Open is the way

JJ, it’s literally been years since we’ve spoken. I’m so sad by that. I love the conversation we had. “Venture capital meets commercial open source software.” We didn’t even call it COSS then. It wasn’t C-O-S-S, which has become the thing, right?

That’s true.

Back then it was like the beginning, the budding of this idea that you’ve done with this fund. I think it was the earliest days, basically, we recorded; if this is accurate, October 17th, 2018.

2018…

Wow, that was literally a month after the fund was started.

I think you had just announced it. Yeah.

Yeah. I want to give you guys the untold story of those days…

Okay, I like that.

…which I may have told snippets of publicly, but…

Drop it.

Let’s hear it. Let’s get the dish.

I had no money at that point.

[laughs]

Really? What do you mean by no money?

Zero.

Zero dollars?

I was just coming off of a failed second startup, where I basically ran out of money and the company died… And sort of leading up to October 2018 when we did that recording, I was doing blogging. I was blogging quite a bit about commercial open source. Open source companies, why are they interesting, they’re sort of different, they’re kind of misunderstood… Lots of nuances about open source itself, but also kind of like the intersection of open source and startups is kind of strange… So the blogging encouraged me to learn more about the market and companies that were raising capital. And if you remember, 2018 - we did that podcast at the end of 2018, but 2018 was a super-eventful year for open source company outcomes. That was the same year Red Hat made the announcement – the transaction didn’t close that year, but they announced that they were getting acquired by IBM. If you remember, GitHub was acquired by Microsoft that year as well. Cloudera and Hortonworks did a big merger, the two big Hadoop companies…

I think the year started off with a company that I was sort of competing with a little bit in my first startup called CoreOS, which was like a minimal operating system; they were acquired by Red Hat. I believe MuleSoft was acquired by Salesforce, which was another pretty big acquisition… So an open source enterprise service bus, middleware technology.

I’m friends with the founders of all these companies now, years later, just kind of building the fund and kind of reaching out to these people and sort of telling them what I’m doing… So like crazy lucky to have made a lot of progress. But yeah, I’ve been lucky enough to get to know Bob Young, the founder of Red Hat, Tom Preston-Werner, the founder of GitHub, Ross Mason, the founder of MuleSoft… I knew Alex Pulvey from CoreOS previously… But a lot of those relationships happened as a result of starting OSS Capital, and kind of evangelizing our points of view to the world, and sort of explaining why we think open source companies are interesting. This is a really profound way of building a business and building a company. And yeah, I think as you were saying earlier, Adam, this COSS acronym, Commercial Open Source, had not yet been kind of clarified in my mind. It wasn’t clear enough to sort of like give it an acronym. And in fact, the name of the fund is kind of like weird, and sort of messy, because it should be like COSS Capital, not OSS Capital. A lot of people are like “Oh, you’re funding open source projects. That sounds so awesome.” I’m like “No, we’re not funding open source projects. We’re funding companies that are built around open source projects.” So much nuance to this, but… Yeah, it’s pretty crazy. It’s been going on five years of having the last conversation, and as I was saying, I did not have any money then… So my previous startup ran out of money. I had little savings, but not enough to make a big splash, or anything.

[08:34] So I tweeted the fund, and I announced the fund when we had, I think, no money under management… And then the following 30 days, I think I was able to raise like 500k or something, from – because the tweet… I actually have the launch tweet framed in my office.

Is that right?

It was like a pretty momentous – the tweet went crazy. I didn’t really have any Twitter followers at the time, but it was like hundreds of retweets, and almost a thousand likes on the tweet, or whatever… And this was like the first time anything like that had happened to me, and so I was like “Wow, this is–” I kind of had an intuition that it was like a good idea to do OSS Capital… But the validation from the tweet was like “Oh shit, I have to actually do this fund now.”

Yeah. Put it out there. Gotta do it.

