Ethical Venture Capital: Why Social Media Needs a Conscience (with Brad Burnham & Zoe Weinberg)

I don't think that it would have been possible for a Twitter to create a business model that was based on

being paid by the people who are receiving the value. So, you know, a subscription model of some kind.

Nobody understood what they were going to get from these services early on.

One of the things that's so kind of strange and annoying about the AI era is somebody could now fall

down an extremist conspiracy rabbit hole without even a community of people coaxing them to do it,

but really just alone with a chatbot. Hi, Ravel. So, this week we have two wonderful guests on

the podcast. We have Brad Burnham from USB, one of the original Web2.0 investors. And then we have

Zoe Weinberg, who is a newer VC based out of New York, her firm Exante, focuses on investments

oriented around giving people agency back and digital freedom. Tell us a little bit about your

relationship with Brad first, and then how you kind of got to meet Zoe as well. Yeah, so I

got to know Brad through various friends and through people who were working on companies.

Union Square Ventures was the lead anchor VC that really believed in Twitter in the early

days. They also invested in Tumblr and Etsy and a ton of other companies. They were sort of

the Web2.0, the early social media investor. What really stuck out to me about Union Square Ventures

is that they had a social conscience. They advocated for platform co-ops and for public benefit

corporations and for the idea that we should have B-Corp accountability and that we should have

a system of businesses that gave back to the community they're supporting. And so that work

from Brad and USB is really interesting. And I got to know Zoe much more recently because

she was one of the principles behind the resident computing manifesto, which we've talked about

on this podcast, which is a call for computing technology and for Silicon Valley and for tech

to build things that make people feel better. Build things that make people's lives

genuinely better as opposed to just building businesses that extract and that make us feel worse.

Zoe is a younger, newer venture capitalist, but both of them are people who say they are investors.

They are VC's and I wanted them to talk about the role of venture capital and the way in which

it can build businesses and how it doesn't necessarily have to be the exploitive negative

experience that so many VCs do. Sometimes I as a far leftist anarchist who believes that we

need a radically different way of organizing society. So many people criticize me for starting

companies and for working with VCs. And the thing that I talk about in this and the reason that

I had them on the show is because it is so much more complicated than that. We exist in a society

we have today. We have the incentives and the businesses and the government structures and

we need to work within this system if we're going to change the system, even if we want to radically

change the system. And so there are people within VC who want a better world, who want to use this

leverage to make a more inclusive society and make a democratic society and make a society that

has human agency. And so that's why I wanted Brad and Zoe to come on because their examples

of venture capital that is having a real positive impact on the world.

Yeah. And I mean, I think to your point or two about your position in this ecosystem and how

people view you, you can't change what you don't understand, right? So I mean, I think that's one

of the things that kind of sets you apart is that you do understand how it works. You have worked

within some of these larger companies. You've been there at the beginning of some of the biggest

ones. You've made these amazing contacts and connections that you've defined a very clear and

differentiated point of view that enables you to create change as you go.

The point of revolution social was to make these conversations public, to make it visible so you

can understand the how and why did we get to the world? You know, we have technology industries,

we have a technology world, we have social media that has shaped everything that's going on in

the world. And yet so few people understand the who and the why and the how it coming to be.

And so I'm having these conversations with all of the principles throughout the stories

to make sense of of how did we create the world we have today and to make it accessible.

Absolutely. Okay. Well, let's take a listen.

Say a revolution social, we're talking to a couple venture capitalists and VCs or venture capital

has such this sort of boogie man reputation about the way in which companies get funded in the

role that they're doing. And so I really wanted to have Brad and Zoe on this because

their VCs union square ventures and X anti are examples of investment firms that are looking at

advancing human agency, advancing a humane vision for society and a main vision for the internet

and the way in which we're building news institutions. So I've known Brad for a long time through

here in circles and Zoe, I got to know through her support of the resident computing manifesto,

which I've talked about several times on this podcast already with other guests,

which is that technology should make us happier. It should make us better. And all of these

businesses that we're doing, it's about increasing human happiness and agency. So welcome, Brad and Zoe.

Thanks for having us. Thank you. If Brad, you can talk a little bit about your background. I know

that you are you're someone who is venture capital and career extends back to the the various

sort of earliest ages of the emergence of the web through union square ventures. There's probably

no BC that I look to that has had a more positive impact in funding really interesting companies.

Funding the companies that became social media that became web 2.0 and advocated for a bunch of

good things. So what if you could talk a little bit about your background and how you came to start

USV and the work you've done? Sure. So the background actually goes back a little further than the

early web. I started a career in technology in 1979 working for ATT and actually got to know

open source and unix and mini computers and you had to learn how to sell open source up against

closed source operating system from IBM. So I've been thinking about that and exposed to that

world for quite a while. In the mid 90s, I ended up working for ATT ventures, a captive venture

capital firm. At one point, I was exposed to working with Fred Wilson on a shared deal in New York.

And then after ATT ventures was brought back in house and that's a story in itself,

I helped friends started an internet advertising company and worked on that and brought

Fred in as an angel investor. And then the two of us went off to once that company was fully

fledged. We went off to start union square ventures. And so that was in the 2003 time frame.

