Instagram Founders Kevin Systrom and Mike Krieger on Why Social Networks Should Be Less…

Yeah. I think it is a rarity. We are second time founders. We’ve worked together for thirteen years now with basically, like, a year break, and that was mostly us having left Instagram, us both having new kids. And pretty much even in COVID, we were working on projects together. I used to say it was because neither of us wanted each other’s jobs, but I think that’s a little bit, like, a reductive answer.

The foundation is, like, a tremendous degree of trust and integrity, and I think that’s the unquestioned part that has when things get crazy and you’re fundraising or things aren’t going well or just all bumps that are inevitable in building a company, to know that you’re not also questioning whether your founder has your back is huge, and I’ve never once had to question that. That’s an enormous kind of foundational piece. I think we both also have the trait, and this is maybe one that we’ve acquired over the years, which is even if we disagree initially, it’s often the case that after a few hours, we’ll see each other’s viewpoint and come back.

Like, I can’t tell you how many times, Kevin’s told me something, and he says, I see the Mike face where he’s, ah, he disagrees with me. But then a couple hours later, I’ll back come and be like, yeah, you know what, I see how that works, and this is different, and vice versa too. It’s useful to have strong opinions, but it’s also great to be able to, with some digestion, see the other person’s perspective, and maybe still disagree, but not feel like you’re just digging in and retrenching, which is I feel like where a lot of partnerships go south.

I like discussing it straight away, but not feeling like you are gonna resolve everything in that moment. We’ve done some angel investing, and I’ve seen partnerships go south where there was some initial disagreement that either got lightly talked about, never fully resolved, or never talked about, and then it never really blows up a couple of months later.

Mike’s

get stuff done. I feel like I’m relatively smart, like I work hard, I went to a good school, the same one Mike went to, we both went to Stanford, both did the same programs. I feel like I’m pretty good at a bunch of stuff, but man, there’s a lot of stuff that I’m just not good at. Mike’s really good at it, and I respect the hell out of that. Whenever I meet people who are just very good at their craft, Mike’s craft is an interesting one because it’s not like he’s an artisan that code.

It’s not like that. His ability to pick up very difficult technical problems very quickly diagnose them and make them right is unparalleled. And Mike, I’m not just trying to blow slug up your butt. I really do mean this. And I like forming partnerships with people where I deeply respect their abilities in that way. I think that’s what it makes it really special. And then on the integrity bit, Mike likes to say I have a bit that flips of my head on people, like whether I can trust them or not.

And it’s pretty true probably if I reflect on it, but when I’m in on the people. And and I deeply invest, people don’t realize this, but I care so much about my close relationships. I’m so deeply invested in them that when people do things that challenge those, I think it can be very hurtful in ways, and that’s happened a bunch of times in my professional career. I even said this to him the other day. I was like, this matters so much to me that even if something questionable were to happen, we push through it and we try hard because it’s not worth losing those thirteen years.

I also don’t wanna take it for granted because Mike and I haven’t done a remote company before. We’ve seen each other how many times in person in the last year?

Five? Maybe half a dozen. Yeah.

Jesus. So my point is, I don’t wanna take that for granted because I think in new situations, it’s like how you think, oh, if I just hire really technical people out of big companies, they’ll work at a startup. It just works because they worked in this other context. I don’t wanna take for granted that a relationship in one context works in another context and will just work in the future. It’s kinda like what people say about marriage, you have to work on it. And I think that’s true in any relationship.

But I feel like you wanna ask about why we haven’t seen each other more in person.

there is a myth that like you can have the same friendship and or relationship ever again because things change. Mike and I both have kids. We both have dogs. We both have wives. We both have homes. None of that existed when we met. We didn’t get invited to fancy conferences. We didn’t do any of that stuff. We were boring kids who liked to drink bourbon at a local place around the corner from our office. We had nothing to do other than work, and life couldn’t be more different now.

So I actually think what’s more special about the partnership slash relationship is that it’s been able to morph through all of this change. Because people lasting one company is one thing, but like it’s one company and then beyond one company, it’s

Yeah. I believe that. I think it that there’s a tremendous amount of empathy. I remember Kevin had kids first, and I look at myself as a mid to late twenties person where my friends were starting to have kids and realize how naive I was a lot about things. The dumbest thing I ever said to somebody that was going out on leave was, do you have a bunch of books signed up that you’re gonna read during your, like, leave? And they’re like, no. I’m gonna this is a second kid.

They’re like, I’m just gonna try to stay alive and see if I can read the news of the day. At the end of the day, I feel good. I’m not gonna read a book. You learn a ton as a first time parent, and Kevin got to do that first. But I think for sure, I think the fact that we looked at each other post Instagram, and we both had that drive to still build. And I think if that were different, then I don’t think we would be working together.

I think that would have made a much harder partnership if one person was like, great. No judgment, but I wanna coast, or I don’t wanna do something, or I wanna do something a little bit less intense. But both of us love building. So when we looked at each other, we were still on that same page. But I would say and this was a really interesting difference from Instagram. With Instagram, Kevin had some initial seed funding contingent on finding a technical cofounder.

I told you. I stalked you. And Steve was like, great. You seem great. And your technical build to try to find a technical cofounder, and that’s where we partnered up. But the process from that to starting was not from us deciding to do it together to starting, wasn’t contingent on us figuring out whether we were good partners. We did some of that, but most of that was actually me getting my work visa. So that was an artificial delay. Whereas doing Artifact this time around, I think we have just grown and seen things.

