Culture & Community as Complex Systems – San Francisco 2025

So, I'm Carter. Um, I've known Beth a long time. Thank you for having me. Um, I have worked in community management and content moderation for nearly uh my entire career. Um, I've worked at startups um doing um art founding.

I've worked in video games. Um and at Google I started our um sort of internal uh community management team which makes sure that our open discussion and productivity um can continue to exist um at Google scale. Um it expanded very rapidly. Um and I care very deeply about community as a common infrastructure uh to get us to find new ideas. It's a great community of people that I see here.

I'm sure we'll get into a little bit more but that's that's who I am. Awesome. So um I titled this talk um community what is it culture and community is a complex system and I just wanted to provide a little grounding um two different definitions of a complex system that uh kind of provide a framing for what I mean by that um one from uh University of Wateroo's Institute for Complexity and Innovation um complex system as a holes that are uh more than the sum of their parts whose behavior cannot be understood without looking at the individual and component and how they interact. And then from Santa Fe Institute, convoluted exhibitions of the adaptive world from cells to societies. And I think as we are thinking about um what does it mean to build out the commons both on a technical and digital infrastructures level on the level of community building and care work on the level of governance practices and how public policy interacts with that I think you really can only understand um the commons and ecosystem building in our complex world as a complex system.

I think it's super interesting to hear the perspective of Carter um who instead of working in this kind of decentralized nebulous like entangled um world of open source projects that many of us are involved in then um he has been community building at um institutions of scale. So, I think it's really cool to hear about um some of the evolutions of governance practices, self-regulating um like ways of evolving and adapting um governance and rule building systems, what that means for content moderation within communities and what we can learn from the centralized world to um make some of our practices more resilient and sustainable in the decentralized chaos system that most of us are operating in. So, uh, yeah, to start off, I would love for you to tell a little bit about your evolution from a active community member, um, to somebody that's building out some of these structures from the inside, like what that looked like for you, what your motivation has been. Yeah, I um I had the the benefit of I we're both from the South. Um, and I had the benefit of growing up as an outgay kid in the South um, which was not fun.

Um, but this was around the time that the internet was truly starting. Um and from a very early age I sort of saw the benefit in safe online communities to bring people together. Um I immediately saw that you had unregulated uh massive communities at scale that were really really unsafe. Uh the AOL chat rooms that anyone could join. Uh but then I saw the benefit of these more regulated uh more structured smaller forums where people were really coming together to help each other and that was my first exposure to online community.

Um, from from there I, you know, I I went to college. Um, and this might sound a little bit strange, but I became like a Google+ influencer. Um, who remembers Google+? It's a thing. Um, I I I got over a million followers there and I got a taste of what it meant to be very public in the eye.

And that's sort of where I saw my community dissolving around me. um where it was no longer a community of like shared ideas, but rather like a a personality, an influencer type thing before the Instagram influencers were really like a this is like when clout.com existed, which was very strange. Um and so from there, Google hired me to run Google+'s content and community management team. Um, so my role was to make sure that people were having a rewarding experience, that you were seeing information that was relevant to you, that you liked it, um, that it was interesting, that we could encourage more content.

And that was a fantastic um, fantastic job. Um, I absolutely loved it. Um, during short shortly thereafter, Google's culture failed at scale. Um, like our internal culture failed at scale. We went from like 20,000 employees to like 189,000 over the course of just a couple of years.

That's ridiculous. That's full-time employees. That's not TVC's, tempers, and contractors. That's full-time Googlers. Um, and some of you may remember sort of uh this like women's walkout moment.

Um, the James Deore moment. Um, the Wired article that said, uh, three years of misery inside of Silicon Valley's happiest company that anonymous people mailed to me. Um, I think, thank you. I'm not sure what that was about or how they got my address, but they they did. And you know, I sort of looked around and I saw, wait a second, this centralized group of people, this small community with uh people that had high trust and shared purpose is no longer that.

And we can't rely on it to be that anymore. We have to change the way that we think about our online communities at Google um and do something differently. Um Google is unique um in the sense that you can talk about almost anything. Um there are tons of communities literally hundreds of thousands inside of the company that are workrelated and non-workrelated and productivity based and non-productivity based and that's what Google values other companies um in Fang, Apple, Meta, they're not the same. Um Netflix is so but Google was like we need to retain this.

So, how do we evolve um our you know ways of operating to make sure that these spaces remain safe and productive? And so, today I I lead a team called the culture technology team and our job is to make sure that all of our internal platforms, policies and systems reflect the employee engagement strategy that Google wants to have. Um that sounds uh you know interesting in itself, but it's just cultural evolution. Um so, how does our code reflect the culture that we want to have? Um and that's what we build.

Yeah. So I think um tying that into this complex system definition of understanding the whole by understanding each of the components of the parts. One element of that that's super interesting to me is like you know if there's a small community of like you know 20 people that are excited about bringing their docs into work versus a you know group that has like 50,000 people of like folks that are in um the Midwest or I don't know this is what I imagine you guys groups to be. I have no idea. But um yeah, how do you h u make sure that the same um like principles and practices and you know governance procedures and content moderation procedures fractalize up from the very smallest tight-knit group maybe of folks that see each other in the office every day or folks that have like an interest in common or work on a team together.