Yeah. I mean, it was very helpful just to kind of get feedback, and stuff. But yeah, so - our biggest investor in the fund I’ll give a shout-out to, for actually the better part of the first two years of the fund… He’s a friend of mine who was an investor in my previous startup, by the name of Clement Peng. And Clement has been an amazing supporter of the fund, and he was the founder of a Sequoia-backed company called Wavefront, that was acquired by VMware for a few hundred million dollars, and so he had some liquidity from that…

And so Clement - I don’t know if I have his permission to share the specifics, but he was, I’ll say, our largest LP in our first fund for the first more than two years of the fund, when we only had like a few million dollars raised. And it actually took – I mean, I’m skipping a lot of history here, but it took close to three years to raise our first proper fund… Which was our first fund, and a $52 million fund. The reason I think it took three years was – 2018 kind of ended, and then 2019, 2020, which was COVID, were really hard. And then 2021 is when we actually raised like most of the money in that first fund. So yeah, almost $50 million of the $52 million fund was like raised in 2021. Unsurprisingly, 2021 was when money was falling out of trees. [laughter]

And you were catching it. Yeah.

I was extremely lucky with timing. And the other thing I think very fortunate and lucky with timing too was if you zoom out in the last five years, which really seems like a decade or more - so much stuff has happened in the last five years - our thesis really ended up becoming proven out and like really correct. Like, there’s a lot of data points… And I actually – it’s kind of funny, I talked to people, we’re like “Yeah, I started this fund five years ago.” It was a really weird, kind of like super-contrarian thing. People in the tech industry found it to be super-interesting, especially people that were sort of coming from an open source kind of disposition, or developers, or building infrastructure stuff… They really got it. But to say that open source companies would be a thing, where there’s going to be like tens of billions of dollars getting invested in that category sounded insane five years ago. It sounded like totally crazy. But for me, I was just very fortunate and lucky to have the experience I had with Kubernetes in particular, going back a decade now, 2014-2015, learning about that community and getting to know how great, transformative open source projects are really built, and just associating with the people that were in that ecosystem very early on, and sort of seeing how the project grew and developed…

[12:16] And kind of like also putting together different ideas around how you could create a company, and how business models work, and how business models are a little different, and all that… I was very, very fortunate.

So now we’re on our third fund… Our funds are all the same size. OSS Capital has remained a very small team. It’s a solo GP fund, so I’m the solo GP. I kind of raise all the money and invest all the money. But I have a CFO now, so that’s a big upgrade…

Yeah.

I hired a CFO about a year ago, and… Yeah, we manage – including a couple SPVs, we manage a couple hundred million dollars now, and we have like 70 or so investments… We’ve led almost 40 rounds of funding at incorporation, meaning we’re like the first investor, first check, and priced the company. First board member, I think more than 15 instances of that… We’ve had one exit, which is exciting…

Congrats.

We sold a company called Remix to Shopify, to Tobias Lütke. And Shopify has completely re-architected their entire web framework around Remix, which - it’s not really correct to call it a sort of Next.js competitor alternative… It’s a React-ish sort of web development framework, and it uses a lot of web fundamentals, and has some really beautiful abstractions… And so that was a company that we invested in and we sold it to Shopify. But we’ve been very lucky in a handful of cases to invest at inception, and help companies build businesses that are producing between 5 and 50 million in revenue… And a huge privilege to kind of take the fund from an idea and a thesis, to attracting founders that honestly have had so many options and ability to take money from any fund out there, that can see how special they are, and take a bet on me and our fund, and OSS Capital’s thesis… And so extremely lucky to have just the privilege of working with – Cal.com is doing incredibly well. We were their first investor. Hoppscotch, which is like an API development platform, millions of users, they’re doing phenomenal… Plane, which is like a monday.com, or kind of ClickUp, also going after kind of JIRA application company… We’re very lucky to be the first investors in Parallel, which is actually a game studio building game experiences, and they’re worth a few billion dollars now… We were the first investors in a company called Liquid AI, which is building an AI model architecture alternative to the transformer, using some scientific kind of breakthrough work that came out of MIT, and Austria, Vienna.

And yeah, I mean, we’ve invested in a lot of different types of things, mostly B2B-oriented kind of software company, technology companies… But we’ve stayed true to our open source roots and core to everything we’re doing… And yeah, very optimistic about the future.

I’m curious about your thesis… I want to know the data points and the thesis. I don’t think I know it for sure, but I want to tangent back one step, because I recall when we were working closely with the GopherCon folks that around this time - were you the person who created KubeCon? I believe you did, right?

Yeah, that’s right. Yeah, I did the logo, the name, I organized the first two events…

Yeah.