That's the background. You asked about our interest in the early web and web 2. I'll say a couple

things and I think we're going to get into a much larger conversation about this. But the couple

things are in the mid 90s when we started investing in web 1, we like probably, I know you

rabble and probably pretty much everybody else involved. Thought we really were democratizing access

to knowledge, wealth and power. We really thought that we were opening up the world. And we fought

for that. We fought for that by investing in companies that were taking advantage of this new

distribution medium to work around the sort of natural monopoly of the existing media industry

that controlled distribution. And it kind of blew things wide open and we had this enormously

optimistic moment where we really thought we were not just going to change the media industry,

but we were going to change society. And we believed in the Arab Spring. We believed that these

technologies would open up, would sort of force a kind of liberalization not just of an economy,

but of also of all of society. I think we'll get into a conversation about what happened and

how we missed that vote and why. And I think that's a really worthwhile conversation.

USV is invested in so many iconic companies, you know, Fred was on the board of Twitter,

but I also remember the work you did with Etsy and I believe Zegato and like delicious was a small

little investment that you guys did. There was this sort of moment where the internet had gone

through boom and bus cycles. And when you founded Union Square Ventures, it was really a down point.

It was a moment in which the internet economy had been sort of laid waste and there's all these

comms that went out of business and a few of them survived like Amazon and Google, but it wasn't

a moment where it felt like there was a wave investment and a wave of money and it made a lot

of sense. You actually set things up during a down period. Yeah, that was actually very fortunate for

us. I mean, it was really hard to raise a fund. Fred and I hadn't worked together before,

most investors, most institutions and diamonds and pension funds and things thought that

you had to have more infrastructure than just two partners. They also had been burnt investing

in first time funds in the late 90s. They kind of got over committed in the late 90s and watched

them all go to zero. And so going out and telling the market that was the two of us that we were

going to invest in this sort of fuzzy applications layer of the web without any hard intellectual

property protection or any patents or anything. And that we were going to be able to business on

that took a long time. But once we were in the market, it was kind of an open field. I mean,

we could Twitter as one example when we first met Evan Jack at obvious corp out in south park

commons. I think it was that office. Yeah. Nobody else. I mean, I think you'd had one other

conversation with benchmark and benchmark said is, you know, if you get paid by that telco is for

your SMS minutes, then come talk to us. But I think one of the things that I am proudest of in

Union Square Ventures is that we were willing to take risks on companies that didn't have an obvious

business model. And certainly at that point, Twitter didn't, we believed that if these networks

would create value for their users, that there would be some way to build a business around that

at some point. And we were perfectly comfortable with that risk. It worked out well. And we could

talk about whether or not the ultimate business model is a good idea or not. Yeah. Well, we'll get

into the problems of the business model and the protocols and not platforms and that whole debate.

I'd love Zoe for you to describe because your path towards being a VC in founding accenti

actually comes from international development work. Yeah. That's right. It's a little bit of an

unusual path to venture. And I should say upfront, you know, in many ways, accenti itself is a

little bit of an unusual venture fund. We're also a very new venture fund. We've just been around

for a couple of years. And our focus is on human agency and digital freedom. So tools that

generally give more control over things like privacy, data, assets, algorithms, infrastructure,

etc. And I should also say, I'm hugely indebted to Brad and the the US, the family for being

incredible mentors and sources of inspiration and wisdom along my whole path there. So I should say

that, you know, starting this fund was also challenging, although maybe in sort of different ways

than what Brad describes, because as far as I know, you know, no other venture funds have really,

you know, sort of explicitly launched with this particular thesis. And I, you know, I think in

many ways, it attracted a lot of like skepticism or like side eye or whatever. But to me, it felt like

the right moment in history and also the right moment in the market to pursue the sort of strategy.

And a lot of that was informed by these earlier chapters of my career that were very far

outside of Silicon Valley and tech. You know, I spent most of my career in investing. I sort of,

you know, at Goldman right out of college, but the period you're talking about rabble was a few

years in which I was working at the the World Bank at the International Finance Corporation,

which is like the private sector investment arm. And I was doing like debt, equity, investing,

and some structured finance. But I became interested in a lot of the domains that I'm focused on now,

because I ended up doing a lot of deals and work in conflict zones. So I lived, I lived in a rock

for a while. I spent time in places like Somalia and South Sudan and DRC. And you know,

I was always there with an investor hat on, but I became very interested in how tech was being

used in a lot of these contexts to do things like surveilling and manipulating and controlling people.

But also there were all these ways in which folks were very resourceful and creative and

were downloading signal to share information or in the case of Iraq, people were using Twitter

in very creative ways to report on what was happening on the front line or, you know, mining Bitcoin

because local currency wasn't reliable. And I just found all of those examples really fascinating.

But it always sort of stayed in the back of my head and I went back to grad school and I sent some

time working in national security and policy. But it definitely planted some seeds that I think

that led me to see that there was this gap in the market a couple years later when I eventually

started the fund. I think that most people don't know how venture capital works. Both of you talked

about raising a fund in VC. You have LPs limited partners. These are the usually larger institutional

investors that are putting together fun. I wonder if you could talk about the fact that your role

as a VC is in some ways both selling to the vision to investors of what you're going to do while

then you evaluate the companies that then pitch you. I mean, that's sort of the one of the fascinating

things. I think people don't realize is that you've got this sort of intermediary where you have

to sell the idea, the vision, who the team is, what the investment philosophy is, how are you going

to get ready and then use that to choose which companies to build to invest in and to build up and

extend so that you can raise subsequent rounds. I know USB is several several rounds. I think,