Spent a good couple of weeks just making sure that we were aligned on what we wanted out of this experience, and that was really important. If we had skipped that, then there’s always a risk of one person being like, oh, I was just excited to build this thing over and not have any real pace and keep its whatever the different avenues might have been. So building that alignment was really different this time. Whereas the first time around with Instagram, building the alignment was, hey, let’s get together at coffee shops and hack together and make sure that we could code with each other and build with each other.

We took that as a given. Now it’s more where we are in our lives.

This is gonna be hard, Mike, because we don’t fail at anything, do we? Are you kidding me? I feel like every day is a constant failure. It’s Even at Instagram, when Instagram was going well, it was a constant failure. Why? I think like Why? That’s the way that companies work. Everything is always broken and you’re always unbreaking them constantly, and they keep on breaking on you. Jeff Bezos likes to say that the greatest thing in the world is that customers are always unhappy because you can always build more stuff for them that makes their life better.

It might sound a little cheesy, but it’s true. Things are just always broken, and you’re always fixing something, and I think the myth that things are ever just in the stasis of good is totally BS. And what people need to realize is that, like Ray Dalio likes to say, it’s not about avoiding suffering. It’s about learning to enjoy the suffering or choosing the right suffering because you’re gonna suffer somehow. Choose the suffering that is most enjoyable, where you get the most reward when you work through that suffering.

Again, I’m paraphrasing here. I like to say choose what game you like to play in life. If I were a basketball player, I’m assuming I’d be practicing my three pointers and I’d be running a lot and I’d be like exhausted. To stretch this to the marathon, it’s no one enjoys training for a marathon. No one enjoys running a marathon. You like winning a marathon. The idea that training is going to be easy or enjoyable, sure, hope you get a runner’s by every now and then, but you just have to get comfortable with the fact that you’re gonna be uncomfortable.

Mike, what do you think?

Yeah. I think that there’s a strong component of that. The thing I was gonna add on the failing bit is I think it actually goes back to partnerships. My wife, we we’ve been together since before Instagram. So she’s seen all of the insanity. And then I was like, I’m starting a new company or we’re launching soon. It was not, oh, don’t do it. It was, hey, let’s make sure we talk about what that means for our schedules or how we think about our time with the kids and all of that.

What commitments do we wanna maintain? So failing in multiple areas while being alone in those areas is very difficult because you don’t have any external signals to help you prioritize or to help you out, honestly. And Caitlin, I think, didn’t even bat an eye. She just rolled over. It’s like, I remember these days. Just go do this in a different room, but it comes back to you.

It’s a constant struggle, so I’m not sure I’ve figured it out. There’s something called the hedonic treadmill where you have your first, I don’t know, $40 bottle of wine when you’re used to drinking crap vodka at the local dive bar. And you’re like, oh my god, this Shiraz is so good or whatever. And then you have your first $80 bottle, and you’re like, oh my god, I could never drink a $40 bottle. And then you have your first dollar bottle, and you’re like, what the heck?

Thousand dollar bottle? And then you go to some fancy party where someone serves you a $10,000 bottle of wine. And it just keeps going, and you’re like, this is ridiculous. This is grape juice. Now to be fair, I enjoy wine, but this is grape juice. And you can go that far on anything. TikTok likes to serve me watch like, I’m not into watches at all. But there are all these watch things where people are walking around with a $150,000 Patek Philippe whatever on their wrist, and it’s just it’s absolutely ridiculous.

And I’m like, why? Why do you care? And it’s just because everyone feels like they have to one up themselves, and you’re right. You get there and you’re like, wait, I thought this was gonna be happiness. And happiness for me is when my little girl runs in and she says, dad, I counted to a 100 for the first time. I’m like, what? We’ve been working on that for months. We get to 70 and all bets are off. You got it. That’s happiness. It’s cheesy, but the little things with family and friends and man, like, money is such a siren song, but I think we’re doing this not because of money.

Of course, it would be great if we made this into, like, multi billion dollar company. But the nice thing is we don’t have to have that as the goal, so we can make building or

Potentially. I think it depends on your personality though. So there’s a difference between being pressured by others and being pressured by yourself, and I took all these like personality assessments, I wanna say five, six years ago, and that thing that stood out was that I feel very little competition with other people, but I feel an enormous amount of, it’s called challenge, which is competition with yourself. And it’s possible we would have moved more quickly to launch or whatever if we had taken money and we had VCs being like, what the heck are you doing?

And it’s possible. Where it’s possible, we’d be moving bigger on having a bigger team more quickly, but I’m not sure it’s the right thing because we are definitely not lazy. Like, we work our butts off. I’m in this chair, gosh, sixteen hours a day at least. The only reason we work from home by the way is it’s the only way to work this hard and see your family. At least that’s my opinion. I don’t actually particularly like working from home, but the fact that I get to work this hard and have a family life is incredible.

And if I were working this hard elsewhere, first of all, I wouldn’t work this hard in an office. It just wouldn’t happen. But I feel like I’m way off the original answer.

But what do you think, Mike? The question on, like, urgency, it’s interesting because it gets to the heart of our partnership too, because we both try to blend these two elements of pragmatism and perfectionism. And if I had to score us, like, on the relative amounts, like, I’d probably buy us a bit more towards pragmatism. Kevin buys us a bit more towards perfectionism. But we both try to keep both in balance. Even with Instagram, I remember we had already built Instagram, and we were in at the time you could have a 100 testers on test flights.