How do those um like how are they symmetrical to the larger mission and the you know complete group that everyone at Google constitutes or maybe there is a difference in there and kind of where is the breaking point of what is it called the Dunar number of like how the groups um yeah like how similar that ends up being. Yeah. So um I think you know community can feel very uh nent at times of like well what is a community? How do we get people together? How do we know when we have a community?

Every company is asking this. Every company wants to have a community to serve some sort of purpose. Um, and what's helpful is to remember that communities, like anything else, also follow like a series of rules. Um, so as communities grow and expand, they eventually form subcommunities, right? Like if you're talking about one kind of electric car, maybe then you start talking about another kind of electric car and all of a sudden you have like an umbrella community, but you're talking about different things.

Discord is a great example of this. Another law is that eventually that people start to conform to the communities that they're a part of. They start to talk a certain way. Um they have like in language and inspe they understand or in jokes. Um but another one is that every single community as it grows bigger needs it naturally forms hierarchy.

Um and I know that we're talking about a lot of like uncentralized stuff here today and I hope this isn't controversial. Um but you know eventually you do need a form of hierarchy to keep yourself focused on a purpose. um at Google that's what we realized um was that you know we had been sort of very trusting um of our employees up until this point and there was no sort of hey here's actually how it has to work beyond like our code of conduct um we leaned into external best practices to apply that internally um as well like if you put 50 people on a desert island they're going to form a community they'll probably form subgroups they'll probably become some form of hierarchy and that's basically what's happening um when you're talking about all these different discussion groups so we wanted to create a set of guidelines, a set of clear purposes, a way to everyone understands how to successfully engage in the community and set the right expectations with them so that you know if something did happen and we needed to remove content for being uh misogynistic or perpetuating harmful stereotypes or communicating with care which a company would care a lot about. Um we were able to do that in a way that folks expected um and create a team by which people were like ah they're the ones helping to enforce it. Now, it's not just that, you know, we need to be the enforcers and be the enforcers alone.

You need to inspire responsible stewardship and responsible ownership of spaces as well. Um, so we, you know, said if you owned a community, you needed to be a volunteer moderator. You needed to step in and we would reward you for that. And so, I think like as we're looking towards what you all are building and what you're getting into, don't shy away from creating some standards and some expectations and some hierarchy in your communities to actually keep yourself focused on the purpose. um you know that's responsible stewardship.

That's not censorship. That's like what we need to be able to do. Definitely. And yeah, one topic that comes up a lot in our community is this concept of the tyranny of structurelessness. So the time that's wasted and the tensions that arise and um the lack of clarity that makes um existing in community difficult for people when there is not that level of definition.

So exploring this tension is very um continually present. Um to that end and as it relates to our topic of funding the commons, one uh way that I feel like hierarchy will always inevitably emerge is around who um holds the budget and who has access to giving budget to others and you know what it looks like you know whether that's building out grant programs or funding a government or you know investing in a company. Um yeah I'm curious to understand is that if that is an aspect of what you guys are dealing with internally so you know grant programs or budget for special projects or um what are some learnings that you guys have had from uh scaling community of how to effectively handle these budget hierarchies so that um budget is allocated towards um yeah as many needs as possible in a way that's still directed and efficient. Yeah. So, um, even though Google has as much money as it does, they're quite tight with the, um, pocketbook strings.

Um, and so, you know, one of the real questions we had when our community failed at scale was why not just shut it down like like why not just say all of our discussion inside of the company needs to be workrelated? Um, and what's the cost associated with that? Um, the cost associated with that was very very high. Um, I I leadership and everyone who was in the room was basically saying like no, we want people to continue to have ideas. We want people to continue to express themselves.

We feed you lunch so that you can go interact with folks. Hopefully a lot of people sit alone. Um but uh you know it it's basically saying when we were setting up the team um we were trying to say you have to fund us. We will cost a certain amount of money. We cost a certain number of engineers but in order to do that here's what you get from it.

Um and the purpose of our community is now that we're able to evolve our culture. We're able to put some controls in place. um and we're able to say, "Hey, here's how that budget will be spent." And I think that that's something that is very cognizant for everyone to keep in mind. Um which is what is the actual benefit of the money?

What's the opportunity cost of not getting the money? And then how can you motivate people around that shared purpose? I think that's where you talk about grant funding and you know getting things back into where it needs to be. Community has that shared purpose. Um and to get what you need and to get that grant funding, it needs that sort of like structure around it so that you can actually accomplish what you want to.

What you don't want is a community that isn't accomplishing the purpose that it's set up to do. Um, yeah, it's great to talk. Yeah, it's great to form connections. Yes, that is part of a community, but like what are you actually here to accomplish as part of this thing that you're a part of? Um, sometimes it can be support and to get help.