Zoe, you're still in your first fund. But I think it would be really fascinating if either of you

could talk about what that experience is like because it's not like the money appears magically

out of somewhere, out of nowhere. Well, one thing I'll just say and then Brad has had many more years

of navigating these different lines. But when I was first starting the fund, I was interested in

all these different models of financing. I briefly explored whether there was a possibility of

creating some sort of interesting hybrid fund structure where you could do venture investments and

potentially use other types of financial instruments or tools and also be able to support

open source projects that might not be venture scale, whether there was just like a more creative

way of doing various types of funding and financing. I think there are models out there that do

exist and it might be we're talking about some of them. But to be totally honest, I think the

advice that I got, which was a little bit sobering at the time, but I think was right, was like,

Zoe, you are a first time fund manager of solo GP raising this kind of mission-oriented fund

with a provocative thesis. You got to keep the structure vanilla. You can't manipulate too many

levers here. I think ultimately that was right. I think there is a lot more room to experiment

in the market more generally around different forms of capital. I think it's a shame that entrepreneurs

feel like there aren't that many options when it comes to raising funding. I wish there were more,

but I think there were also practical reasons that it did make sense to raise exanties of

venture fund. I think actually in retrospect, as I've had more time to reflect on it, I think there's

been a lot of utility to having a lot of clarity around exactly what it is that we're doing and what

types of returns we're pursuing. Maybe it's better that we didn't end up in one of those hybrid

structures, but that's not to say that I didn't explore that in the really earliest phases.

To invest in, but they are choosing which funds to invest in. It is always useful, at least,

to remind entrepreneurs that venture capitalists had to raise the money too. They put up with a lot

of crap, frankly. In our first fund, I mentioned it was tough, but we did 80 meetings. We flew

across the country and had people cancel after that. We showed up at their doorstep. We had people

fall asleep on us. We had people walk out on us. It just feels just as bad to be out there raising

the capital. Eventually, if you've had a couple of funds that have done well, it becomes a lot

easier and that certainly has been the case with Union Square. I'm actually, I bristle a little bit

when you described us as intermediaries because I thought we were all in the business of eliminating

those intermediaries, but yes, we are intermediaries. You have this idea that somehow

what we were doing with the internet is disintermediation. This idea that we allow these people,

everybody to connect directly with each other, but what we've seen in all the marketplaces,

you guys, we're really important behind Etsy, that intermediary role, that market maker,

that connecting is both really hard to do and also a really valuable thing.

Early on, the marketplaces that emerge, not just marketplaces for goods, but marketplaces for

information, Twitter, Tumblr, they are really important. It does take something to kick start

those marketplaces. Etsy was kind of interesting because they found a niche in serving crafters

who, the crafters were the buyers as well as the sellers early on. They were buying from each other,

so it started the flywheel moving in that way, but it is hard. I think we now pretty much understand

what it takes to start a marketplace. I think it's actually quite difficult to start a new marketplace

in a place where there are existing marketplaces because moving people to a new venue is difficult,

but I also think that we may overestimate the value, particularly in information spaces.

I think it's interesting to watch Reddit complaining about the degree to which they have been

ingested by large language models when they're not actually ever asking the redditors whether or not

they had an interest being discovered by language models or not. They basically decided

to extract rent in the middle there, and I think what we're going to see is later and later

wait marketplaces over time that extract less from the actual creators of that value.

And is that because with agentech programming, the cost of building the technologies

of equivalent stuff becomes lower, so the cost only becomes how do you convince people?

It's this sort of, I've seen this argument that says, and I'll be curious what the two of you think,

that as we move to a world of more agents, then the value of the door dash versus Uber Eats

goes down because the agent will just pick the lowest cost. Does the fact that technology

creation gets cheaper and cheaper changed the dynamics around these marketplaces?

Which were the modes? The thing that still might be in short supply is trust and reliability,

and I think those sort of things will improve as agents themselves become more functional,

and as we feel more safe and secure in deploying them in our everyday lives. But I think that's

where you'll continue to see whether it's an intermediary or a platform or a marketplace play

a role is where there's some sort of built-in guarantee around trust and reliability that makes it

more potentially paying a premium. I think maybe this is a little bit orthogonal,

but we've been talking to a lot of various AI product builders about what sort of software do

you build in-house versus buy? I think we are going to see this bifurcation where there are lots

of tools that you can kind of spin up easily yourself through vibe coding or whatever, and I'm glad

to see the cost of software coming down so meaningfully for those types of simple applications.

But I do think there's a whole host of categories where it really just doesn't make sense to

vibe code it yourself. Trying to recreate Slack is quite complicated, and there might be lots of

reasons that it actually makes sense to continue to buy certain types of software, and so I think

that's the... I have a friend who has this great analogy around bakeries. There's certain things

like a biscotti that you should just make at home. It's so easy to make, whatever you make at home

is going to be better, and it's like foolproof, but a croissant is really hard to make. It's not

even worth it for you to try to make the croissant at home, even if you follow all the directions,

you should just buy it. I think that's essentially what we're going to see with software, and that's

also where I think a marketplace that can offer something that puts it in the croissant

category are the ones that are likely to persist. I don't know if you would agree with that,

Bren. I definitely agree that that's true, and I really like the frame of, is this...

It's going to become more and more important to have trust in the content that you're consuming

or the relationship you're creating, and so I definitely agree with that. I don't know, though,

that I think we're talking about a much bigger question. I believe that we are beginning...

We're at the cusp of a transition that's as fundamental and as profound as the transition to

the original internet, and I think it's going to change things in ways that are unexpected,

and really important, and it's a moment that we have to pay attention to because

we have an opportunity now to shape something given what we've already learned about how things could

go wrong. It's not going to be here forever. This moment is going to come and go, but I think it's

worth maybe digging into, first of all, what happened with the original promise of the internet,

how we lost it, and then digging into how to make sure that we don't lose the promise of AI.