We had made 80 people using it. And having conversations, and we were always like, was it ready? This thing could be better. We had 16 filters. Are these the right ones? We had all these different conversations about whether it was time to launch. I know there’s entrepreneurship cliches. If you’re not embarrassed by the first thing you ship, you waited too long, which like is maybe a useful thing to put on a poster, but in the details, it’s a little hard to say. It’s like, do we have this feature ready?

I look back at view one of Instagram, I was really proud of the polish around the core features, and there was stuff that was comically off. The second something got more than 30 likes on a post, it totally broke the UI on launch because we didn’t summarize it, so you would just have everybody’s names. It would be a wall of text this big. All sorts of stuff that this is, in some ways, table stakes. But on the hand, there’s probably another week or two of dev, and I think timing can matter, and getting out there and starting to react from actual customers really matters.

And then with Artifact, the interesting thing is, I feel like core UI, we probably could have shipped several months ago, but this product lives or dies by its recommendation. And so getting the recommendations into the spot where we were, like, ready to bring things on, bring more people on was the key. So I think you’re always struggling with this question of, are you ready? Do you go? I think this time around, it was not a function of urgency or work ethic. I think it was a combination of more work just needed to be done to feel ready on the ML side, but also, man, it’s our second time around.

And as much as I’d like to pretend that this thing will be judged purely on its own merits, there’s both the pro and the con of it. It’s Mike and Kevin’s second thing. It’s great to have that attention, but it also gave us some pause around, yeah, we’re like, we wanna shift something that feels like it’s thought through, is ready, and people can use it day one and not feel like, man, they really phoned it in, or man, what have they been doing? And I think that was a helpful aspect of trying to calibrate how ready do you wanna be before going out.

Given

single day, I still feel it right now. First of all, we overbuilt the product. There’s parts of this product that exist. You can DM people. You can post socially, like Facebook or Twitter. There’s entire pieces of this app that people can use. It’s just locked away and not launched. Mike is being miced. We could have launched a year and a half ago. How do we not build that stuff?

He’s No. In fact, I was the one who was like, we should not launch this, and the entire team’s like, you’re fucking crazy. We spent all this time building like, are you insane? That was one of those moments I think I got Mike faced for thirty seconds. And then when it dawned on all of us about how hard it was going to be to both launch this product and do moderation at the level we’re gonna need to do with the amount of attention on this product, I think it made sense.

But my point was more that if anything, think we should have started with the kernel of the idea, launched that thing really quickly, and then gotten feedback. And what we’re seeing now, we chose the worst idea by the way for a second company because the characteristic of this company is that you need lots of data to make the recommendations good, but you can only have lots of data if people like your product and come to it and use it. And there’s all this pressure that the thing you put out definitionally is gonna suck day one.

Do you have to take this pain that a lot of people are gonna sign up and be like, I don’t get it. I see like a feed of news that I can get anywhere else. Why is this a thing? The good news is that a good portion of those people actually come back every day, and a good portion of those people are producing data that makes the recommendations for the next set of people really good, but there’s a big portion of people who there’s a big effing question mark over their head when they sign up day one, and that’s painful for us because we wanna build a product where people sign up and they go, oh.

Because they did at Instagram. They signed up and they went, wow, filters.

think what people don’t understand is we’re not building a new service. Like, maybe I shouldn’t be saying this as much as I say it. Well, no. I’m interested in core recommendations. I’m interested in the science of utilizing people’s tastes and profiles to make sure that the world serves them. And I’m not talking just like Pandora, and I’m not talking product recommendations. It’s all of the above. The idea that you can take behavior or data and use it for a person’s benefit rather than for the company’s benefit is fascinating to me.

And it just so happens that news is a great place to start because there’s a ton of information out on the web. Machine learning is very good at analyzing it, and we can build portraits very quickly. By the way, this is the crazy part. People like to think they’re really unique. They’re like, oh, I’m like super unique. I have all these super unique interests. The core thesis behind recommendations is actually there are like groups of people that tend to think a lot alike, not just ideologically.

We’re not talking, like, left, right, politically. I’m saying, like, if you’re in San Francisco, we can almost guarantee you’re gonna enjoy reading about startups from TechCrunch. It’s not that hard to guess. You’re generally not wrong. Or if you’re super into the Atlantic and you’re reading a bunch of mental health and wellness stuff, we know there are a bunch of associated categories with that and publishers with that. So your initial recommendations actually can be pretty spot on, which is pretty crazy. But to get to the level we want to get to, you do have to be much bigger.

Where my interest in hi fi audio systems is represented or Japanese architecture is represented, we’re not quite there yet. Basically, we chose an idea where the ideal, we know what it needs to be, but bootstrapping it, that activation energy required is just, really hard. But because we are who we are, we can get some amount of press. We have hundreds of thousands of people sign up the first day. Enough people can get in that you can bootstrap it, where if you’re just some pair of this is not a company we could have started out of Dogpatch Labs in 2010, that we couldn’t have done it.

So it is an unfair advantage to get some eyeballs. The hope and the bet is that you can take that initial attention, parlay it into just good enough recommendations for a certain set of people, and then basically bootstrap and expand, and then hopefully we can expand our purview over time.

theory where basically products are hired by the consumer to do something for them in their lives. It’s almost like having a digital assistant, and you have to ask why your product’s getting hired. And if your employee gets new skills that help you do that job or similar jobs better, then you’re excited. But imagine having employees who’s like, hey, I’m gonna serve you really entertaining photos and videos from your friends every single day so you can feel more connected to them. And they’re like, and now you can shop.