At GopherCon – Brian and Eric were too busy, but they were thinking of doing it. And I knew them, and still do… I haven’t caught up with them in a while. I should fix that. But it was kind of Kelsey and I and a couple of other people, we were sort of like “This is an obvious idea. Someone should just do this.” And then I think I just like woke up and I was like “Look, here’s the name, this is the thing, whatever.” And then I just kind of got involved. But it was a very obvious idea. I was just also fortunate and lucky with timing around that. But yeah.

[16:22] Settle the bet then, because I’ve been calling it KubeCon… Some people call it KubeCon. Obviously, it has a longer version now, which is KubeCon Cloud-Native Con. That’s a mouthful. But is it KubeCon, or is it KoobCon?

I like the diversity of pronunciation for different things…

Oh…!

And I’m not going to pick one.

[laughs] He refuses.

For my personal pronunciation, I’m trying to remember what I would say… I think I would say KoobCon. I don’t know if I would say KubeCon, like pronouncing the cube… I’d say KoobCon, K-O-O-B. That’s probably how I would say it. That’s probably how I still say it, but… I mean, some people go KoobyCon…

Is that right? I’ve never heard that one. [laughter] I’ve never heard that one.

To each their own, I guess. I know I’ve argued both sides. Depending on whatever side Adam’s on, I just pick the opposite and I tell him “You’re saying it wrong, dude.”

Yeah, because he has to.

I have to.

I’m all for diversity. Pronounce it how you like.

There you go.

Okay. Well, let’s go back then to the real meat here, which I did want to talk about, because I was like “Gosh.” You mentioned Kubernetes… I recall KubeCon, I recall you being the person who created it… And now it’s just massive. It’s a massive conference. I mean, you were very, very early in this idea, obviously… And I know you didn’t found the CNCF, but KubeCon became their conference. It’s the epicenter of the CNCF and all they’re doing… So every year there’s – that’s what’s happening. That’s the place to be. And we were there recently, this last year – wasn’t it, Jerod? This last year? Not this calendar year, but –

Yeah. KubeCon North America.

Right. In Chicago. And the expo hall floor was like a circus.

Massive.

It was so wild. But anyways, let’s go into your thesis. What is the thesis that you drive upon for this fund, this company?

So it actually took me probably – so the exoskeleton of the thesis has never changed, and it’s kind of been very consistent. I’ve been describing it in slightly different ways. I would say about a year and a half in, maybe even a year in - I’ve got to look at my tweets… Or zeets. How do we pronounce that?

Posts…

[laughs] I think zeets just sounds kind of like perverse, or like an STD or something. It doesn’t sound like –

Yeah, it sounds like some sort of a skin abnormality, or something.

Yeah. Some disease. I’m just going to say tweet. I came up with one phrase – because people ask me this question all the time, “What is the thesis?” It’s so strange. And the phrase - I want to say probably late 2019, maybe even early 2020 - is “Open source is eating software faster than software is eating the world.” So that’s been the phrase. And that’s kind of been the slogan, and the whole mantra or whatever. And I think it does really well tightly kind of articulate what the fund is about, which is like, it’s easy to explain things to people when you can kind of build on an existing, well-understood dynamic or concept or whatever… It’s sort of like reasoning from analogy, which I think often is not that helpful, because you’re not really deeply understanding something if you reason by analogy. But Mark Andreessen is very famous for saying software is eating the world, and this is kind of the Andreessen Horowitz thesis, right? And they’re very known for that, so people have a strong understanding… So I was just kind of like trying to build on that to say “Well, actually, we think instead of software eating the world, which is kind of true, we think something more important and more kind of disruptive is actually happening, which is that open source”, which really means something very specific… It’s like a really like kind of specific phenomenon, and important… “Open source is eating software itself at a rate that is faster than software is eating the world.”

[20:23] So I’m beating the drum on this for - probably another year or two years go by… And then as I was raising our second fund, which was sort of like 2022, 2023, I started to visualize this. I started to kind of like write down… I think I was like having breakfast one day or something, and I was just trying to write down two waves. There’s a really big wave, which is like the software eating the world wave, and then there’s a smaller wave, which is the open source eating software wave. They overlap. They’re sort of super-imposed on each other.