Other times it can be to get a grant. And so my my advice sort of more generally is to think about how do you very specifically hone people in and then show them the reward of that as well as the opportunity costs of not organizing around the opportunity. Yeah. So that brings up two related questions for me that are yeah continually come up as being very pertinent in the space which is uh the first we'll hear later today from um Devanch and Kevin and some of the other people that have been working on this um effort to uh integrate more AI agents into determining grant funding and um to streamline some of those processes. And I think um yeah, you guys have a unique perspective on how to maybe use AI agents in a way that um goes toward achieving that purpose without necessarily um yeah like uh being too restrictive.

Yeah, I would be curious to know if you're able to share about it what that experience has looked like, you know, toward building some like successful programs like that. Yeah, I mean I I think uh AI agents are awesome. I couldn't imagine if I if I said that they weren't um I probably wouldn't have a job tomorrow. Um but you know when I think about AI I think it should take away the parts that you don't like to do right now. Um you know Agentic is fascinating and deep research is fascinating and Gemini 2.

0 is like blowing my mind every day. Um but it it's not the thing that's going to give you the great idea. Um it's not the thing that's going to tell you all of your life perspective and all of the ways that you have a framing that's in your head today. Um that doesn't exist. It's not data.

It's not in a database. That's like who you are as a person and the connections you formed. And we don't write everything down yet. Um, and so I use these tools and these platforms to take away the burden of the thing that like takes away from your cup, right? Like the the the boring parts, the research for me, the research parts.

Other people may really enjoy comparison and research and gathering things together. I I don't. Um, and it's uniquely gifted at that. Um, in order so you can free yourself up to do the interesting parts that you want to do. Um, for me that's communications and change management.

Um, it's coming up with like, okay, well, how can we actually get this accomplished now that we have like all the facts down? Um, I wouldn't think about it as something that solves everything for us. Um, but it opens up the opportunity for us to do the most impactful work without the burdensome work um that's there today. I I had it um I don't know if anyone here uses like notebook LM. Um, it's like my favorite thing in the world.

Um, I fed it like a couple of vision presentations and like notes that I had and said, "Hey, could you just write the intro to this PRD for this product I'm working on because I hate writing the intro. It's like my leiser." And it did and I was like, "Oh, okay, great. Now I can actually like really focus on this fun part instead of procrastinating because I know that that intro is in the way first." So that that's I I mean that's just the way I think about it.

I think there are a lot of unique and interesting ways to think about it. Um, but have it take away the stuff that takes away from your cup. Yeah. And I think uh one interrelated concept with that is um being able to effectively prompt engineer and figure out whether uh that is actually happening versus making things more Byzantine and complex is to have really um clear metrics for success and for directional growth. So um having had experience of um deeply understanding all of these complex communities, what are some uh advice that you have for effectively um determining metrics in some of these areas that are very difficult to define like you know for example the success of this conference like there you a lot of attempts at trying to have metrics of it how many people came how many people you know followed about it on Twitter like it's hard sometimes to truly capture the essence of what you want to with success by defining metrics especially in um you know more nebulous communities that are about finding joy and connection.

So yeah any thoughts you have on that? Yeah um a lot um I I think a lot of times when companies are starting community management teams community management teams will say things like we're going to form connection or we're going to get people together or we're going to uh you know help them feel good. That's not like an awesome reason to be honest. Um, that's not anything that any marketing team or CEO is going to be like, "Oh my god, thank you so much." Like marketers are talking about impressions and conversions.

And when you're talking about human connection, you need to figure out what the purpose is of the community that you're forming. Sometimes that's successful support queries, right? Um, but it needs to be and other times it could be um helping people create new content for a platform that you have. you need a concrete metric when you have a community that is directly contributing to your company's goals. Um, if you don't have that, you just have like a group of people.

Um, and it will eventually become defunded or dep prioritized as things get tough. Um, so my advice is to really think about what's that like one thing by getting people together that we can drive and move forward. Um, don't be afraid of qualitative metrics either. Those are valuable. Um, so for here, if you wanted to like send out a survey of like did you find a new connection, that can be the metric and that's okay.

Um it doesn't always have to be well through this we've gotten um like four new projects over the the the line and we could have directly contributed 2% to like the global climate budget. Like that's kind of hard to prove as you get more abstract. Um so don't be afraid of very clear very direct metrics and just align um with the people that you're working on and the value of your community um and make sure that it serves that business function. Um at Google that business function is innovation. um being able to say, hey, you know, it's risk mitigation as well.

Um there's a real reputational risk um to going too heavy-handed um or going too lax. And so you want to help the company, you know, strategically respond to like all these external pressures. And that's I think the main takeaway, you know, to how that relates to our community building and our um interconnected ecosystem is to have the same clarity of vision and purpose so that then evaluation goals and metrics can fit smoothly into that as um yeah, as like a company that very specifically is able to define those things. So yeah, that's super useful. Um, I don't know if we have enough time for one more question that I know is um super um Oh, okay.

The the coveted before lunch spot. I can only see him out of the corner of my eye, so I'm like you can't quite see the sheet. Well, in that case, come chat to us at lunch and and I'll be at the happy hour if anyone wants to chat. So, cool. Thank you all.

Thank you. [Applause] [Music]