You’re like, hold on. That’s that’s not why I hired you. Or if it feels unrelated, it might work if you’re big enough, but at the same time, it might confuse people. And I think people can only hold a complex enough idea of your product in their head, but it has to be simple. And if you complexify it, complicate it, they just generally don’t want to open you. It’s almost like in music. I’ve been reading a lot about music, and they talk a lot about if you use certain tones, people will just stop listening to your music.

In lo fi music, there’s in chords, there’s the thirteenth. The thirteenth is, oh gosh, I’m gonna every music lover’s gonna crucify me on this. I wanna say it’s thirteen seventy nope. Not thirteen seventy tones. Anyway, it’s above the root chord by steps basically. And it sounds dissonant, and there’s all this data that says if you include it in your song, because you’re playing this certain chord, like it sounds dissonant and people stop listening. And I feel like products are that way too. They can just grate at you if they include things that are dissonant to the core job, so you need to be very careful about forming that core job and only adding things in when they support the core job.

But right now, we’re about two weeks old. We’re gonna be like that core job and that core only. We are vanilla ice cream, but we’re gonna be the best vanilla ice cream you’ve ever had. You’re gonna have toppings later, because toppings are great, but like, we’re gonna get to the toppings. I mean, saying we’ll get into the toppings How do I get an award? I worked so many metaphors in there. There was there was the sprinkles. There was court.

you. If you visit our app store page, what does it say, It’s like personalized news for you?

Yeah. Exactly. Artifact personalized news.

So we’re very clear to people what we are today, but it’s almost if you listen to music versus you’re talking to someone who understands music theory, you just talk about it in two different ways, and you have to understand our ambition is not to make is not just to have a fun news app forever. We think fundamentally machine learning is changing the way people consume social products, and we wanna ride that wave and build the best social product using machine learning we can. But it’s just so happens that we think there’s this enormous unmet need in news that we can start on.

Every app that is successful in social starts off as a utility, and it becomes a large social behemoth over time. Now, I’m not sure if that’s our future. Maybe we’ll take a different turn. Maybe we’ll go in a different direction. But when they started Apple, if they had just said, we make desktop computers, imagine how limiting that would be. But everything I’ve read is Steve was like, we make great products, which is like very broad. But you can’t tell a consumer, we make great products. Okay.

What are you gonna do for me? But you can have that ambition, and that’s, I think, the contrast that I’m trying to highlight.

I think that’s also where launching helps. As I would describe the product to people over dinner, in the months leading up to launch, I’m like, people are made up of a lot of different interests. We had a really fun World Cup thread going inside the beta, or maybe you’re really into a particular type of architecture or a particular trend in fashion. And that’s a lot of sentences, and the App Store gives you 200 characters. So it was actually very clarifying to be able to say. And I would actually go out of my way not to call it news because I think I think people are made up of a plethora of interests beyond that.

You also need to be able to explain things very simply in the App Store, and you have to pick a category. There’s no category in the App Store for apps that are generally about your interest in content. It’s there’s a news and magazines one, and that is the thing that most closely describes the core of what we do. So I think that’s always that sort of play around your full ambitions or even the texture of what you’re doing, and then also, great. Give me the five words that describe what you do quickly.

And you kinda have to have both, it was like a video game. Good video games, you start with the basic sword, and you can’t do very much. And then as you explore, you’re like, oh, actually, now I can do this other thing. And then you level up and you go through it. I think good products let you build into having more advanced functionality.

Abandonment, I feel, is a very binary way to look at it. I don’t think you’re gonna see a complete abandonment, but I will say, like, why do we use social networks initially? Because social networks were great filters. When I say social networks, the people you follow on a social network. They’re great filters, first filters for content that’s relevant to you. And that fundamental thesis is that your friends are into the stuff that you’re into, And I would argue it is the greatest mistake of social networks in the last twenty years is to believe that is true.

Because it turns out Mike is into very different yes. I like electronic music, Harry. That’s true. Mike, you like f one. And if you show electronic music stories to Mike and f one stories to me, that’s terrible just because we’re friends. By way, our social graph couldn’t be stronger and more overlapping. But your content is terrible for me, Mike. Now, yes, of course, it’s median voter theory. It’s this stuff in the middle that satisfies both you and me, Mike. It happens to be tech news. It’s like middle of the road stuff on chat GBT.

Great. But what about all that, like, amazing texture that makes us people? That’s the thing that I think TikTok or ByteDance, I guess I would say, unlocked. With TikTok, they realized that there’s so much more to people than just their social graph. Your social graph is, of course, a useful signal, but to say that is the core signal, and when I say core signal, I log on to Facebook circa, I don’t know, 2006 or whenever news feed came out. I can’t remember. 2006 is fine because seven was the iPhone.

You log on and you’re seeing posts from your friends. And so, okay, I guess that’s relevant if I’m in college because I wonder what my friends are up to this weekend. But over time, that doesn’t scale. Over time, that becomes weird when your angry uncle is posting about Trump and you’re just like, okay, this isn’t for me. Long story short, I think that social networks need to become less social, and I think that something that will make up for it is this idea of learned association or learned affinity through your behavior.

And I hope what we will be able to do is show that algorithms can be used to work for the user rather than for the company, and that will end up being a great experience long term. There are all sorts of things you have to watch out for in this direction. The same way there are things to watch out for in social networks around filter bubbles and around some user having way too much distribution and using that distribution to post things that aren’t true or There are all sorts of issues you have to work through.

It’s not squeaky clean, but it is very exciting because every network is moving this way, but it feels like when networks were moving to mobile in 2007, 2008, they half stepped. Remember Facebook had a mobile app that was like actually written in HTML, and it wasn’t really native, and then all these native apps came out during social networks, and was like, oh, that’s what this should be. I feel like machine learning is exactly the same wave we’re seeing right now. It’s a parallel wave. I don’t think it’s a complete annihilation of the social network, but I do believe that unconnected content is incredibly important.