So the big wave is like this huge wave. And then if you look at the horizontal axis, it’s time. So a number of years, a number of decades, or whatever… And then the vertical axis, kind of up and down is the size of the market, and how much money, how much industry size kind of has been to sort of like eaten; how much of software has eaten the world.

So it’s like tens of trillions of dollars or whatever on the vertical axis. And then on the horizontal axis, it starts around 1970, maybe 1960 or something… And then sort of the wave kind of goes – Oracle, Microsoft were founded, and then what other software companies and software industry were sort of created… So it’s like this line that develops, and you sort of put logos on it… And then the bottom wave is open source eating software. So how do you visualize that? But again, it’s in the context of companies.

And so obviously, I put Red Hat in there, and I put JBoss, and whatever. A bunch of other companies. And then as time goes on, you’ve got like MongoDB, and Talend, and MuleSoft, and GitLab, and HashiCorp… All these other companies. Databricks… And there’s more and more. Odoo, Grafana… There’s a lot of these companies now.

And my thesis in visualizing this extremely oversimplified image of the slogan, the visualization of this slogan, was that the rate of growth of the small wave, that small curve - the rate of growth is faster and expanding at a faster rate than the rate of growth of the large wave. Because the rate of growth of the large wave has sort of hit this kind of like asymptote or sort of like scale in which you can continue to digitize more industries, and like software is eating more and more things… But it’s almost pervasive at this point. Yes, there’s software for so many parts of the world… And now if you really look at the fundamentals of software and sort of like what it really means, there’s two ways of building and sort of releasing software to the world. You can build something that’s proprietary, where the person who created the technology or the software is solely in control of the ability to make the software better, and improve it, and dictate can you trust it or not… The other way is open source.

And if you look at open source in the context of Linux, there’s so many reasons why Linux succeeded in basically becoming the dominant server-side operating system. By the same token, there’s so many reasons the Android smartphone platform succeeded on smartphones just in terms of the number of smartphones. There’s many reasons why languages and databases are increasingly becoming dominant as open source open platforms.

Now we see things like LLaMA, which is actually not technically open source, but I think there’s some forgiveness that we have to kind of offer there, because neural networks are not source code. They’re a different artifact. And we’ve blogged about this and stuff last year, actually.

[24:04] So the idea, the principles of open source have remained unchanged. There’s a lot of resistance to that in the world of capitalism, and the world of business, particularly by financier types and VCs. I don’t consider myself a VC. I consider myself something a little different, but technically my job is as a VC. But we have a philosophy, we have a view that the world across all technology categories, starting with software - but it’ll eventually sort of like permeate across other things, too - is moving more towards transparency and systems that you can actually trust and verify that you can trust them, and systems that are user-empowered, user-owned, user-controlled, and permissionless. And those things are totally orthogonal to any particular type of technology. You don’t need to be a blockchain, you don’t need to be crypto to have those properties. You do need to adhere to a certain set of principles though.

And so that’s kind of structurally what the fund is about. How we’ve been sort of prosecuting these views has kind of been relegated to software company startups, and open source companies, largely, with a handful of exceptions… But I plan on doing this for the rest of my life, and it’s extremely fulfilling and comforting to see more and more evidence of the thesis actually playing out to be more and more correct. I mean, I can certainly imagine a world where I wake up every day for years on end, and the world is just more and more proprietary, and more locked down, and more centralized, and there’s less hope of any of these ideas having any long-lasting potential. That would be really depressing. [laughter] But so far, I see a lot of evidence on the other side, so it’s kind of cool.

Break: [25:55]

Are you a Silicon Valley the TV show fan, by any chance?

I’ve watched a few of the episodes. I’m not like deeply immersed in it, but I know a lot of the references and stuff, yeah.

I’m going to give you a reference that’s not going to be a plot issue for you if you decide to watch it… And don’t be offended by this either.

Okay.

This is a crucial point, I think, that we have to ask ourselves if we agree with that thesis. There’s a moment in the show when – there’s a character named Jin Yang. He is part of the hacker hostel that Erl Bachman created in episode one, season one, that all the hackers that eventually became Pied Piper created. They were part of this place.

