, what is a really effective E8 doing, right? There are occasionally, of course, the prodigy coders, who are just far more effective than the average person in producing a ton of hard code, right?

• 15:43Speaker 00

  But I think a lot of very senior individual contributors, they are actually great at leading teams. You know, like, they have a vision for technically where the product needs to go. They understand the infrastructure and the weak point. They have a plan for staged migration to the next better design, right? And people love working with them because they are good and compelling at coaching others and whatnot. That's a lot of what makes a great typical sort of E8 individual contributor, in which case like this diamond shape happens, right? Which is, once you're senior enough, a lot of the skills once again overlap. Are you good at communicating with others? Do you have vision? Do you understand the business strategy and how it pertains to the technology you're building? Those things, once again, become shared.

• 16:25Speaker 00

  So I think there's a part in the middle where it's most difficult to make the change.

• 16:29Ryan Peterman?

  Yeah, that is a, that's a really interesting idea. And I, I see that as well with the high-level archetype I sees, especially the, you know, tech lead, generalist, the skill set in leading an initiative that AD engineers are contributing to. There's probably a lot of leadership skill overlaps with the director, so I see that. But I wonder, though, for some of the other archetypes, like, let's say you went deep into the, you're a specialist in AI or something. I'm wondering, maybe that diamond doesn't come back.

• 17:06Speaker 00

  Yeah, for some archetypes, it definitely doesn't come back, you know. There are some people that are super deep, like, let's say you're into like query optimization for SQL queries, right? You can go very, very deep on that. And I think in a way, some archetypes going that deep on the technology might not need to have developed all the other skills that are related to leading teams and whatnot. It may not be a good transition into managing.

• 17:32Ryan Peterman?

  You mentioned that you navigated the conversation, you were in E9, and you sought out a demotion to E7. And for those who don't know, E9 is a distinguished engineer and E7 would be a senior staff engineer, which are both crazy high levels. How did you navigate that conversation, like requesting a demotion?

• 17:56Speaker 00

  Yeah, yeah, this is a great question. I was even concerned about it when I was first promoted from an E8 to an E9 as the Site Director of London. What I told Facebook's CTO back then, Shrep, was, you know, I'm very concerned that because I like being an individual contributor, I'm very concerned that I cannot maintain this level of performance when I want to move back. And what he said to me during my promotion to E9 was, hey, let's worry about that problem when it happens. Like, let's not not promote you because you think one day you're going to want to go back to coding. So when I went back to coding, I very much understood that I was at a level that I in no way could meet the expectations.

• 18:38Speaker 00

  And so I think under those conditions, it felt honestly great for me to ask to be moved back to E7, which is the level I joined the company at like eight years before that, you know? Because I knew that I could perform at that level. And I definitely did not want to be compensated more than I deserved relative to my peers on a team. It feels so much better to be performing well at a level than it feels to be like barely treading water at a level that you feel a lot of imposter syndrome in and that you're not even sure that you'll stay long term qualified for, right? So I actually very much welcomed the change. But I think also, many people don't make a change that big, first of all.

• 19:22Speaker 00

  And second of all, many people don't want, don't want to. Take all that comes with that, right? Like, for instance, I went from being the site director of a site that was probably, you know, four or 500 people, down to reporting to a manager who reported to a manager who was then part of the management of the site, right? And so, like, you have to have an ego that is very willing to say like, "Hey, I'm happy to be just a member of a team down here, um, doing my own work." And you also have to be sensitive to other funny things like, you have to be careful not to speak up about your opinions about all sorts of things that you used to have opinions about when it came to running the site, because you need to support the new people running the site in being successful. Like, you need to support the new leaders, right?

• 20:07Speaker 00

  By playing your role on the team as opposed to using your outsize influence from the past to get things done. And so I think both of those things were changes that I was, uh, you know, willing to try to learn to navigate, and I very much welcomed being leveled appropriately relative to my contributions.

• 20:25Ryan Peterman?

  Makes sense. And you talked about the, uh, expectations being too high for E9. Can you give a high level? Like, what, what does it generally mean for? What is an E7 typically doing? Or E8 typically doing? Or E9 typically doing?

• 20:41Speaker 00

  Yeah, great question. And I have to caveat my opinion about this relative to, you know, I was working at Meta starting 15 years ago, right? Um, and so my data about how those levels map is going to be very coarse screen, right? But in general, I think here a few important concepts, just when it comes to all those levels, right? Is it is about the

• 21:06Speaker 00

  scope of the person's influence. So like, how big or how long, uh, or how valuable of a project can you single-handedly guarantee the delivery of? Okay, so like, an intern can not even guarantee the delivery of a feature necessarily, right? Uh, or they might guarantee the delivery of a small feature on time, right? A senior developer, like, let's say, an E5, E6 in a company, should be able to deliver the work of about 10 people, 10 to 15 people, right? If you're regularly influencing the work of 10 to 15 people, and you can reliably deliver things of that scope when you are the one in charge for

• 21:47Speaker 00

  that delivery, whether it's technical delivery, or a product feature area, or whatever it is. Usually at E7, once again, my data is probably very stale. I'd expect a person to regularly impact the work of maybe 50 people. You know, so like you would be a person that is driving technical direction for maybe around a team of about 50 people. You are regularly consulted on key changes on that team by the product managers, by the engineering managers, like you are viewed as that influential, right? Typically, you are able to personally deliver projects on the scope of like six months. Meaning like, hey, if I tell you this is your goal and I don't have time to speak to you for the next six months and I send you off into, you know, the wild of coding,

• 22:31Speaker 00

  you should be able to come back in six months with this thing done, right? Like that is the scale and the scope of what I'd expect a person to personally be able to do, right? As you get to eight and nine, those things just multiply out. And I think there's both a quantitative change to it. So meaning like I'd expect a director to regularly influence probably the work of 100 people. You know, like that, that's around the scale that I would say seems deserving, right? On the IC side, as you know, people can be ICs that are large scope in many different ways. They can be large because they're very deep, right? They're narrow, but they're very deep, right? They're deep in a way qualitatively that others, they can do things others can't do.

• 23:15Speaker 00

  So like a great E8 that's deep qualitatively, that is deserving of it, for instance, you cannot simply replace him or her with like four E6s and expect the same work. Like there, there's a level of work at which no amount of adding people to levels below is going to get the quality of work that you need, right? So there are some jobs like that. There are some people that are very broad, so they are a jack of all trades. This type of person is especially useful for nascent teams or startups, right, for small groups, because generalists are super valuable there. And that's also because typically you cannot hire three E6s that basically can outperform one great generalist E8 that can like do a little bit of everything, right?