Mine, I think what doesn’t go away is the desire to talk about it with people you care about when there is a connection. I think inverting the discovery versus sort of conversation piece is really key. What determines what an artifact or other things that are recommendation driven is not who you’re friends with or who you happen to follow or who you knew in college. It’s we think is relevant to you. But sure, I’m really interested in housing policy in California. Whenever I get a great housing policy article, like, instantly, I have a housing policy group in the DM chat.

But even for the people that aren’t in social area of the product yet, we’ve seen people sharing it out to Twitter to start a conversation or sharing it to a WhatsApp group and have a conversation. It can build relationships and strengthen them once you have more things to talk about. And then I have a friend who we’re both really into the warriors. And pretty much once a day, something will come through my feed that’s about some either warriors rumor or story, and it’s just like another moment to have another conversation with him.

The social component saw the very strong conversational piece. It just doesn’t need to gate discovery as it has traditionally.

I was gonna say, one of the questions you had sent over earlier was like, man, isn’t news and actually, no. You said social. You were even nicer. Isn’t social like a bad place to invest right now? There hasn’t been a new social thing really, what did you say, in last ten years maybe? Yeah. Yeah. I said I said it was be real, but I don’t know. I saw some headline this morning on Artifact about how like be reals are down 95% or whatever, and I guess, like, social gets a bad rap.

This is gonna be heresy to a VC’s ears, but I think investing in verticals and spaces is BS. Yes, in general, generative AI is gonna produce something pretty cool. Yes, These verticalized, like, the sharing economy. It’s like, okay, cool. That was a while ago, but I’m just naming these trends that have come and gone throughout the years. A 100% of this is team. A 100%. You can take a team of people and say, hey, go build a new car car company, and you’re like, oh, cars.

But I don’t know, with the right team, with Elon in charge, it’d be worth a lot. Cars are inherently unsexy as a business, and I guess the last chapter hasn’t been written on that valuation. My point is, I think almost everything here is about team plus the area. And what’s cool about news and social is the general let’s talk social for a second. The general lack of the skills that it takes to grow social things. The one thing that I think Facebook should be remembered for as a professional development like incubator is its ability to teach young PMs the scientific method of growth.

How do you grow things? Maybe to a fault, by way, with some of the tactics. But using the scientific methods, run AB tests or multivariate tests, using statistical power, using key values to understand whether or not you have significance in a test and whether you can loop your way towards better things in social. Man, that is so important, and it is so hard to just get a group of people together and get lucky like we did with Instagram, because we had no insight into any of that until we got to Facebook.

My point here is that with the right team and the right skills and the right history, you can make pretty, quote unquote, unsexy areas become sexy again. I can’t tell you how many people when we started Instagram were like, images? There’s this famous graphic at the time that had money amount in exits in video versus versus photos. It was you have this huge YouTube one and then nothing in photos. And it’s like, okay, you can either read that as that is a pattern and that you can extrapolate that or you

If there was a VC that was simply focused on discovering amazing talent rather than special ideas or special verticals, I often think you should ask Steve Anderson this. Was the test, hey, I’ll give you 50 k for this company if you can find a technical cofounder. Was that him being like, you need someone technical, so, yeah, that derisks the idea. Or was it a test of can you attract the right person? Can you team build? And is this person good? Can you find a really good person as your first hire?

And I got really lucky finding Mike in that coffee shop that day. And then the fact that Mike didn’t go running for the hills when I told him about the company we were building and he was excited. I’m not putting it all on me because of course, Mike helped build the company from the beginning as well, but I do think there’s a an important quality in early teams, which is can you form a great team? It’s why you see so many ex Stripe people who are fantastic go off and start their own companies or ex PayPal people.

Like, I often say to Mike when we’re hiring or firing, do you feel like this person three years from now can go start a great company or go be the CTO or COO somewhere? And if that’s not true, why are they here? And that’s just the common element with every great team from the beginning is they attract that group of people.

I think prelaunch, it’s funny because amazingly, the beta was really quiet. People signed an NDA, but it’s only as good as the person decides it to be. So it was really quiet, which actually meant that recruiting wise, we leaned on our current network. That has pros and cons. I think they’re wonderful people that we worked with at Instagram, and it’s really different in the startup. There’s high variance of outcome, and there’s time that it takes to get there. I honestly just in full transparency, learning to screen for is the people that will thrive in that environment or at least tolerate it enough to get to the other side versus wanting the solution or wanting the fallback.

We did a six month performance cycles at Facebook, at least they were when we were there. I think they’ve changed this a little bit. We’re on six month increments. People would sometimes be like, I’m signing up for a six month project. If it doesn’t work out, I can apply for a transfer. It’s this is really different. We’re in this together, and we’re in this for the long run. I think we’re learning still about what it’s like to have that conversation upfront and to have people join for that reason.

And then the other piece too, and this is a skill that I learned the first time around at Instagram, is tailoring the solution to the problem space. There are ways in which building for an Instagram, not just from a community size perspective or from a team size perspective, are so different from our team has seven people. We’re just getting started. There’s pieces of process or technology that are the right thing for your multi office 2,000 engineer team that would just not feel right here. So it’s about rightsizing.