Yeah, I mean, I think that people forget that before we had the web,

we had something like an internet. It was AOL and Compuserve and Prodigy. We had these online

platforms that you would directly install their specific software, it directly connect over

your modem, you'd be able to do it, and the entire world of what you saw on AOL was controlled by

that company. Now, people could post forums and chats and all sorts of things like that, but

it was an entirely locked down thing, and the excitement of the web, the browser, was that anybody

could put up a web page. It became a permissionless protocol that was really powerful,

looking at what happened and where we're going is so useful, especially if you look at the social

media era, where it felt like from about 2004, 2005 to 2012, there was this sort of moment of

huge amount of Cambrian explosion of experimentation. A lot of your early investments, Brad, were in

those specific companies that had a very open and participatory vision for what the internet

would look like. Yes, and it was very expansive, and people believed in the promise, and people weren't

trying to negotiate for their cut up front. They just thought that they would build something,

it would create value. I probably overused the Cambrian explosion model as a way of thinking

about what happened then. But basically, an asteroid hits the Earth, dinosaurs go away, mammals

emerge, all the niches are up for grabs. Nobody is serving those needs, and so all of a sudden,

there's an opportunity. You're not competing with an incumbent in an existing space. There's a new

space that was created by technology, and you have an opportunity to move into that space,

and you could grow very, very quickly there. That is the phenomenon that I think we're about

to see a repeat of. Those are the moments when I think venture capital can be very productive and

can create a lot of value by supporting, particularly by supporting things that may not be as obvious

as some of the foundational stuff, but this stuff that is going to ultimately create value for

consumers. It's fascinating to me, because when I look at the history, I feel like there's a

period of the Web 2.0 error where you get these companies. And eventually, Google, Apple,

and Facebook consolidate around this in Twitter, and eventually Snapchat, and then eventually,

we get TikTok. You have this moment, which initially it was very open. It was very easy to

build on top of it. You had the FA conference, you had Farmville and Zingha, you had Twitter with

its Open API. Then in finding business models, you had this moment where they go away from being

open, and they discover that they can just make a tremendous amount of money controlling the

experience, controlling the algorithm, controlling the ad delivery. And so that is when Facebook shuts

down the online games, that's when Twitter shuts down its Open API, and you as one of the voices

of the Twitter being open. I wonder if you could talk about what that was experience-wise,

and the dynamics that pushed them to not find a business model that stayed open, but instead

kind of consolidated all of Twitter inside the company. Before Brad answers that, I wanted to

say, I think like this rabble, you're totally hitting on exactly the double-edged sort of venture,

which is like, it allows entrepreneurs to take these huge risks. It's like the ultimate risk

capital for experimental ideas, which is amazing. Also, it is the incentives created by venture

capital that I think also caused a lot of it to then go downhill. You mentioned it was the pressures

of monetization and realizing that actually there were these huge profit engines underneath a lot

of these social media platforms that fueled their kind of downfall or fueled the closing of them.

Where did that pressure come from? Frankly, I think a lot of it came from DC, and I get asked

about this a lot because one of the broader objectives of Xante is to think about how do you

move beyond surveillance capitalism, and the irony is not lost on me that venture itself is part

of what created it. So I think we have to be intellectually honest about that, and also come up

with a compelling reason why this time is going to be different. But Brad, please go back to the

Twitter example because I think it's such a good one. So I don't think that history could have

been different. I don't think that it would have been possible for Twitter to create a business

model that was based on being paid by the people who are receiving the value. So, you know,

subscription model of some kind. The reason for that is nobody understood what they were going to

get from these services early on. And so advertising was hugely important for the early

internet because it allowed people to sort of wander around and explore different things and

not feel like they were, you know, having to decide whether or not to pay $9.99 a month or

whatever before they could have a taste of what it is that they were going to get. So I think

advertising came into focus as the obvious business model. And, you know, once you have

advertising as the obvious business model, all of a sudden you have to control the experience,

the user experience because you want to be able to insert the ads. And so I want to know

that people saw the ads. You want to be able to tell the advertiser that somebody saw their ad,

right? There was this moment. I forget exactly when was it 2013 or something or before that

that Bill Gross came up with Uber Media. And he did something that we now call in the crypto

space of empire attack. He tried to recruit and pay some top level talent to move over to his

and he used the fire hose to allow people to interact with, you know, their existing fanbase

on Twitter. And it was a kind of brilliant vampire attack. It showed the potential weakness of

an open system in terms of its ability to sort of, you know, control. Yeah, Twitter was very

vulnerable to attack. Yeah. I don't know that there could have been a different history there

that that, you know, I don't think we could have launched Twitter with a subscription model.

And I don't think we could have kept Twitter open if there wasn't a way to prevent a vampire

attack. You know, one of the things that's fascinating is very quickly as that moment is happening,

as Twitter and Facebook consolidate around the business model, that's when we start getting

the emergence of crypto. That's when we start getting the emergence of a payment system and a

finance system and people starting to experiment with other ways of monetizing or building businesses

around that. And I know both of you have spent a bunch of time investing in that and looking at

that by definition blockchain project cryptocurrency are permissionless tech. So I wonder if you could

talk about like, you know, there's this thing, which is some people think all crypto is scams or it's

all going to save the world. And it's not the main thing that we're talking about now. But in that

era of 2012. So by the time Bitcoin gets enough established that people actually start knowing about

it through the launch of a theory on through the launch of all sorts of ICOs and everything else,

it becomes this other model for how business could be done on the internet that isn't advertising.