• 24:00Speaker 00

  So there's a quantization question there. So there's both a quantified difference, meaning scythe of team influence. But I think the qualitative difference is very, very important because Those qualitative differences often cannot be articulated well in a bullet point list that your HR team sends you. And in fact, if you go and try to do that bullet point list, you'll often be disappointed that you're still not at the level and you'll be very confused why not. It's because people at those levels excel for very different reasons and it takes the judgment of a person above those levels. And these are hard judgments, right? To decide whether or not a person is deserving.

• 24:42Speaker 00

  So like, let's say you get promoted from an EA to an E9. When I worked back at the company, right? E9 wasn't even a level when I was first promoted to EA. Like, it didn't exist as a concept. Later, the company, like, it got hugely important people in the company, like John Carmack as a good example, as an individual contributor, right? Like, what would you level him at, right? So, the company eventually had to invent these levels. I think there was a period of time when it was discussed, when I was part of the conversation of the first promotions into E9. And my memory, if my memory serves correctly, the company was probably several thousand people at this point, and the company was considering promoting three engineers into E9, you know?

• 25:27Speaker 00

  And so, when you have that small of a sample set, it is very hard to create a bulleted to-do list of like, "Oh, if you're EA, do these four things and then we will like move you to E9." Every one of those three people were very different in shape. They were different archetypes, right? They were successful for different reasons. And so, I do think that as you get higher up in level, it's almost like how you become a professor somewhere, right? Like, it's not like they have a bullet list that says, "You publish 10 papers over three years in journals of this quality and you're guaranteed tenure," right? At some level, it's all the other professors believe you're a professor. Then you're kind of a professor, right? Same with these leadership level.

• 26:08Speaker 00

  At like, eight and above, in my opinion, a lot of it is, would other E8s think you are an E8? You know, for whatever reasons they think that. Uh, and those reasons are different for people, uh, different archetypes. And they are hard to articulate. And so for anybody who's looking for a punch card to get there, that's why it's hard to describe how to do it.

• 26:30Ryan Peterman?

  With that dynamic, then I kind of wonder if there could be a chicken and egg problem. Um, so imagine you are a excellent E7, but there are no E8s or E9s in your vicinity. Could, you know, could you even get that qualitative input that says, hey, you're solving problems that, you know, E7s can't? Cause you're just at the top of where you are.

• 26:56Speaker 00

  Yeah, totally, especially at small companies. So that happens all the time to people who join small companies or weak teams.

• 27:03Ryan Peterman?

  Right?

• 27:03Speaker 00

  Both things can happen. So let's talk about the latter first. Uh, Microsoft's a huge company. You know, when I worked there, it was 140,000 people. Now, before the layoffs, it was about 240,000. At 140,000 people, this was like 15 years ago. 140,000 people is more than the populations of 40 countries in the world. So you can imagine like what that feels like to work in a company that's the size of a country. Um, in a company that size, there are always weak teams. And the problem with weak teams is on a weak team, you may think you are awesome, but that is because you work with very bad peers. Okay? So like that will lead to funny situations where people get overleveled, for instance. Um, or it'll lead to situations where people simply do not have someone they can observe

• 27:47Speaker 00

  that they can learn from because they're on a weak team. They're like the best. It's like how you buy real estate, right? You never want to be the most expensive house on the block because you have no, no room to grow up, right? Like you, you're already at the top. You don't want to be the worst house on the block. Nobody likes you, right? So you kind of want to be the middle house, right? And so you should find a block where you're the middle house. Um, when it comes to those higher levels, I do agree that it's very hard to find great examples of those people. And furthermore, each of those examples excel in different ways that might not be well matched to what you're good at.

• 28:21Ryan Peterman?

  Right?

• 28:21Speaker 00

  So I feel the more important thing at those levels is not necessarily another person to observe, but a great coach that can give you feedback. Someone who is a decision maker in the promotion of, let's say, an eight to a nine. And they can tell you like, hey, qualitatively, you're way off from where we need you to be. And then you can have a meaningful discussion about like, oh, what things can I do that would be more like that? Right? So I think it's much better to get that qualitative opinion than to necessarily find the right people to model after.

• 28:56Ryan Peterman?

  So when you got promoted to E9, what was the project or the scope? Just to give people a sense of, you know, what does it look like if something gets you to that level?

• 29:09Speaker 00

  Yeah, I had brought 12 people with me to start the engineering portion of the London office of Meta. Right? And so we started off pretty small like that. And over a period of about four or five years, we grew it to a site of probably 400 or 500 people. Right? And so through that growth, I managed several key transitions on the site. So when you start off with a site with 12 engineers, that's super easy. Like a great engineering manager at a level six could do a solid job running that. Right? But as you grow with each order of magnitude difference in the size of the team, the complexity of the problems becomes much harder and the solutions become much less black or white.

• 29:53Speaker 00

  They become always kind of gray in some direction. And so I think the promotion was primarily because I was handling a much wider variety of things and much more key strategic decisions. Let me give you a concrete example. One of the last meetings I had as a site lead was I was trying to convince the people in Menlo Park, the number of interns they should hire and the headcount they should have in London this year for this summer. Because I had done the math on the H1B anticipated volume for the following year for full time offers to matriculate into Menlo Park. Does that make sense? So I was having a conversation 18 months before the actual thing was going to happen.

• 30:38Speaker 00

  And I had used data to derive what actions we should take today that would only play out 18 months then. Right? No one else was thinking about that, right? When I was more junior as an E8 in the site, I was the first one of all of Meta sites to really look at not only ladder levels of all the engineers, but the pyramid of seniority. And to really think about like, hey, do we actually have the right seniority of people? Are we growing people at a commensurate engineering like growth velocity, right? And so things like that, which are also much more about strategic things than they are about sort of tangible, concrete things to deliver. So I think there are a variety of those sorts of decisions that become larger and larger in scope,

• 31:23Speaker 00

  either in the breadth of people they impact or in the timeline forward in which you're expected to think.

• 31:30Ryan Peterman?

  Is there something that you learned that really made a difference in upholding a strong engineering culture? Because you had a very unique opportunity to take, it's almost like you started a company within a company. What is it that really made a difference in making sure that site was very strong from an engineering perspective?

• 31:53Speaker 00

  Yeah, great question. I think one huge help to me before starting the site was not only did I observe other sites. So I was the second person hired in Meta Seattle. And Meta Seattle was Meta's first office outside of its headquarters back then in Palo Alto. So I had seen what it felt like to grow from four desks into 120 desks by the time I left. And I had learned from some of the things we learned from that were mistakes in how we grew that site, right? But more importantly, before I actually started the engineering office in London, I interviewed site directors from other companies at different sites. For instance, one super helpful site director was the former site director of Google London, David Singleton, who became Stripe's CTO sometime after that.