And honestly, even for ourselves, taking what you know, which is all the accumulated bits of knowledge both on team growth and on the technology engineering side and filtering it through, is it right for our stage? Is it right for where we are? And can we take the best of that without all of a sudden acting like we are a team that is either much further along on the journey than we are. And we tried running an AB test on our beta group, and then two weeks later, we were like, let it run, and we did all the math, and we’re like, to detect a stat sig difference in this beta group, you would have had to basically make the feature twice as good, effectively.

There’s very few things that you’re gonna ever test that are that good. When we launched Explore and made it personalized rather than just the most popular photos, like, that was a two x delta in usage. There’s very few of them. Stories, equivalent. So it’s actually been really fun to now be public, and we ran our first test and actually got stats like results. We’re like, okay. You have to remind yourselves, like, what tools and techniques from this toolkit that we built up are right for where we are now?

And that’s the question we’ll always have to ask, and it’s about growing on that journey with our team too.

I think your your premise is wrong though. I think that, of course, if you’re gonna get married to someone, if you’re gonna start a co founding relationship, you can’t fly through co founders. I guess maybe some people do. I think the premise or the constraints that you have to 100% trust every single person is just not right. I also think it’s right, like in hiring, you just need to hire a bunch of people and realize your hit rate’s not gonna be a 100%. That’s okay. This isn’t a forever thing.

Who holds a job forever? No one. Very few people. And instead, just optimize for minimizing the chance of getting it wrong, subject to moving quickly enough. And if it’s wrong, move really fucking quickly about making the change. And I think actually that last part is the part people have the most problem with, not actually trusting people. It’s when it’s not working moving quickly because it’s so hard to screen at a few interviews and a coffee or whatever whether or not you can trust someone for life or whether they’re the right person for this chapter of the company.

It’s just not possible. So you just gotta move on it with the best of intentions on both sides and move really quickly if it’s not working. Did you move quickly when it wasn’t working? I certainly tried a bunch of times, which I think there are probably like three examples in the last year and a half I can name where that happened. Yes. And then there were times where we didn’t or I didn’t, and I think we regretted that. It’s just a lesson you’re learning over and over again running a company.

You need to be a compassionate leader. You need the right person in the right spot, and you need to help them be great. And if it’s not the right fit for whatever reason, you move on. I think that’s something a second time founder will do that a first time founder has a lot of trouble doing.

You have to go against basic human psychology and empathy, because you think, like, your first instinct maybe as a first time founder is, oh, the compassionate thing is, even though it’s not working, we’ll try to make it work and have the fifth conversation. And then the second time, you’re like, maybe the empathetic thing is realize when it’s not working and make that move.

Can I talk about music for a second? Because I have to bring it in again. I heard this thing on NPR this morning. It was so fascinating. I have to find this paper. But there were these researchers. This is one of these tough pieces that NPR throws in the middle of their morning show, but it was actually super interesting about how basically these researchers figured out how much of a clip of a song they have to play you before you can form a very strong opinion of whether you hate it or love it.

They found that the shortest they could do was five seconds, but I’m gonna quote this wrong and I need to actually find this paper. But it seemed like to me the difference between five and ten and thirty seconds was very small. All you need is five seconds, and you’re like, I got it. This is great, or I got it. This is not great. In general, with talent, you can tell very quickly in a working scenario whether someone generally is gonna work or isn’t gonna work. There’s that middle ground with where you don’t know, but in a small startup, I feel like it’s fairly short.

Just getting a better spidey sense for detecting that in the hiring process is what keeps you from having all this pain of churn throughout time. Isn’t it like a conman? In general, your goal as a founder is to figure out how quickly you can detect whether or not someone’s gonna work in a role without having hired them yet to save them the pain and you the pain of having to make a hard change. But then being a founder, making the hard

Harry, sometimes that’s a feature, not a bug. Sometimes you want, like, someone who’s not highly volatile cracking jokes all the time. It just depends what you want. I don’t know. Like in your scenario, if you have to spend a bunch of time with them, travel with them in a small office, then I guess personal chemistry really matters. But Mike, we’ve worked with plenty of people where we didn’t have personal chemistry, but they were great at what they did and that’s fine. And in fact, as a small company, one one of our mistakes in Instagram is hiring too many mission hipsters that wore plaid and really like drinking bourbon.

That was stupid. We should have just been hiring the greatest people out there. I think it depends on the the size of team and the dynamic. I don’t think being a fun person should be a requirement for a job. I think it’s okay to be boring.

Mike, do you agree? Or I I like Yeah. I think it goes it comes down to you kinda want the team composition to have at least a sizable group of people where you draw energy from them. I think if your whole team is made up of people that you’re like, oh, they’re fine, but, like, they’re good at their jobs, but they’re not, like, sparking energy, I think your own day to day is not gonna be exciting. But I don’t think that has to be everybody, and I think I think you can find that balance around.

Some people are quietly executing and doing that, and they’re probably not gonna bring you an idea that’s gonna blow your mind, but they’re executing really well on the things that they’re doing. I would just be wary of an entire team of that would probably be not super fun to work with.

In

I think the different phases of the company require different kind of people. I’ve always been happy with people who have the sort of raw talent and desire to learn, and don’t see boundaries as a reason to stop working on something. That’s a repeated theme. The people I love working with are the ones that don’t come up if they’re iOS engineers and they start having to do back end work. They don’t throw up their hands and go, that’s not my problem. Please help me here. It’s the people that you’re like, you know what?

Like, one of our engineers was like, oh, I haven’t really done marketing stuff before, but somebody needs to put together our Google Play Store page. I’m gonna go learn what other companies are doing. I’m gonna put together a proposal. Like, that desire to not see your role as strictly defined by, like, the four bullet points that you think you do, That’s to me the kind deep I repeated working with. Now, once you scale, there’s sometimes reasons why there are boundaries between teams. There’s more process. You wouldn’t want your excellent mobile engineer be like, you know what?