We early on started investing in businesses because we saw the possibility of bootstrapping

a network by creating an economic incentive that was more of a fair distribution of value among

the creators and consumers instead of an intermediary that's extracting rent for creating the platform.

We had the opportunity for, you know, if you were if the transactions within the network were in

the currency and you had to hold the currencies as a user of the network. And as more people came

into the network, the theory was the network would grow in value and the participants,

both creators and consumers on the network would benefit from holding the currency that was

the underpinning of the network. And it felt like a novel approach that was very different than

equity, which is where, you know, the venture capital business has been equity owned by definition

means that there's a third party that is making an investment and has to be repaid as Zoe was

saying earlier. But in the crypto model, it seemed like almost a magic opportunity to bootstrapping

network and allow the participants to participate in that value creation. We got caught up in all

the scams and they overwhelmed all of the sincere players in the market that were looking to,

you know, try a new model. And it was a big distraction from, you know, what the underlying

opportunity really was. I actually still believe that we will see this value and we will see

this model play out, but we still have a ways to go to sort of work through the, yeah, distaste

that everybody has been left with. Just to add on there a little bit, I mean, I, yeah, I do

think that there have been these, you know, notable developments in technology around things like

blockchain or advanced cryptography that actually provide these meaningful alternatives to big tech.

And that's, and big tech slash the kind of platform era. And that is exciting in so many ways.

I think sometimes there's been a fixation on decentralization for the sake of decentralization.

And I think we've learned lots of lessons about the times in which that can add a lot of benefit

and value and also the ways in which decentralization and community governance can be really,

really hard. And there are some real benefits to centralizing certain types of processes or

decision-making or technological development generally. But to me, I think there's like the

technology itself. And then there also, if anything, I would say the intense interest in crypto

and web three in the last handful of years to me indicates something about the cultural moment

we're in and the zeitgeist around wanting to have more ownership over over your data,

over your assets, over the over the networks you participate in. And I think listening to that

signal is worthwhile. And thinking about how some of that signal can be captured in the AI

moment that we're in now, I think is also interesting. You know, I was thinking a little bit earlier

when you, when both of you were talking about the optimism and excitement there was in the kind of

early days of social media, which I remember, of course, as well. And, and I found myself wondering

in that moment, like, was there also a kind of like festering backlash that could have been

perceived in that moment or not? And maybe there wasn't. Maybe it was really just like

pure optimism. I think today we're seeing something very different. I see indications all the time

that the AI backlash is here. I think it is linked to that cultural undercurrent around who owns what

and who has control over what it means to be human today. Yeah, I mean, I remember the trying to

sign people up at the very early days of Twitter. And I went to my friends that I like, here's

this thing. And they're like, why should I get random stupid text messages from my friends about

what they're eating for breakfast or things like that? At least for me in the early days. And,

you know, and then I went and worked with the flicker team and others. There was this attitude that

like this was vapid and superficial and, you know, not real. And then over the web three era was

very utopian in its language and use of it. And very exciting about that kind of thing.

Then it felt like it sort of got turned around and then people are like, oh, it's all a scam.

And so it was like, either all amazing or all a scam. And then social media went from nothing.

Like, this is stupid and superficial to, oh, my God, it's taking down governments.

And there was sort of an anxiety, you know, there was an anxiety about putting your credit card

on the web. There was an anxiety about exploitation of children and things like that. And, you know,

there was a kind of moral panic back then that did accompany the emergence of this and,

and sort of the lack of control. And there was actually a counter revolution, I guess,

that was launched by the media companies basically saying, you know, we can deliver a safe space

for your kids. And you just have to sign up to these, you know, stop online privacy act and the

protect intellectual property act. And so that was a real effort to capitalize on that moral panic

to bring back. The same thing is happening with AI today. I think crypto is a little bit different.

It didn't touch as many people as directly. If you weren't involved, it didn't really affect you

quite as much. You heard about it. You read about it. But with AI now, I think there is a panic

about, you know, the potential loss of jobs, the social dislocation, the degree to which people

who worked hard went to school, got a big degree and all of a sudden feel realized that they're

vulnerable, you know, now feel quite a bit of panic. But it has always been true that technology

is neutral, you know, that it isn't the technology. It's the way the technology is implemented,

and that we have a choice about how we implement that technology. And, you know, that's the thing

that I think we need to be focusing on now. This brings up a fascinating thing like the role

of technology in its neutrality or not, because so you have this quote, which is quite literally

technology is it neutral and you can't divorce it from the intended unintended consequences.

It's like we are building all of these technologies, be it a genetic programming or blockchain

or cryptocurrency or the original existence of the web in these protocols. It is neither neutral

or lacking neutrality. Like it exists within the context. I'm wondering if you could talk about

some of your sort of investment philosophy of Zoe about what Brad said. Yeah, I don't think

technology is neutral. And I felt like there was a period particularly in tech in which it became

the sort of fallback position where engineers would say, oh, you know, I'm just building the tools.

It's, you know, it's how people use it, right? And it's just the AI. You know, it's like technology

for the most part is purpose built. And when something is purpose built to serve some sort of

aim or objective, like that aim and objective is embedded with a whole set of values and potential

outcomes and consequences that have to be accounted for. That's like, you know, basically how I

feel. I think what that means from my perspective, when I'm looking at a company or startup is

we are thinking about first order consequences, but we are always trying to think about the second

and third order consequences. No, that can be very difficult because we don't know exactly how

technology spaces are going to play out. In large part, we actually, you know, we have to have a

lot of trust in teams, right? We're a minority investor on the cap table. In our case, we don't

take board seats. And so there does have to be this real, you know, sort of values alignment with

the people that you're backing that when faced with these kind of novel decisions about who to

sell to or how the technology is used probably in these ways that are totally unforeseen, you know,

that they will make choices that that we could all stand behind. Yeah, it's a tricky thing.