• 32:39Speaker 00

  He gave me some great advice. The best advice of which was that Google had experimented with what they call landing teams. How many people you bring to a site to bootstrap the culture and whatnot. Our decisions on landing teams were influenced by two things. One thing was Shrep. Shrep said he wanted two full interview loops of people, meaning he wanted enough engineers for two full loops, okay? So that estimates about 10 to 12 people, right? The thing I learned from David Singleton was that Google had experimented and Google's takeaway was make the landing team commit to two years minimum. Now you'll find many companies that approach this differently. They'll have a landing team show up for the first three months, the first six months, and then the site is on their own, right?

• 33:21Speaker 00

  I think that that's a mistake because it is very difficult to maintain culture in those early growth periods where the percentage of new people is like commensurate with the percentage of old people that have been around for like more than three months, right? And so I brought 12 people with me and I asked them to commit for two years. And that made a humongous difference. The other big difference culturally with the 12 people I brought is I interviewed probably nearly 50 people in the company that volunteered to go and I chose the 12, right? And one of the ways I chose it is I interviewed not only the people, but somewhat their peers and definitely their manager when it came to how they were good culture carriers for the company.

• 34:03Speaker 00

  Like, would this be a good person to add to the culture of the site? And I honestly really lucked out. The 12 people we brought with us were amazing, were amazing people for building and maintaining the culture of the company. They had been at the company long enough to know what its culture was, right? I also asked that each one of them be able to bring over a project of some sort because you also need someone that's able to lead work, right? And then you have exceptional talents, like people like Ben Matthews, who were individual contributors that were very happy to work on whatever the site needed, but could code the crap out of things, you know, was just very, very productive and very astute when it came to how to hire. So I really lucked out with the people that we ended up choosing to build the site.

• 34:48Speaker 00

  And so much of it was due to learning from others who've done it before.

• 34:52Ryan Peterman?

  I see. And you said that you wanted people to commit for two years. Does that, is that a verbal thing or is there a, you know, you get some additional equity for staying or something like that?

• 35:05Speaker 00

  There's no compensation to it. So people got paid whatever they got paid before, right? And we didn't make them sign anything that said it was two years. But I basically had a serious conversation with them of like, hey, I'm only looking for people who are committed to at least being there two years, right? So if that doesn't fit in with your life plan, this is probably not the job for you. Right? And so those folks sort of elected, self-elected off, right? We had one person who needed to go back sooner than their two-year. One or two people needed to go back sooner than their two-year time commitment. And those were understandable reasons. And we just talked about it when it happened and I was very supportive of them transitioning. But the surprise even to me is far more people stayed much longer than that.

• 35:49Speaker 00

  In fact, from the original landing team, I want to say probably about four people are still in London right now, 15 years later. Like, they are basically British citizens, right? And when they joined me in London, that was not the plan. They were not joining me because they sneakily thought they would become British citizens. Like, they ended up loving London, they loved the office, they stayed committed, right? So I think in the end, we also were very fortunate to have people that were committed like that without any sort of real enforcement of it.

• 36:16Ryan Peterman?

  You mentioned that Shrep was your manager at some point. And when I was digging through things, I saw you had worked with Boz as well. And I imagine you had some proximity to Mark Zuckerberg. Do you have any favorite stories working with them? I was very curious because they're all so legendary.

• 36:34Speaker 00

  Yeah. You know, I was very fortunate in working for a lot of great people at Facebook. So my first manager was Boz, right? And Boz eventually became the CTO. My manager during the years of leading London was Shrep, the then CTO. Both gave me sort of great advice and guidance.

• 36:58Speaker 00

  I only talked one-on-one with Zuck a few times. So I don't really have too much insight there other than he felt to me like he led the teams. He felt very genuine to me in how he led the teams, you know? One thing that impressed me most about Zuck is, and this is true of Boz as well, is you gotta remember when I joined the company, I was 33 years old, right? The median age of the company back then was 27. And so I was the oldest guy in the Seattle office for the first seven months. Like a 33-year-old was literally the oldest guy, right? And so most of the people I worked for were actually younger than me, right?

• 37:37Speaker 00

  But I was very impressed with both Boz and Zuck in their ability and willingness to personally grow. Like in each of those cases, if you talk to people that have worked with Boz for more than eight or 10 years, right? Or with Zuck for a good number of years, they can tell you exactly how they grew over time. It was very obvious that they were not only open to coaching and probably seeking it on their own, right? But making a deliberate effort to improve. Like small, concrete example, if you look at how Zuck ran AMAs when I first joined versus even two years into my time at Facebook, where I'd randomly guess he was probably like 28 or something. He had greatly improved how he communicated at AMAs.

• 38:21Speaker 00

  So like there were really neat thing. I'm pretty sure, for instance, small thing, but I noticed, right? I'm pretty sure someone coached him on how to hold his head when he was speaking in public. So when I first joined Facebook, Zuck's natural pose when talking was head slightly upward. So he would look slightly like this at the crowd, right? But through the years, you saw him deliberately correct that. It was very obvious. He like deliberately corrected it to where now he faces the crowd straight on, right? And so I was always very impressed by the things they were very willing to improve in themselves and put effort into.

• 38:56Ryan Peterman?

  You mentioned some of the highest level ICs. The qualitative feedback from people around them is a really important aspect. I'm curious, from your perspective, having worked with a lot of Meta's highest-profile ICs.

• 39:13Ryan Peterman?

  Is there someone that stands out to you that consistently impressed you?

• 39:17Speaker 00

  One person that I worked only ever so indirectly with, so barely with, John Carmack, obviously legendary, right? But the thing when you work in Oculus that you realize is not only is he super prolific in coding, okay? So he produced so much code, but he had an ability to drop into your code base and product after not looking at it for like six months and give actual concrete advice about what to do. Like I was working on a very tiny product called Oculus 360 Photos, very simple product. And just one random day, like John dropped into the code base and he tried it out and he like had had insights about like the distortion on the corners of the photos, like when you turned.

• 40:02Speaker 00

  And he had some performance ideas for like, hey, I think we can speed up performance like this. And this is all without anybody, A, asking, right? But B, like he said some really smart things that just no one on the team had thought of. So that was impressive, but I didn't work with him as directly. A great example of an interesting shaped person, Scott Renfro. So I worked with him a bit more because he came out for a rotation in London for six months. One thing that you'll find consistently is peer feedback about Scott's work on a team is super glowing. Like when Scott reviews your PR, right? Even if you do something very dumb, he will put it in a very sensitive way not to shame you or anything.