I rethought how we should do our scaling when you’re, like, serving over a billion people. So you start creating more of that, but especially in our scale, and honestly, the next few years, it’s raw talent, ability to learn quickly, ability to not define themselves strictly by the role that they were brought in to do, and that leaves you out. But if you have only that and you don’t have any people who have expertise, it can also lead you down this other side of things where you’re like, just be enthusiastic about an area isn’t necessarily gonna determine that you’re gonna be able to be good at it.

So it’s that community of expertise, interest, and selflessness is a kind of core combination that I seek.

Definitely more breaks throughout the day and more thinking about what am I building for on a twenty year horizon and less sleep.

Kevin? Prioritization, only doing the shit that matters because you’re probably gonna get cut off in the middle of doing it anyway. So, you know, reporters in the old days, you’d like, you have to write the most important stuff in the first paragraph because it might get cut off. Basically, doing that in life because you might get cut off. So my son last week was running around. I was in the middle of something super important, and he just fell and busted open his chin and needed to go to get stitches, and he is fine now.

He got three stitches. But my entire day, gone. Just because I had to go do the

It’s it’s gotten to the point where it’s not fun to have the argument with Dave anymore about this. But he doesn’t live in F F, so of course he feels that way. And I live in F F, so of course I feel that way. I think San Francisco has been through so many downturns. And if you look at this one alone, you can draw the line down and say everything’s gonna just shrink. But it has been through so much and it always rebounds and comes back way harder than it was

true

I think the best universities in the world generally for computer science are in this area. People passionate and making next generation AI algos and stuff are all from this area. They all mix together. Literally everything interesting is happening right here. I guess if you consider Bitcoin and crypto super interesting, I guess you can bet on Miami. But I’m gonna bet on the next chapter of companies starting their core here. Though I will say the biggest difference going forward, you’re gonna see huge remote component, especially for your most talented people.

They’ll be able to work basically from anywhere. And I think that is a fundamental difference, but there’s no way San Francisco is going away as a city. And it’s like the center of great startups. Basically, bottom line is that I think it’s easy to be anti San Francisco right now. But mark my words, look back five years from now, and I we are even seeing the glimmers of it now. Just all these companies getting started. If you look at history, what is it? It never repeats perfectly, but it rhymes.

And it’s just there’s not a single time where San Francisco hasn’t come back with a vengeance. Boom and bust periods. There are cities in The US where fundamental changes have happened to their industries, and the cities have gone away, Detroit being one of them. Prior exists for that being true. I cannot make the same parallels today for remote work or any of the shifts that have happened during the pandemic that you’re going to see that fundamental shift to the community here, which is nerds. It’s foggy.

It’s cold. You get to work hard. There are no distractions. We only have three restaurants and stay open past eleven. I think there’s one club in the city. There’s literally nothing fun to do past 9PM in the city, so people work their butts off. And I think it’s a great center of that. It’s the kingdom of nerds, and I love it. You literally have dinner at 6PM. It’s bizarre. It’s a Yeah. It turns out it’s a lot like what having kids is.

The SF thing is one. We stayed here and doubled down on San Francisco. Goes back to the brilliant jerk thing, like, that you can tolerate some amount of the wrong person, but right for the job, and it’ll get you further along. I keep learning this lesson. It doesn’t work, and you need to cut that off, and it might move slowly in that first week or two, and then you’ll go forward, and then you’ll build from there. The other one is, and it’s ironic as a second time founder, is that past success is a prediction of future.

Like, you have to work even harder. You have to be just as thoughtful about the team. You have to work to both fight your own bias. Have I done this before. No. You earned yourself the right to be in the conversation. It doesn’t predict anything about how well you’ll succeed.

challenge that with saying that with the lens of seeing my friends who haven’t had billion dollar exits, who are perfectly happy with their families growing them, maybe not in Downtown San Francisco and a high rise or whatever, but actually the thing that matters most is that you meet a partner or a group of friends that carry you throughout life, and then you grow a family, or you choose to live the life you wanna live. And I see people who are VCs for a little bit. They make 5,000,000, and they retire to the Sierra Nevada, and they buy a huge plot of land, and they live like kings and they’re super happy.

I just don’t actually see a meaningful difference in happiness between any of these people as long as you’re just enjoying your life and doing what you love. So forget about the outcome. I don’t think people should worry so much. That being said, there’s a lot to worry about to get to the level of being okay. So I wanna be clear. There are a lot of people in the world that like work three jobs to support a large family, and I get why people worry. My point is, if you’re 20 years old and you’re working hard, you don’t have to worry.

Are you gonna become director by three years from now? Are you gonna be on the cover of Forbes? Are you none of that stuff matters. So I just I think worrying a little bit less. There’s this great article that I was served on Artifact yesterday about doing things that are a little bit more irrational, a little bit more often, and I agree with that. Just don’t always do the thing that makes complete financial sense or complete career sense. Be a little bit random. And actually, machine learning would tell us that, which is that being a little random on the margin is actually optimal in the long run.

So have fun.

new friends? I’m sorry. But if you’re a parent, I feel like I get told this the first five years of your kid’s life. It’s just literally everything else goes away, and you’re just like changing diapers, going to the hospital, giving drugs, waking up at 4AM. What are new friends, Harry? I like, I’m excited for the moment where I come up for air and you start hanging out with people again. Please tell

If it feels hype y, if there’s at all a glimmer of, like, I don’t love this area, but lots of people are excited about man, I looked back. I did a little reflection. It was, like, mid last year. I was like, there were, like, a few either LP positions I took or individual investments. Was like, man, I did this because it felt like this was a thing that a lot of people were excited about. And I wasn’t I have to keep relearning that lesson. And my life is relearning lessons.