And Jin Yang wanted these guys to be his friends; he wanted to be in it so badly. And he tried to create his own startup a couple of times… He was never really about being the software creator and a startup maker, but he wanted to be the friend. He wanted to be in the crew. He wanted to be in the clique. So that was his main thing. And later in the show he’s standing in front of this whiteboard, and what he had begun to do - I won’t ruin any plot for you. There’s more in there, and I’m going to paraphrase it… He was sitting in front of this board that had a bunch of Chinese writing on it. And I don’t read Chinese, so I’m sorry, I couldn’t read it all. It’s probably Mandarin, as a matter of fact.

Okay.

And what it was was clones of US companies that he would take and create into China.

[31:12]

“Jin Yang, are you copying all those companies for the Chinese markets?”

“Oh no.”

And when I say this, I don’t mean this offensive, so please don’t be offensive, but… You know, I’m – full disclosure, Jerod and I are both small angel investors in cal.com. We use cal.com. We like cal.com.

Amazing.

Yup.

Prior to using cal.com - I’m going to say it five more times - is we were users of, not investors of, of Calendly, the other most prominent.

Yes.

And so the reason why I bring this analogy up is Jin Yang was famously standing in front of this whiteboard and his intention was to simply clone well-known companies, and just take them to China, right?

Yes.

So is part of this, not so much thesis, but part of this, are you a proponent of – let’s take an idea of a Monday.com, for example, or a Calendly.com. And say, “Okay, let’s find a way to create this as an open source way.”

100% unabashedly yes. And it’s a brilliant strategy. By the way, there’s a gentleman by the name of Neil Shen who recently broke off from Sequoia, who literally did that and made I think $50 billion bringing Western SaaS consumer internet companies to China, and creating Sequoia China, which recently had to like split off from the mainland company for geopolitical reasons.

Yes, but that’s not the whole story. So I think that there is actually a lot of truth in that. That’s a funny example of the Silicon Valley show. I should watch that episode. It sounds entertaining.

Oh, yeah. You should watch the whole entire thing just so you have the context of that one episode.

Just so when you talk to Adam, he can tell you things with less words, and you can say “Yeah, I know exactly what you’re talking about.”

We can catch up and have inner dialogue about it. I think that there’s a ton of validity to that example. And I think that there’s no shame in that at all. By the way, I want to kind of go back a little bit further in history and sort of say… SaaS – so one other way I describe our thesis… And this has also been super-helpful for me. There’s been an evolution of the distribution models, let’s say, or the innovation models in which you have ideas that manifest into technology, in the software industry or the technology industry.

In the beginning, you literally had to physically sell – I’m carrying a pen here, but maybe my unplugged keyboard is a better example. You had to like physically burn – I don’t even think CDs existed then. They didn’t. They were sold on floppies. Oracle’s first software was literally shipped in boxes, with like very large floppy disk series…

The 5-inch ones.

…and in order to upgrade the software, you would have to order, and they would have to ship you a new series of floppies, from which you would need to kind of like physically install, and you would own those physical like pieces of technology that would give you an upgrade. And then the pricing model was all built around that, too. It was perpetual licensing, where you would like basically buy that version, and then you would have maintenance that you’d pay for the version, and in order to upgrade a new version of the software you’d have to physically spend a new upfront amount for the new version, and then do that over and over again.

[34:18] Most of the – I would say probably 99% with some interpolation synthesis, mixing and matching of ideas… Most of the categories in software, including SaaS, which have become a little bit more like fractal domain-specific… Cal.com has almost like one feature, and wow, it’s a billion-dollar company. That just happens because of economies of scale, and humans using more and more technology. But most of these ideas were implemented in the seventies, eighties and nineties, even in the mainframe days. Scheduling, booking things… I mean, this is like calendar functionality.