• 40:46Speaker 00

  So he'll say it in a way that you understand what he's saying is like, hey, this thing could be done differently, right? But he'll never publicly shame you or anything like that. If you look at peer feedback for him, it is overwhelmingly like Scott takes time out of his day to make me great. So he was an example where I think his seniority was well deserved in the company because it was very obvious that having him on the team would make your entire team much better. So he was a force multiplier for teams. And later on, he led these huge things for Meta. I'm pretty sure he had something to do with GDPR at some level and like a few other things like huge infrastructure wide like massive changes.

• 41:29Speaker 00

  And so he was capable both personally of guaranteeing a huge scope of work, right? But he was also a force multiplier for the teams he was on. So kind of a slam dunk case of someone who deserved to be rewarded well for that work.

• 41:43Ryan Peterman?

  Coming to the end of your your tenure at at Facebook or yeah, it was Facebook at the time. I I remember I joined in twenty eighteen and I had been there for some amount of months. And I heard in the echoes of workplace that someone had bought the entire company coffee for a day. And, you know, now I I've read a lot of your posts and I know that was you. What what made you want to buy everyone at the company coffee for a day and spend over twenty three thousand dollars?

• 42:16Speaker 00

  Yeah, great question. I'll I'll start with the original story, which was several years into my time at Facebook. Facebook had moved into its current location in Menlo Park. Hacker Square wasn't fully paved yet. Eventually, a filled coffee opened there. And one Christmas, I just thought, well, let me buy everybody a coffee for Christmas, right? So for that Christmas, for one day, I had paid for coffee for everyone. And I just felt it was such a joyous thing to see how happy people were to get a free coffee from someone, right? And after that day, the CEO of Phil's had written me to say that his employees really enjoyed the the day a whole lot. Because it had never happened in their history, like something so crazy, right?

• 43:01Speaker 00

  And so I felt like it was a sort of a pay it forward sort of thing that made me feel great, that made others feel great. When I left the company, I really felt like, you know, I loved the company so much. And the company gave me so much over time, like gave me the opportunity to do really interesting things. Gave me incredible coaching, right? When I joined the company, a brick of money fell on my head, right? Like, I did not join Facebook. People forget, when I joined Facebook, the big talk was still like Friendster was coming back. G plus was going to roll over Facebook. Like, there was all sorts of talk. And eventually like Facebook would never make money on mobile like there was a lot of like rock throwing, right? And so when I joined the company, I had no expectations for what would happen. I had left Microsoft after working at Microsoft 12 years.

• 43:47Speaker 00

  The stock had gone up 15 cents in those 12 years, you know, so I had sold my seven eighths of unvested like remaining shares for $1,600. And I went to the company store and bought like an Xbox 360, a driving wheel and Forza 3. And I was thrilled, right? So the fact that a brick of money hit my head, not by my own design, but really just by being lucky, right? Made me so grateful, you know, what I did when I first, when Facebook IPOed, I gave 100% of what I made from the IPO to charity, you know, because I felt like it was not a thing that I deserved, honestly. And I don't mean to poo poo myself, like all I mean is like, I was lucky enough to be born in a timeframe

• 44:30Speaker 00

  where my interest in computer science actually paid something versus if I worked for a blacksmith 400 years ago, right? And I happen to have migrated to America, land of opportunity, right? So I felt super lucky. And so what I wanted to do was, given the great experience that fills that first time with Christmas, when I left the company, I felt like I want to give something back, you know, for all the things the company has done for me. And so what I did was I coordinated with the major sites that had employees. So back then it was probably five different sites I thought of coordinating with. And I just called the individual owners of the coffee shops in those places and arranged a way to pay for a coffee for a day. And it was super fun and really rewarding to do, and I'm really glad that I did it.

• 45:14Ryan Peterman?

  Yeah, I mean, it definitely left an impression. I think everyone, me and all my friends were kind of new grads at the time and, you know, it's kind of this fun little thing. But I don't remember if it was attached to your name at the time. I think we just thought it was someone nice that Facebook was paying, so, you know, it was appreciated for sure. So when you decided to leave Facebook, I'm curious because it sounds like everything was going quite well. What was your rationale for leaving?

• 45:43Speaker 00

  Yeah, great question. The biggest thing on my mind back then was I was very concerned about the rising income gap in America. And so I wanted to start a B Corp to bridge the rising income gap. I had a few ideas for how to do that. Back then in 2018, I was also, like, this doesn't sound revolutionary now, but back then I was complaining to coworkers that I think were going to experience technical unemployment. You know, and back then that term hadn't even really come around, but I was convinced that robots would replace people and that our society needed to change quickly to adapt to them. And so most of my coworkers disagreed, but I felt very concerned about it. So I thought I should actually do something, like start a B Corp to try to do that.

• 46:26Speaker 00

  And so one of the main reasons that I stepped down was that I felt like Facebook had grown to a point where it would be fine, honestly, without me, like plenty more than fine. It was doing really well. And I was concerned about this thing happening to America. I think another thing, if I were to evaluate myself having left the company at the same level as when I joined, right? Is I feel like I was much more useful as an individual contributor when Facebook was small. When I joined as the second person hired in Seattle, Facebook had 500 engineers globally at that point, right? But that meant that I, along with two other engineers, could ship the entire video calling feature,

• 47:08Speaker 00

  you know, from nothing to shipping in a period of probably six months, right? I'm a generalist, not a specialist. So I was also able to like help with the negotiation of things like what does the Skype contract say? I helped negotiate and hire and manage the contractors that were testing on it. So I was capable of doing a very wide variety of things. Eight years after that time, the company had grown to I mean, just thousands of engineers, right? The company had grown to a point where I think a generalist like me is much less useful, you know, than specialists. And so I also felt that my unique value contribution was no longer the same.

• 47:48Speaker 00

  You know, I was a very generic coder and I wasn't contributing anything that was exceptional, honestly. And so I also felt like I was much less effective for the company.

• 47:59Ryan Peterman?

  Would you say that's generally true? So, you know, taking it out of big tech, imagine someone wants to work at startups. Would you say the generalist's skill set is the way to go? And, you know, specialists can have an edge at the largest companies?

• 48:14Speaker 00

  Yeah, that's my intuition is that most small companies value generalists more than specialists. Like every small company can use definitely one specialist, like no matter what domain you're in, right? You probably want at least one person who really knows that that domain super well, right? But I think most of the times what you deal with in small companies is quantization effects. Like if your company is large enough, you can hire a PR comms person whose full-time job, 40 hours a week, is to handle comms, right? But before you're that large, it's very difficult to hire a part-time, three-hours-a-week comms person, right? And so you've kind of tapped some product manager on your team that's pretty good with communication and you ask them to handle it for a while, right?