If we love to hear your response, Brad. Well, yeah, I mean, it's interesting to find Zoe and me

on opposite sides of this. I think we probably agree at the end of the day, but I think we're maybe

talking about things slightly differently. So I'll use as an example early on when we started

investing in networks. We had an intuitive understanding that there was some value there,

but we didn't understand it as network effects. The more time we spent with these companies that

we were investors in, the better we understood that the more people in our network, the more valuable

that network. And as we began to understand that more deeply, we realized there was some real

defensibility there. What we didn't understand is that it was going to lead to the consolidation

around Google Apple Facebook Amazon. We didn't understand that it was so powerful that and there

was so much value in the scale of liquidity within a network that there was no way that a startup

could compete with those dominant players. And so we, you know, when I say technology is neutral,

we didn't set out to say, okay, we're going to we're going to support technologies that will

lead to this consolidation. We, I think one of the things that Zoe was saying is that we have a

moral responsibility to try and anticipate what the consequences of our investments are. And I think

we do. But I also, you know, I'm not going to lose a whole lot of sleep that we missed that. I

would lose sleep if we didn't learn from that lesson and approach AI differently.

I mean, to give us some credit because you're not tuning your own horn enough, but like,

you guys did a lot of the groundwork to get the concept of B corpse built into the public

benefit laws. You know, the idea that a venture capital firm can invest in a public benefit

corporation. You know, now we see anthropic as this massive public benefit corporation,

where it's in their bylaws that their impact on the environment and their workers and society

have as much value and need to be considered at the same level of just the stakeholders.

Before USB went and fought for that and lobbied with a larger movement of public benefit corporation

folks, you couldn't do that. You could have to justify these other public goods as a benefit for

shareholder value. So you guys have changed the industry in this way. I mean, I remember the very

first time when I was talking to Chad Dickerson over on Kelly Nellie and McCray were at Etsy and

they were taking it public. The idea that they could take a venture back public company public

that was a public benefit corporation was a radical departure from this other which says you have

to take an account to this stuff. Now, blue sky and anthropic and all these others are public

benefit corporations and it seems normal. But I think that's one of the lessons that we had

from that sort of web 2.0 era is that the underlying rules need to change them.

I think that it was important to establish a framework whereby you could not be sued for

supporting a community or your employees or other stakeholders. That was probably good thing to do.

But I don't know that we should overstate it. I think markets are really powerful.

Markets drive a lot of behavior. I actually think there's kind of a funny anecdote about

I think it was the year we invested in Twitter. We invited Biz Stone to present to our LPs

in our annual meeting and he presented and everybody was very excited about this hot new young

company. Nobody understood it but everybody was sort of feeling good that we had gotten into it.

And then toward the end of the meeting, I was at the lectern delivering some update or something

and Biz Stone raises his hand at the back of the room and he says does USB have a mission?

I think I forget exactly what the language was but it was very definitely focused on the idea

that we should have a purpose and that purpose should be above and beyond equity returns.

And I'm standing there in front of all of the people who invested us.

I love the equity returns.

Trying to think on my feet and I said, you know, no, but we believe that in this era of

open innovation and these large networks that the networks themselves have to attract an audience

based on a set of values. And so we support those companies that commit to this set of values

that attracts that audience because we think that's the only way you can actually deliver a return

to the people in this room. And so I think that's true. And the thing that we're going to need to do

in the AI era is to understand what the market forces are and whether or not we have market

failures, I have actually come to think of a network effect as a market failure.

Kind of by definition, it reduces competition. It prevents new entrants. It is a kind of

natural monopoly. It is a kind of market failure. And so what is the ultimate source of that market

failure? It's a misunderstanding of property rights. You know, it's a misunderstanding of who

controls who owns the data that's created as people interact with these intermediaries.

And I personally have the view that when people talk about ownership of the data that they

create as they interact and the need to ensure that ownership, that's actually not necessary.

What's needed is a equal right and equal interest in the data. It takes my partner Nick Grossman

likes to say, you know, when you have two players in a network, you know, it takes two people to

create a baby. And now we're just basically arguing about custody. Yeah. And it shouldn't be one

person that has that custody. And so if individual users have an equal interest in their data,

if individual users could do whatever the platform does with their data, with their own data,

we would undo a lot of the market power associated with network effects. And how that interacts with

the agentic economy is that that is going to be a very, very important battle that gets fought

over the degree to which I have the ability to access the data I create as I interact with Facebook

or Google or wherever else, or one of the large language bottles. And so if we get that right,

I think we don't need to necessarily legislate or regulate an outcome. We just need to allow that

market to work. Yeah. I mean, in some ways, if we legislate too much right now, we end up

locking in the network effect so that a few incumbent players get regulatory capture.