I was recently watching the oral history of Mike Markkula, who was the first CEO of Apple. He was a guy that Sequoia founder, Don Valentine, introduced to Steve Wozniak and Steve Jobs to basically help them with the business plan, and like run the company. And Mike was also a really brilliant programmer, and he wrote Fortran and basic programs that would actually run on the early Apple computers. And the collection of those programs came with a credit, sort of a copyright credit at the bottom, which was Johnny Appleseed. You might remember this from the early Apple computers, this Johnny Appleseed name. So Johnny Appleseed was Mike Markkula. And Mike in the computer history museum describes this – it’s on YouTube, you can watch it. It’s a really great, multi-hour oral history. Mike actually personally wrote the programs for the applications that would ship in the early Apple computers. So including the calculator, calendar app, bookkeeping app… A lot of these basic – a very basic, primitive word processor…

And so I just think back to like the most fundamental, basic things that humans do as part of their work or their lives have been re-implemented and re-implemented so many times in the history of technology, going back to literally the fifties and sixties, that when we have new ways of distributing and innovating in the core technology, what happens is you just reinvent and you re-implement what happened, what came before, with shinier and newer, sexier, more sort of societally-relevant and interesting abstractions. And then those things end up taking off. And oftentimes people are given a choice. There’s sort of like the free version of the thing, maybe it’s open source and it’s like less sophisticated, but you can totally control it… Then there’s the proprietary version of the thing, which is much more polished and sexy and packaged. This is so fascinatingly – for the first time, and sort of skipping to the AI stuff, we’ve seen this incredibly shockingly surprising almost acceleration of equivalency. This actually happened this week - I mean, the latest LLaMA models are state-of-the-art. And by all accounts, the benchmarks are pretty much indistinguishable, and the capital required to produce that is similar. Mark Zuckerberg literally did an interview a couple of days ago saying they’re spending tens of billions of dollars building LLaMA, and they’re giving it away completely. And in fact, they’re making – Yann LeCun was just tweeting about this the other day… They’re making their licensing even more permissive. You can do distillation, quantization… You can use the outputs from LLaMA models to train other models… It’s almost like an Apache or MIT-level permissiveness. The artifact is very different for a neural network that has massive amount of information compression. Not in the Pied Piper sense, by the way, but in the neural network, AI, tokenization of –

There you go. At least you got some of it.

…massive amounts of data sense.

For sure. And they’re literally being super-permissive, because Meta doesn’t actually have a need to monetize these AI models, unlike Anthropic and OpenAI. They can just give it away. But there’s actually a choice now in the industry where you can say “I’m going to use a proprietary AI model”, and all these trade-offs. Or I can just take LLaMA, and it’s a few gigabytes, and I can run it on some servers and get the same capability, and it’s 50% to 100% less expensive.

[38:10] I can’t remember – and I obsessed and spent all of my time on this for like five plus years now, a lot more really, thinking about it… I can’t remember a time where there was this equivalency that happened that fast. Cal.com is actually one example. I think Cal.com reached and exceeded the Calendly parity feature set maybe a year ago. And they’ve actually out-innovated and added a lot more things that Calendly hasn’t been able to ship, because they’re sort of like a consumer-oriented, consumer-focused SaaS company primarily. And Cal.com is more like GitLab, where they’re selling on-premise subscriptions, and it’s much more customizable and extensible and so on. They’re almost kind of two different businesses. To your point, the sort of copycat dynamic is pretty clear.

So I don’t actually push back on that at all. I think there’s a lot of validity to the responsiveness to an existing proprietary incumbent - maybe not necessarily inventing the category, but being first to market with that particular technology. And then over a pretty short window of time you see a massive upswell of developers and interest in “Look, we want to have something that we can control, that we know we can trust, and that we can extend, and that we can use.” And you have people just building it.

I can’t tell you how many times I see – there’s some new, shiny SaaS thing that’s proprietary, and then you look at Hacker News… Because I look at Hacker News every day. I find so much stuff on Hacker News. My life is basically Hacker News and Twitter. But you see something on Hacker News, the open source alternative to this thing… And some of these projects are really successful. They’re taking off massively, they have huge amounts of users. And I think this whole wave of AI we’re going through, and LLMs getting better with code generation and stuff - we’re probably only going to see an acceleration in open source alternatives.

Just imagine, even Riverside… Riverside’s a good example. I don’t want to pick on Riverside. I think it’s a good company, and the founders are talented, and stuff… But this is a piece of software. If you wanted feature equivalency to this and you use the existing AI models that generate code, and are getting better every day, you can get to a state where you release an open alternative probably 10 or 20 times faster than you would have been able to do even a year ago or two years ago.