• 48:56Speaker 00

  So because of these quantization effects, I do think generalists are more purposeable in those cases of fractionalization, right? You can fractionalize the generalist to a bunch of different roles. But eventually, a company gets so large that each fraction of work that needs to be done is probably, you know, warranting one or more employees full-time on it. In which case, having the generalist is kind of like a lot of wasted energy, right? Because they are only, you know, only 15% of their skill set pertains to the actual work you're doing.

• 49:25Ryan Peterman?

  So at some point, once you left Facebook and then you pursued the B Corp, I saw that you did, you came back to the industry for a little bit to join OpenAI. I'm curious, what's the, the story behind you coming back to industry?

• 49:44Speaker 00

  Sometime after quitting Facebook, I ended up starting a local nonprofit in Seattle called Outarray that builds free software for global health, you know? And this was funded by the Gates Foundation. And the software is deployed, I think, in Nigeria, Kenya, and South Africa. And it helps with assisting rapid testing for malaria. Okay. That work was simply because I didn't expect to have a grant from the Gates Foundation. I felt very fortunate and lucky to have been chosen to get a grant for that. And I felt like building some free software for global health seemed like a great idea. And so I really enjoyed that work. After that, I took a break for a while where I was unemployed.

• 50:26Speaker 00

  And I think part of that during that time, I had started using a lot more of GitHub Copilot in VS Code back then, trying to evangelize to friends that, hey, this AI thing, like it's, it's useful, but I had a hard time convincing people to join back then to pay for it. And so when AI began to really take off, you know, sometime in the timeframe of maybe ChatGPT three, five, or four, I felt like, wow, this is, this is a huge transition point in the industry. I feel like I missed two major transition points because I was at Microsoft. First of all, I missed the whole dot com thing because Microsoft as a whole missed the internet revolution, right? And so that was a massive thing. The second thing was the mobile revolution. Microsoft largely missed that as well.

• 51:10Speaker 00

  And I was at Microsoft when that happened. And so I felt very fortunate to have caught the social network revolution, right? Which was a huge transition point in humanity. But I increasingly felt like AI would potentially be as big or maybe bigger than the internet's impact on humanity. Because from a universe perspective, we might be the first species to create something smarter than us. And from like an evolutionary perspective, that's just really weird. Like that goes second order, like, once that happens, right? It's like, what is actually happening in the universe, right? We might answer the Fermi question, right? And so I felt like this transition was something I didn't want to miss. The reason I joined OpenAI was what I learned from working at Facebook is I would

• 51:53Speaker 00

  much rather join the market leader or nobody at all. And here's why. I think the market leader has room to experiment because it can afford to fail at a few things. This is what I discovered at Facebook. With a big enough market lead, you can take some brave big shots. Because like if two things fail, and Facebook had some big colossal failures too, right? But you have enough of a lead to do it. Usually, the market follower, like Google Plus as a good example, right? Usually, the market follower, their formula has to be match all of the leaders' checkboxes and have one unique thing you think is going to be differentiating. In the case of Google Plus, the thing was circles. This idea that you might want to manage circles of friends that are different, okay?

• 52:36Speaker 00

  But to have a social network, you had to do all the basics that every other social network did, right? So I feel like especially number twos, number threes might be able to do crazy hail marys, but number twos, the problem with number twos is their whole schedule and timeline is dictated by the leader. Like, they have to be fast follow, right? Whereas number one gets to blow a lot of time on crazy things, and I love the craziness. So that's why when I look to join AI, I only apply to OpenAI. My plan was I either join that or I'm still going to write my own software.

• 53:10Ryan Peterman?

  I see. And when you joined OpenAI, how did it compare to Facebook?

• 53:18Speaker 00

  Great question. I had joined Facebook at a different stage in its growth. When I joined Facebook, it was 500 engineers, right? When I joined OpenAI, I want to say it was about 120 engineers. So like qualitatively, that is very different. So with that caveat in mind, it is really the highest talent density of any group of engineers I've worked with. And so it was really amazing. And that's after, when I shifted from Microsoft to Facebook, I felt like Facebook was a huge jump in quality of engineer on average of the person that I worked with. And there are just amazing people, especially in early Facebook. It was like full of amazing people. I felt like OpenAI was even a notch above that.

• 54:01Speaker 00

  And so it was really amazing, the people I got a chance to work with and to learn from.

• 54:07Ryan Peterman?

  What are the things that made you see how amazing it was? Was it the technical abilities or something else that really stood out to you?

• 54:17Speaker 00

  Yeah, people had exceptional technical capabilities. I would say one thing that Facebook always had a lot of that I believe OpenAI is going to get increasingly better at is I think Facebook always focused a lot on product. You know, like, what is the product? Why do people use it? And I think there's a way in which, like, because Facebook did not honestly have much magic sauce beyond a network effect and a great product people love using, right? It causes you to work on it. Whereas, I would say, in the early ChatGPT days, when there were no viable competitors, right? You kind of could have a lousy product, people use it anyway, right?

• 54:56Speaker 00

  And because like you, you might remember like a year or two ago, it's very common for like every fourth conversation thread to have an error, like a server error. And then you like reload the thread and you rerun it, right? You can get away with that when nobody else is producing anything near as compelling. But I think as competition heats up, people will begin to see that ultimately products are what make things. Not pure research, but actual research shipping in compelling products is what makes for success.

• 55:26Ryan Peterman?

  Yeah, yeah, definitely. I agree. The early days of ChatGPT, it was so special that you you put up with it. And now people are catching up a bit.

• 55:39Ryan Peterman?

  One thing I wanted to ask you, and I think this is because I, you know, am a fan of your writing and I've read so much of it. What I've seen is that your your writing ability is a superpower of yours. And I know writing is important for software engineers. It gets you a lot of leverage. I'm curious, how how did you develop that skill? Were you were you writing from a young age? Or is that something you developed at Microsoft? Yeah, what tips do you have when it comes to writing?

• 56:07Speaker 00

  I was bad at writing at a young age. The main reason that I was bad at writing when I was, even through college, is I very much did not respect the value of good communication. Like, I came from a culture that respected getting As in everything. And like, I felt like there were hard sciences and then everything else was fluff. So if you made me take history, if you made me take English, I felt like this was a waste of time. These are subjects with no right answer. It's all like people's opinions, right? And so I had a very sort of low attitude toward the soft skills. And so I did not respect the soft skills, and so I didn't invest in them. So I was a bad writer, right? Over time, I think I lucked out in that when I started work at Microsoft,

• 56:52Speaker 00

  in some of the spare time I had, I just started rereading some of the classics. You know, and these are things that in high school, when assigned, I really hated them, right? It's since these are all the classic like Hemingway's and then Great Gatsby and like stuff like this. And I started to see these books in a different light. Like I was a bit older, so I could appreciate more of what was said in it. And I also started to really develop a love of the language, right? I think it was Virginia Woolf that said, to write well, you need to read well. And so I think first of all is you gotta read well, right? And so the types of sentences that someone like David Foster Wallace, DFW, can put out, the types of sentences that he can put out are amazing. If you read Alice Monroe's short stories, right? The language is beautiful.