I'm very curious why so you have to think about this because when I see it, I'm like,

this is not too similar to the cable companies needing to like tear up our streets in order to

bring cable television to people or, you know, radio broadcasters who have to use the public

airways. The only way these large language models can exist is if they have the vast majority

of human creation, like human ideas creation. Some of them are open source, but a lot of them are

banking all their return on the investment on the idea that they will be the winner. They will have

some exclusive access to all this stuff. I wonder if you as an investor in smaller, more nimble

companies that are existing in this AI-agentic era, what do you think about it and how do you

talk about these companies? Well, I want to go back a little bit to this question about the value

of data, these are the users. And I think we're witnessing a big shift where in the earlier era,

data collection was valuable because of advertising. And there are some utility

for users in that. In the sense that maybe you'll get targeted with ads that are a little bit more

germane to your life and might help you find a product you want or whatever. But I would argue

that that utility was somewhat limited. And for the most part, people didn't always love it or

found it creepy or whatever. Today, the value of data collection is really around, you know,

the training of models, right? Like all of the kind of data broker or data collection companies

of the AI era, including like the Mercores of the world that are focused on generating new set

data sets are not selling primarily to advertisers, they're selling to model producers. And that,

I think, is interesting because I do think that the utility for end users, if you have control over

where that data lives, is much, much greater than the utility associated with advertising.

Potentially, it means that you can have AI services that are much more personalized, much more

tailored to you that speak in your voice, that know how you like to consume information. I mean,

all of the things that we can, you know, come up with together. And so then that to me,

the big question I know, Brad, thinks about this a lot too, is like, where does that data live,

and like who has access to it and who gets to control it, right? But I do think the underlying

value of the data itself to the user is totally different today than it was in that earlier era.

Yeah, to be specific, I think it's about, it's not just training the models, right? It's also

how we interact with the models and the degree to which we can personalize as interaction.

So it's about the context window and it's about the memory. And I think we should acknowledge

the reality that we are headed to two possible futures. We are headed to one possible future

where the large language models essentially create something that looks a little bit like an

app store. They get a lot of wrappers that, you know, around those models that create specific

values and specific market segments. But basically, they absorb not just the training data that they

ingest by calling the internet, but the information that every end user in parts has to interact

with them. And they would like to use that in the same way that Apple has kind of used their

app store to sort of identify where value is being created and then suck it into the core, right?

And that's kind of the playbook. And if we end up with a world where there are four or five

large language models that have very intimate relationships with very large numbers of users,

we're going to need to regulate those. It's going to under its own ultimately look a lot like

the telecommunications industry. And we can already see, you know, the degree to which there's

a kind of inherent corruption between the political power and the economic power that sort of feeds

each other. You can see this in in Bezos controlling the well, you know, the Washington post

editorials and things like that. It's already we already watch it. And so how do we prevent that?

We have to understand the nature of this market power comes from the data. If the consumers

have control over a data asset and doesn't have to be exclusive, they have to be able to take

that context memory and move it from model to model. We end up in a market where it's possible that

the large language models become commoditized that open source catches up that we end up with a

larger number of smaller companies competing in an open marketplace, which is a great thing for venture

capital, but it wouldn't create the kind of outcomes that we've begun to think are inevitable

as we look at winner take most, you know, models in the in the web to era.

Yeah, I mean, it's fascinating. We see that moat that they're trying to construct, which is like

your chats with Claude or with, you know, chat GPT or Gemini or Groc, they don't expose the model

that is constructed about you, the history of it. There's there are all these documents probably

marked down under the hood that each of these agents are constructing as they interact with you

or open claw instances and they get more and more information. If you can't export it, then

we're building exactly the same kind of lock in that will exist. And, you know, that's for

individuals, but it's also for companies like if you built your entire company stack around

integration with a bunch of clawed agents and the knowledge of how those agents work exists

on some server that you don't have access to, then you can't switch.

Yeah, and I would argue it goes beyond lock in. I mean, that is a problem, of course, but also just

that the quality and type of data that's collected, I think, allows for a level of psychological

manipulation of individuals that we just have never seen before and we don't know what that

looks like. I mean, I did a lot of work around disinformation a few years ago and obviously that

was all focused on news feeds and on social media and kind of niche conspiratorial communities that

people might end up in. And I think one of the things that's so kind of strange and

annoying about the AI era is, you know, somebody could now fall down an extremist conspiracy

rabbit hole without even a community of people coaxing them to do it, but really just alone with

a chatbot, right? And it becomes this kind of like echo chamber of one, which we've never really

seen before. And I don't think we fully understand the consequences. So it's lock in, but it's also

like, you know, in some ways you're really handing over potentially like the keys to a lot of your

a lot of your psychology to a third party in a way that is, yeah, it's just scary and a little

unprecedented. So the really difficult problem here is to figure out how to approach that

without being patronizing and without trying to construct a reality for people.

How do you give people both individual agency and an opportunity to basically manage their own

choices and not be pulled into this rabbit hole that Zoe's talking about? And I think it comes

down to you have to allow individuals to control context and memory and to be able to interact

with multiple service providers on the back end. And I think if you do that, I'm a believer in

people. You know, I believe people will do not just what they want, but what they want to want.

And I think that that, you know, we will actually shape the market. The dominant economic powers

in this space today, the ones that have raised enormous amounts of venture capital, don't want

that outcome. They want to have everybody within a single environment and a data store that they

control. And one of the mechanisms that I think they're going to, it's going to re-emerge is the

question of what does it mean to be represented by a bot? And can that bot actually have access to

your accounts? And if they try and prevent a bot that you've authorized to have access to your

accounts, you won't be able to use that bot to essentially test a couple of different markets,

bring back a proposal, curate your experience for your benefit. You know, they will be able to

prevent you from accessing those underlying services. I think that's a fight that we need to

see play out pretty soon. It reminds me a lot, you know, I have this this happy built divine,

which is bringing back old vines and as an open way and filtering out AI slop. The metaphor I

use in talking about is it's like podcasting. And early days in podcasting, the podcaster said,

it's my feed. It's my media host. You can't intermediate and insert your own ads or everything

else. Like the downloads have to come back from me. And the podcasters got together and they

organized on a few concrete things about it being always they get to host their own feeds and

they get to stand the standards and they get to host their own content. And that let them build

out ads. And so podcasting defended itself from sort of the centralization that we had now.