So my thesis - I could be totally wrong about this; only time will tell. My view is that the innate desire that people have to want freedom, to want control themselves, and to know that they can actually trust the thing that they’re using - that deep, almost primal human desire to not be in a cage and trapped by something… As long as the technologies that exist make it easier to create those outcomes, we’re going to see more of open alternatives and open platforms. That’s just my view.

Yeah, I think so too. I think that the hacker spirit is very much alive inside of the thought of “I like proprietary thing X. I would like it more if it was open source, therefore I’m going to go create an open source version of X.” Whether you call that a clone or a copycat, those can be kind of diminutives. But a lot of work goes into that, a lot of talent goes into that, a lot of labor. And when you set it free, and it’s probably not quite as good, but maybe eventually it gets there, maybe it doesn’t, but it provides a different core set of values… I think there’s no shame in that game whatsoever. Whether or not you’re 10 to 20 times faster or better with code gen, I don’t know. I think the verdict is still out on that one. But there’s certainly a lot of money being poured into making that a reality, and I certainly would appreciate a world in which I could build things 10 to 20 times faster with code gen.

[42:09] I’m just surprised that you can invest money in all of this and make money at the end of the day. That’s just kind of where I go back to. I’m with you on a lot of the stuff… And this goes all the way back - I looked it up - 2016, Nadia Ekbal first came on our show, talked about funding in open source. This was after she had written that post on Medium called “How I Stumbled Upon the Internet’s Biggest Blind Spot.” This is before her Roads and Bridges paper, and all this stuff. It’s at the very beginning of her short career in open source. She’s moved on to other areas of interest. But I remember her saying, “Wouldn’t it be cool if VCs could just invest into this stuff?” And I remember telling her, “That’s nuts. They want returns over everything. That’s the VC way.” And I was interested when you said you consider yourself slightly different than a VC. I think you probably are. I’d love to hear how you consider yourself. But has it been the hardest part of this, convincing people there are sizable returns on the other side of these open source companies?

I’ve learned a lot in the last five years of doing this full-time, and I would say having enough capital to really properly test things - three and a half years - you can absolutely make huge returns investing systematically in open source companies. You will not make any returns investing in open source, but you have to invest in companies. And I think this is an extremely important thing to clarify. Capitalism is fundamental to both a free market economy, a free and open and free-loving society, and a productive global world market economy, and all of that.

I think that there’s a lot of arguments and debates, “What kind of capitalism? What type of capitalism?” “This and that type of capitalism…” I got caught up on that for a little bit for a while. There’s like stakeholder capitalism, and shareholder capitalism, and all this kind of stuff. Look, it’s just capitalism. Build something great, charge money for it, make more than you spend. Keep doing that consistently. It’s not complicated.

My learning in making a lot of investments over the last few years has been truly great founders who are skilled in a multifaceted way at selling, building, recruiting, hiring, setting priorities, leading and driving teams relentlessly, are extremely rare people. Peer of Cal.com, by the way, is one of those people. We were very fortunate, because that was the largest seed check that we wrote out of our first fund, a $6 million check.

By the way, at the time - I did mention this publicly the other day, but I’ll mention it again, because I have like fragments of comments on the internet… At the time we invested in Cal.com, it was a little bit less than half of the fund. Because I was still raising the fund.

That’s a big bet.

Which by most accounts is bats**t crazy… But today has turned out to be not a bad investment. It’s still very early. The number of open source founders I’ve invested in – so the thing I was kind of trying to get to there… The number of founders I’ve invested in where we sort of like really heavily encouraged them to start a company, and/or they weren’t really on the founder path themselves, and didn’t really have that kind of like innate – like, the way I express my mission in my life is to be a capitalist and start a company, and I just so happened to be a really great open source developer as well. Times in which we’ve kind of like over-rotated on our thesis, and been a little bit too intellectually obsessed with “We want to fund so many open source companies”, we might not do as well. We might actually lose our money. And I think the reason behind that is really capable, great founders are extremely rare people. And it is extremely hard to build a company that figures out a business model and ends up sustainably charging money for something that solves a customer problem. And you have to keep continuing to do that.

[46:10] So I think part of our journey is really just going to be having better judgment and discernment over which founders are worthy of this venture capital rocket fuel thing, which I find really powerful and awesome in extremely small doses. When you do larger doses, it is catastrophically harmful, toxic, and net bad for the world, with extremely few exceptions. Literally, unless you’re building some very complex material science –

Yeah, deep tech kind of thing.