• 57:35Speaker 00

  A pilgrim at Tinker Creek, like there are all sorts of amazing books where when you read it, you feel like, wow, the language here is exceptional. So I then started to really respect that that was a skill, you know? And that was a valuable skill, every bit as valuable as a hard skill, like a science skill. And so then I started working at it. And a lot of writing better, I think, is not only reading well, but really being willing to rewrite multiple times. So many of the posts that you've seen me make, I have rewritten multiple times before posting them. Right? And so being willing to have some things on the, you know, cutting room floor, right? Is also pretty important. Even now, when I read people's substacks, for instance,

• 58:18Speaker 00

  I once in a while do notice something and I try to remember it. Like, oh, this way of saying something rhetorically was very clever. So I'm going to try to remember that this is one way you can communicate an idea. And so I think that has also helped. But most of all, I think, was my willingness to not disrespect the skill and to actually see it as an essential skill for, even for scientists, for technical people. I think it's hugely important.

• 58:45Ryan Peterman?

  Yeah, definitely. And what you said about growing up and not respecting the softer subjects definitely resonates. I have very similar value system when I was growing up. And now I see the value of writing, and so that's why I'm kind of into it.

• 59:03Ryan Peterman?

  Some of the stuff that you wrote too, for instance, your essay, and I guess this is only, only people at Meta could, could see it, but your essay, as good as it gets, your, the creativity to kind of jump around and give a non-chronological narrative, I thought was really cool. You're jumping around your life and just getting the feel of a bunch of things. And some of the things that you communicate are not, it's not even a full, you know, narrative bit, but it somehow gets the idea across. I really liked that. And I'm curious, do you have an editor when you were writing, or that's all from you continuously reading and rereading?

• 59:44Speaker 00

  Yeah. That thing was a moment of inspiration where I thought, how do I tell this story? You know, like, what is a compelling way to tell this story? And most often, like, the compelling way is chronologically, actually, because that's the way that makes sense, right? But some stories, I think, work better out of chronological order, because you see a different arc that's developing in a person's life. An arc that's actually like threaded, right? And so that's what I felt for that particular story. I've not had an editor for any of those, but one place where I did hire an editor, it was great. I hired a lady named Rebecca Ambrose to edit my podcast, Peak Salvation. And the big thing about that is I knew I wanted to do a mini-series,

• 1:00:29Speaker 00

  and I wanted the words to be well-structured. I wanted the story to be told well. And I knew that a professional would do a much better job at editing. And there I learned a ton from Rebecca. Like, if you listen to those episodes, you can tell that, oh, someone who works in the profession has, has like given feedback on this, right? And so I have gotten feedback in those cases, but nothing ever that I've posted professionally was edited by another person.

• 1:00:55Ryan Peterman?

  I see. Yeah, your essays are legendary. I think I, sometimes I go back to visit them and I'll see they're still alive. People are commenting saying, oh, I just found this. This is great. Or, oh, you should reread this. Big fan. So glad you put those out.

• 1:01:10Speaker 00

  Thank you.

• 1:01:12Ryan Peterman?

  The, the great, by the way, the great post explorer, I, I, it was alive as of maybe two-ish years ago. I think someone took it down at this point. I tried to, before we did this, I was like looking for it. I'm curious, how did that side project come along?

• 1:01:28Speaker 00

  Yeah, so I wrote a thing inside the company Facebook called the great post explorer, which was meant to like rank and expose people to some of the most commented on or the most reacted to posts of all time within the intranet of Facebook, right? The main reason I wrote that was I benefited from discovering serendipitously great things written by people who came before me in Facebook that I wish were somehow cataloged somewhere. And not only cataloged, but like crowdsourced in how they were categorized because I wanted to be able to navigate it by categories, right? And so I built that thing with that hope. And so the whole purpose of it was the realization that,

• 1:02:08Speaker 00

  you know, Slack has a worst case of this, but, but Facebook has this, which is like posts fade temporally. But in my opinion, posts values, the value of a post is often temporal. But in the cases of some posts, the value is not temporal. Like this is just like some podcasts, you know, our evergreen in some podcasts talk about current news. And so to me, I wrote the great posts explorer as a, as a chance for people to discover some amazingly timeless things that people have posted in the past.

• 1:02:39Ryan Peterman?

  Part of writing well is reading quality stuff. And that's also what made me want to go deep in there because you essentially created a curated set of quality posts where all of Meta was helping you curate. And so, yeah, I loved that when it was up. I'm not sure what happened to it. Coming to the end, I just wanted to ask you a few reflections on your career that might be helpful for some people. So the first thing I wanted to ask was, when you were young, you prided yourself on fast execution speed. That was something that set you apart, at least from I can tell in your, your writing. And then later you were more of a leader. You slowed down in the execution.

• 1:03:22Ryan Peterman?

  I'm curious, do you, do you find that software engineering is like being an athlete from the perspective of you have some performance peak at, I don't know exactly what, for an athlete, it's probably late twenties or early thirties. And then there's a performance decay. Or, or do you think that, you know, you continue to develop experience in software engineering doesn't have that cap?

• 1:03:50Speaker 00

  Yeah.

• 1:03:52Speaker 00

  I feel like you can get more and more valuable over time. Now, everybody has an asymptote and everybody's curve is probably asymptotic. Meaning like you probably have a phase in your career where growth is easy. And then everybody has a final level, right? Everybody's final level is different, by the way, right? And, and so as you approach that, you should feel like the growth is slowing. However, just as with work hours beyond fifty or sixty, right? I believe each incremental growth can still be growth. So like, I don't think you have to stay fixed and you're a flat line forever. I think here's the thing. When I was young, I would outrun everybody, but I was running serpentine, basically. Right?

• 1:04:33Speaker 00

  As I've gotten more experience, the idea is to jog in a straight line, right? And how does a senior experienced person jog in a straight line? Part of it is, an experienced person usually gets a spidey sense for when things aren't headed the right direction and they either don't pursue that thing, or they convince that team to stop that early. Like, they basically save an entire team's worth of time because they're like, hey, I feel like we're heading this direction. I've seen this in the past. One of the things that can happen is this. Concrete example, I had a friend who worked at Microsoft, who was telling me that the product he was working on, the vice president of the organization, met with them every week to triage the bugs towards shipping. And I told him, look, if your VP is in triage meetings on a weekly basis so that you can ship on time,

• 1:05:18Speaker 00

  your project is nowhere near shipping on time. Like this thing is in a huge flame ball. Like this is a huge problem, right? I don't know how people that experience can live through that and not see, you know, how obvious it was to me that that thing is like a bad sign, right? This is a tell that your project is not on track. But I think you develop that sense of like, oh, this is worth investing in. This is not worth investing in. I think the other way you grow your scope over time is I think you begin to value everybody running in the same direction, even if the direction is three degrees off ideal. You know, when I was young, I had this idealism to me. So I felt like there was a right direction to go. And like two degrees off was like wrong, you know?