And now in the AI era, we have all of the spectrum from people running open claw on their own

Mac minis through everything exists in these context window control systems. I would love

for both of you sort of as a bit of a conclusion, thinking about the AI era, thinking about agents,

thinking about what we're doing going forward. And how do we build business models? How do we build

businesses that are anti-authoritarian that are pro democracy that empower users that build the

world that is inclusive that we want given the complicated systems and power dynamics that are

going on? You know, I will say that I think there's been this kind of like false tradeoff that

people sometimes sort of gesture towards, which is that more openness means less safety when it

comes to AI. And I understand the kinds of generally, right? Like if you have AI that's proliferating

and doesn't have, you know, many guardrails or whatever, like, yes, things can go wrong for sure.

But I actually don't think that that's the right way to think about it because that will almost

always lead you to a conclusion that we should lock down more that we should have, you know,

less open source that fewer people should be in control. It's like a nuclear arms race or something.

This only only a government can control this totally. And that creates its own very meaningful

risks, which are largely anti-democratic ones, right? You know, I've been talking to a few folks

about this concept of like, what is anti-fascist AI, right? And like, how can we how can we be building

around that? I'm definitely keeping an eye generally on the development of open source LMS.

And I'm interested in companies, you know, I don't think it's received as much attention as it should,

but like there's a company in New York called Reflection, which is building like the US deep seek.

And Nvidia, you know, invested $2 billion or something like that for them to do that. And,

you know, I'm really interested to see how that space evolves because I want to find these ways

that you can thread the need all between safety and security and reliability, you know, not,

not necessarily have total chaos, but also do it in a way that does for allow for that free

and open internet and openness that we all want to see embedded into this next era of AI.

Rahul, you mentioned podcasting and you mentioned the way in which that evolved. It evolved

in that way for a couple of reasons. One is that it was relatively inexpensive to launch a podcast.

And the other is that it used existing infrastructure, existing standards like RSS.

And so when we talk about understanding network effects and the way in which that leads to market

power, it's also possible to have network effects around standards. It's just creating liquidity

if everybody's using the same thing, you know, so I think you can achieve some of what so is looking

for a combination of reasonable safety and freedom if you agree on an infrastructure that you're

going to use to share information. And that's essentially what we got by accident with the

original internet. And so I think there is hope. I think it's also possible that the right

outcome might almost be inevitable, you know, as we start to understand more about what it means

to have swarms of agents working on your behalf and how productive that can be, you distance

yourself from the direct interaction with the web. And now you have an agent interacting on

your behalf and curating and summarizing and organizing on your behalf. And one of the things

that I think they are going to do is strip out ads. Unless it's immediately relevant, why would

they show that? And so I'm trying to understand what it would mean to advertise to an agent. Like on

the one hand, you know, there's this whole business emerging that is the equivalent of search

edge optimization that, you know, where people want to make sure that they're included in the

large language model so that their offer is represented when people ask about it. And so

maybe, you know, there's going to be advertising that, you know, advertising to agents, you know,

basically saying, come here, you know, that content marketing for LLMs. Right. Content marketing.

But I don't know why I don't think an agent's going to buy things. You know, I mean, obviously

agents will buy things on behalf of consumers. And maybe you can influence an agent's choice.

But it's not going to be advertising as we know it. So I think if you look at the amount of

money and the degree to which the dominant network, social media networks out there today are

dependent on advertising, I think there's a fundamental change coming. And as it hits, I think we,

again, get back to this Cambrian explosion where I think everything's going to be up for grabs.

And it's going to be a very exciting moment to be investing in small companies. And so I look

forward to that. Yeah. Thank you so much. I mean, I agree that it's in a very exciting moment because

it feels like it's very rarely we at these moments where all of it is possible and the structures of

what the companies are going to be and what the software and the structure of the technology is

going to be. And then it feels like that there was a long time with social media where it didn't

feel like there was any option. I mean, I think I even pitched Nick over at USB back in 2018 or

the idea of building a protocol based, see centralized social media app. And he's like, yeah,

this is a great idea. I love you. I'd happy to keep talking. But there's no business there.

And in 2018, there was not a business. But in the era today with AI and us not having settled down,

there's a real set of options. It's like, is this a liberatory thing? Is this a thing that protects

democracy and human rights and individual agency? Or is it a thing that becomes a mass surveillance

state? And I'm very thankful for both of you for the work you're doing, for the investing,

for thinking about the social implications of the companies that you're helping fund. Also,

thank you so much for coming on this podcast and sharing your vision and history.

Thanks for having me. Thanks again to Zoe Weinberg and Brad Burnham for coming on Revolution

at Social. Our intro and outro music is a remix of the Italian anti-fascist folk song,

Bella Ciao, courtesy of White Records. This episode was edited and produced by Eric Johnson from

LightningPod.fm. Our executive producer is me, Alice Chan, from Flock Marketing.

Don't forget to subscribe to the podcast and share this episode if you loved it.

To sign up for email newsletter, visit revolution.social.

And for more interviews like this one, follow revolution.social on Apple podcasts, Spotify, YouTube,

nostusfoundton.fm, or wherever you get your podcasts.

And a paraphrase, the sabbatistas, some common Dante Marcos, we need a social media revolution

to make revolution possible. We'll be back next week. See you then.