I think everybody’s is. And hopefully, you get better at, on margin, making fewer of the same mistakes over and over again. So that’s a really big one. And two, it’s such a cliche, but there’s the founders where you’re like, they were gonna run through the wall of any problem that they hit, and then the ones that are like, oh, actually, they’re really polished, and they came out of a Facebook or an Instagram, and they’ve got like a decent first demo. But you ask yourself, the second something goes wrong here, are they the kind of person that are gonna be able to make the hard decisions and move quickly?

Like, have they had that adversity in their careers? And or can you imagine themselves in that adversarial situation? And, again, I look back, and again, so so much of this is a hindsight thinking, so you can learn less as you wanna learn, and then try to adapt them for the next round, and we’ve been investing way less now that we’ve been super busy with Artifact, but yeah.

someone leave Artifact a while ago, and the reason was I feel like I’m being told to do my homework before I can play. And I was like, hold on here. Of all the things that have literally ever been said to me in my career, that was that tops the charts of millennial.

to me, it seems so out there culturally when you join a company that I literally spend my day working. I don’t get to play. I wish I could play. I don’t get to play. Who gets to play? But actually, I realized in the hindsight, it’s a totally fair comment because there actually are lots of companies that pay millions of dollars to play. Because on the off chance that one of those playful ideas becomes the next big thing, they’re willing to spend that. But at a startup, you can’t.

There’s no playing. You work hard every single day to get stuff out. But I will tell you, it’s changed so much and But I feel like every generation has complaints about next generation, so I think it’s a bit trite to have a complaint like that. But you just you gotta scream for people.

And it can’t be about what espresso machine you have in the office and what perks you have, and those things are all fun, but there’s that famous wartime general, peacetime general thing. Man, this stuff’s just hard, and that’s why the startup process just chews up and spits out people. No process I’ve seen before. It is just a grinder. It goes through people like crazy because it either screens you out because it’s too hard work wise. The trade offs are too crazy, or the volatility and the emotions are too much.

The volatility and the, like, we’re gonna succeed, oh, we’re gonna fail, oh, we’re gonna succeed, oh, we’re gonna fail, and not knowing what is true. That volatility can kill people. We can just chew them out and spit them out. You need to be okay with the fact that this stuff happens, but Mike and I have the, like, amazing advantage of literally having thought this way before and just knowing there’s a saying in flying. I learned to fly after leaving Instagram, but just trust the instruments. If it says you’re sinking, don’t think about how you feel because you’re gonna feel you’re like you’re leaning left or you’re rising or falling.

Have you ever been in a plane with the window down and you feel something? You’re like, that feels weird. Just trust the instruments. My god, we have instruments, dashboards, etcetera. Just trust the instruments. Forget about what people are saying. Forget about what issue you had yesterday with the server. Trust the instruments. It doesn’t make it any easier. You’re looking at it and you’re like, Mike, let’s trust the instruments. You’re like, I don’t want it. It doesn’t feel good. But if you just learn to trust the instruments, you’ll be okay.

And so Mike Maples has the best term for this. He calls them ZIFs, zero interest founders, and I love it. But zero interest founders, it’s amazing. It’s like you’ve only experienced life in the zero interest market environment. Everything is free. Capital is flowing. Everything is great. Man, there’s a reason why people who have seen things in history do better than those who are seeing things for the first time because you start to learn there are patterns. And accesses, Mike, you were like, oh, if I’ve done investments that feel trendy or hip, and I was like, literally every VC’s crypto portfolio.

Literally all of them. People are like, I don’t know, but it’s blockchain, so I’m gonna throw money at it. And it turns out that some companies might actually succeed and do very well in that category. But you need to be careful about, one, riding these trends and two, just extrapolating the present because zero interest doesn’t last forever, and then all of a sudden musical chairs, like, music’s over, just take your seat. There’s not one left for you. But that’s why having this advantage of being through these cycles before it helps.

But, man, you make the same mistake over and over again even though you’ve been through it before, but you just you’re quicker to fix it because you know you’re making a mistake.

an espresso machine on our first Instagram office. That was the first big purchase I made. Remember, Mike?

Absolutely. We had it in there. We had a copy, and he taught us how to pull a good espresso shot. So important. He’s in remote work. Not

I’d love for it to be among a portfolio of product solving fundamental question of how do we help technology serve you and tailor it to you and make sense of the craziness of the world and the complexity of the world in a way that makes things feel fun and simple and tailor made. I’m

gonna only talk about Artifact and forget about all the other stuff that I wanna do because Mike just covered it. For Artifact specifically, I want a world where publishers trust technology companies or at least this technology company where there’s a partnership where it’s, I’m gonna say win win is a cheesy term, but where we build something great and we’re able to support the publishing industry, but also maybe spawn an an entire generation of independent publishers and give them access to a large audience and to publish to a world that they wouldn’t have been able to publish to otherwise.

And we create a system where people look at algorithms and portraits and all this stuff, not with a skeptical eye, but with excitement, and they see what it can do for them. But, man, if there were two rivers to swim hard against, it would be trusting algorithms and the news industry just like all the challenges. But you don’t take on small challenges when you’re a second time founder, so I’m excited to go after it, and I think it’ll take at least five years to get even close to that point.