• 1:06:03Speaker 00

  But the problem is a whole team pulling 98 degrees correct versus like half the team pulling 100% correct and the other half debating them on like which is correct. I think the team pulling together is going to get further, you know? And like that was wisdom over time of saying like, even though I don't agree with 100% of what, this is what Amazon means by disagree and commit, right? It's like there are times you need to disagree and commit. And like that is valuable in itself. And so I think some of that wisdom I continue to develop. But when it comes to, for instance, a young person's ability to outwork me is definitely true now. Like I do not have the physical stamina I had when I was 24. If you woke me up at 3am today to start coding, I would code very slowly at 3am.

• 1:06:48Speaker 00

  Like when I was 23, if you woke me up at 3, it was no problem. I would just get going, right? So I do think that some things get weaker and hopefully if we're growing, like other things get stronger.

• 1:07:00Ryan Peterman?

  At the beginning of your career, it sounds like you, you were kind of leapfrogging people to some extent. You became a very young manager. And management is often seen as a profession where your reports need to respect you and there might be some unusual dynamic because of the age difference. Let's say you have older people reporting to you. I'm curious, how did you win the respect of these people as such a young manager?

• 1:07:26Speaker 00

  Yeah, this was difficult and I think that I both lacked humility when I was younger and I was overly ambitious for myself. And so as a concrete example, I would not have worked for my 24-year-old self, you know, but I was so ambitious that I was one of the team's strongest engineers. So when I asked to manage people, people were like, well, he's a good engineer. Like, let's not, you know, make him unhappy. Let's like have him manage people, right? In that first year, I had to, as a 24-year-old, I had to fire a 40-year-old and that was very difficult. Like, firing people never gets easier. Like, hopefully, like, hopefully you aren't like the eye of Sauron and you delight in firing people. But like, that was

• 1:08:06Speaker 00

  exceptionally difficult. And it was, in my opinion, made more difficult by my self-awareness of just how junior I was relative to this person. You know what I mean? And so, um, now, he did deserve to be fired. He was not performing to the level of expectations that we had of him, right? But he was also a person that was two levels above what my level was and I was managing him, right? And so like, I was too immature to know how to manage people more experienced than me. Like, that's its own sort of maturity is how do you manage someone who's a higher level than you are, you know, like, how do you even do that successfully? It's a tricky thing, right?

• 1:08:44Speaker 00

  So I did a very poor job of managing and so I was a, a, a pretty bad manager. But I was so obsessed with growth that I was also a huge sponge for learning. So I was constantly getting mentors to grow in some area. I would ask for feedback, like, how can I do this better? I would read a lot of nonfiction about things like managing teams, about what conversations, like, all sorts of things, right? So I was doing a lot of self-improvement over time. So I do think that I grew over time, but I was always like, the start was pretty rough and, and I've had a lot of things where I've managed poorly. Even in my last year of managing, uh, people, you know, I

• 1:09:24Speaker 00

  made some huge mistakes. So, for instance, uh, one of my biggest weaknesses when managing people is I am very bad at giving feedback, you know? So, like, I, um, Pull punches when I give feedback. I resist giving feedback because I hate the awkward feeling of confrontation. Right. And so I resist giving feedback. When I give the feedback, I try to downplay how important it actually is. And so people don't understand how serious it is. And so I continue to be bad at doing that. I think over the years, I've gone from very bad to like pathable, you know? But that is not an area where I expect to become, you know, excellent. Whereas I do think I am much better at motivating action.

• 1:10:07Speaker 00

  So I think when I lead a team, whether I'm a manager on a team or I'm an individual contributor, you know, I only join teams that I believe in. And when I believe in something, I'm very passionate about it. And I think communicating that passion and that vision excites people to do the work. So I think I am good at that, but there still are a lot of things to to improve.

• 1:10:25Ryan Peterman?

  We talked a little bit about generalist versus specialist. And it sounds like you identify as a generalist. I'm curious, how did that play out over your career? And let's say there's some new grad or someone earlier in their career. They're they're thinking, should I become a specialist or a generalist? How would you make that decision?

• 1:10:46Speaker 00

  Yeah, that I think is a tricky decision, depending on how well the person knows themselves. So there's the occasional exceptional person, like these prodigies in chess, for instance. Right. They will have been a prodigy by the time they're eight or nine years old. And so they're obviously fit to play chess. That person should specialize because that's like unnaturally unique talent. Right. Most people aren't like that. Most people do not like from age seven have a solid like thing of like, I am very good at this. I should do this. Right. And so for most people, I would caution against early binding too much to a specialty. You know, there there are two dangers to binding too early. One is, how do you know you're a specialist and not a generalist without being a generalist? Right. So that's part one.

• 1:11:29Speaker 00

  Part two is you have an idiot savant risk, right? It's like you bound so early to a specialization. Like Erdesh, a famous mathematician, even as an adult, did not know how to cut his own grapefruit with a knife. So his mom would cut his grapefruit. Okay, but this was an amazing. Amazing mathematician that published tons of papers.

• 1:11:48Ryan Peterman?

  Right.

• 1:11:49Speaker 00

  But he specialized so early, he can't even cut fruit. Okay. So like that to me feels like over specialization. But he did some amazing work in mathematics, right? So I feel like if you are twenty two, twenty three, starting your career, I would in general encourage at least dabble in a few things before you like die hard commit. One last risk to committing to a specialty way early in your career is what if that specialty is going away? You know, especially in the age of AI. This is something to think about. It's like if you join some company and you're diehard committed to like uh uh technology A, right? Um, what if in three years that thing becomes irrelevant? And that's all you know, you know, you're like the Cobalt person like hoping like Y two K happens again, right?

• 1:12:33Speaker 00

  It's because Cobalt not used anywhere, but that's your specialty. So I think specializations come with that danger, that generalists don't suffer that danger.

• 1:12:43Ryan Peterman?

  I guess the last question that I'm curious about is after this career that you've had, if you were able to go back to the beginning when you just entered the industry and you give yourself some advice based on everything that you know today, what would that advice be?

• 1:13:02Speaker 00

  Yeah, great question. I have a few thoughts that I share. One, Roy Disney said, decisions are easy when your values are clear to you, right? Decisions for me a lot of times were hard because I didn't have clear values. If you know exactly where you're going, decisions toward getting there become a lot easier. So one advice I would give my younger self is spend a little more time thinking about what you actually want, right? Before you commit to doing these things. I think another thing is I often feel like I was the dog that caught the car, you know? Like I was convinced when I joined Microsoft that what I wanted to be was a dev manager. Okay. And so for my first eight years or so, I would take any job that would get me a step toward that.

• 1:13:45Speaker 00

  So one, decisions were easy then because it was like, well, does this job get me a step closer? Yes, it does. Then I'll take it, right? But once I caught that car, you know, once I hit that level, once I recognized that, hey, E seven might be my terminal level. This might be the highest I ever get in my career.

• 1:14:05Speaker 00

  The problem with peaking early, you know, because I hit that level when I was probably, I don't know, twenty, you know, thirty years old or something like this.

• 1:14:14Speaker 00

  The problem is, you're like a child actor. Like, the question is, what are you gonna do with the rest of your life? Like, if your whole plan is acting, and if nobody hires you for acting, you're gonna be in a world of hurt. Like, you are not gonna enjoy the next forty years of work. And so for me, I was the dog that caught that car. And I had no mental model for like, once I became the death manager, which I did become, right, at a very young age. Once I became a level sixty-seven, and I became that at a relatively young age, it was sort of like, well, like, what happens next? And I really went through a period of pretty serious depression, because I felt like I had lost a purpose to life. Like, life seemed to have so much meaning and direction when I had a clear goal. But I caught the car, right? And then it was like, what to do?

• 1:14:58Speaker 00

  So the other thing I would caution my young self about is to just be sure you actually want the thing you want. You know what I mean? Like, one thing I tell people now is, you know, would you wanna be Warren Buffett? You know? So, many people love the billions of dollars. But he also is in his late nineties. Like, would you want a few billion dollars and be in your late nineties? Like, is that worth it? Is that what you want? Like, this is an interesting question, right? So I think for me now, I think a lot more about, am I even going a direction I wanna be at when I get there? Right? One final thought that I'll leave with you with the sleeping bag business is, sometimes you bend the reed, sometimes you break the reed, right?

• 1:15:42Speaker 00

  Sometimes things break and they aren't fixable. So when I had sleeping bags in my office, my fiancee came out to Seattle to visit me. During that time, when she visited me in Seattle from Maryland, during that time, I would see her every evening at eleven PM, when work ended. That was when I would go see her. So basically, I got to eat like a late Denny's dinner with her from eleven PM to like midnight. And that was my plan for my visiting fiancee, right? I am very fortunate and lucky that she is my wife now. But anybody I tell that story to, you could bet nine out of 10 times that the story that ends with, that's how I got the engagement ring back, right? Is how that story ends, right?

• 1:16:27Speaker 00

  I did not realize that some reads bend and some reads break. And I was clearly bending something to an extreme that most things would have broken, right? And so I do think that in one's life, back to your point of like, what should a diehard person who only cares about career do? Like, what is the absolute fastest way to like get there, right? I would say, A, be sure that's what you want, right? Don't be the dog that caught that thing and then you regret catching the thing, right? So, A, be sure that's really what you want. And part B is like, you know, be sure you're comfortable with other things breaking, you know? Because like that is what it will take to get there if that's truly what you want. So my advice to my younger self about that period of time would have been, you know,

• 1:17:12Speaker 00

  in net, getting to level 67, getting to an E7 when you're 30 versus 38 in the big arc, doesn't make any difference. Like beyond 38, I still have 30 years of work to go, right? So it's like, how fast do I want to be at my terminal level? Like, what's the real plan there? Versus can I keep a healthy relationship with my spouse, with my kids, right? That's important. And so that's what I would advise back then.

• 1:17:37Ryan Peterman?

  Yeah, it's funny because I, I think maybe this is a type of advice where, looking back, it's obvious, but I wonder if you right now could talk to that Philip that was grinding, you know, at 26 or 27, would you even accept the advice? It's hard to say. I feel like a lot of people who are fully focused may not.

• 1:18:01Speaker 00

  Well, I feel like it's easy to say. I feel like for me, the answer would have been no, I would not have accepted the advice. Because here's the thing, like, everybody tells you money doesn't buy happiness, right? Like, when, when someone with $3 million tells you, trust me, money will not buy you happiness. If you're an unhappy person now, you're going to be an unhappy person that can spend a lot of money, you know, 10 years hence, right? But nobody believes it, right? Like, everybody thinks, oh, yeah, yeah, yeah, but when I'm wealthy, I'm going to be happy, right? Everybody thinks they're the exception to that rule. So I agree with you completely. Like, advice feels good to give because I feel like maybe I'm saving someone from something. But in reality, 90% of the times, You have to be ready to receive the advice for the advice to have any impact.

• 1:18:46Speaker 00

  And I was at a maturity level that was nowhere near ready, nowhere near ready to receive that advice. So I would not have benefited from my own advice, for sure.

• 1:18:55Ryan Peterman?

  Well, yeah, Philip, thank you so much for this. I mean, I was really looking forward to this conversation and I learned so much. I really appreciate your time. This is awesome. I'm gonna polish this up and make it really good for others so they can enjoy it too. Is there anything that you wanted to shout out to the audience? Maybe we can redirect them to something that you're working on?

• 1:19:18Speaker 00

  One of the things I might recommend, just because I'm still concerned about technical unemployment, is people might give my podcast mini-series, Peak Salvation, a spin. It was about my time working at Amazon's flagship warehouse over peak season, from Black Friday to Christmas, right? I feel like in there, I was able to explore a lot more of my concerns around automation, what it means to society as a whole, the income gap in America and its rise, right? I do think that as Americans, we need to get together and like figure out how we help the average person through this change. I think, as you know, in technology, the change is going to be massive. It's going to come sooner than people expect. And then very bad things can happen if we don't help the whole society adapt to it.

• 1:20:03Speaker 00

  So I would encourage people to like, think about that. Maybe give the podcast a listen, which might give you different perspectives about it. And hopefully, you know, figure out a way to contribute to making that better.

• 1:20:15Ryan Peterman?

  I will put that in the show notes. So if you're interested in listening to that, take a look at the show notes. I'll put it in there.

• 1:20:21Speaker 00

  Cool.

• 1:20:21Ryan Peterman?

  Okay, well, thanks so much.

• 1:20:22Speaker 00

  Thanks for your time today, Ryan. This was great.

• 1:20:24Ryan Peterman?

  Oh, yeah, I mean, thank you. You're the one that I want to thank. This is really great. I've been wanting to have this conversation for a while. So really, really appreciate your time.

• 1:20:34Speaker 00

  Yeah, and I appreciate the effort that you put into making these valuable to other people. So thank you for putting that investment in.