…physics-intense type of company. When you’re talking about the world of bits, raising tons of venture capital, in my opinion, is a scam. I’ve actually said that publicly. And I think that if you were to just siphon the size of venture capital and focus it on the bits companies - not the ones that are doing hard atoms type stuff, but just bits - it should be 95 percent smaller of an asset class, a cottage industry, a service industry, very much focused on the absolute best possible founders. And you do not need a lot of capital to make a huge, huge, huge, huge impact. That’s my learning. Can you make money with open source companies? Absolutely. Will there be more of them in all these different categories? Yes. Will there be 10 or 20 false positives in each possible category? Probably. Perhaps because maybe we overhyped the category. Because I’ve been doing a conference, and blogging, and tweeting about it, and there’s more funding going into these companies now, and all the VCs love it now too, and all this. That’s overfunding. That’s inefficient. That’s bad. It’s actually bad. In fact, open source companies, if they’re, again, focused on a pure bits business and not shipping a hardware product or something, they can be the most capital-efficient companies you can imagine.

Because if you have a huge open source community that is like – this is kind of a bad way of saying it, it’s kind of distasteful or whatever, but it’s true… Like, effectively subsidizing massive engineering costs that don’t have to show up on the balance sheet of the company, and also subsidizing go to market, marketing, lead gen, customer acquisition… Because the moment you flip the switch on a product that solves a customer problem, you have a huge amount of that user base that experiences the pain of that problem saying “Look, sign me up. I love your product, and I’ve been using your open source thing for the last five or six years. It solves a fundamental problem in our business, and we’ve been looking for the thing that you are selling for the last five years. We’ve just discovered it. Please sell it to us. Take our money.” That is an extremely efficient sales process, that rivals any other type of sales process you see with proprietary technology companies, that have to spend a ton of money on advertising, on sales, go to market, and all this kind of stuff.

And there’s a lot of other aspects to it too, that I didn’t even talk about. But it is very hard, and it’s very nuanced as well. One thing I think that is still very unsolved, frustratingly, and painful, is that I continue to see people who are not philosophically aligned, or remotely as interested in sort of the essence of this stuff as I am, who are very successful VCs, take the exact same playbook that we have, and then when the companies get to a point of 50, 100 million in revenue, a few hundred million in revenue, they really start to heavily encourage the founders to change the license of the core technology. I find that to be unnecessary, duplicitous, harmful, two-faced, and despicable, to say the least. I could probably talk a lot longer.

We have a phrase, and you can take this if you like. You may have heard it. “Rug pull, not cool.”

That’s right.

Yes. It’s also a good way of saying it. That’s effectively what OpenAI and Sam Altman did, by the way. A.k.a. Scam Altcoin is my handle penname for Sam Altman.

[50:09] So I was going to ask you your opinion on the license changers of the world, but it sounds like you just laid that one out there.

Yeah. I think it’s just a bad thing to do, because - look, here’s a couple of data points; just a couple of data points. When MongoDB went public, a couple of years later they changed their license. When Elastic went public, a couple of years later they changed their license. When HashiCorp went public, a couple of years later they changed it. Okay, there’s a pattern there. These companies aren’t doing it when they’re at the most juicy, sensitive parts of the growth curve. They’re doing it when they’re in the saturated part of the growth curve. Because the board, principally - I do not think this emanates from the founders, because I know a lot of those people, and I know what they care about, and they’re actually fantastic founders, but they also happen to be like really hardcore open source people. The board members come up with a narrative that says “Look, we are deserving of having the right of exclusively capturing the value around this particular technology or this particular product.” That goes against the entire premise of open source, which is that anyone can capture the value permissionlessly. And so the board uses this kind of like weird, kind of backwards justification of “We’ve invested so much, we’ve given away so much. This is a huge ecosystem. Now is the right time to do this.” And so the pattern is they’ve done it well after hundreds of millions in revenue, after their public businesses - like really healthy, impressive businesses, very impressive businesses - and they do that after they’re very significant, and without exception, massive negative response from their communities.

But if you look at their revenue numbers - the other data points that I would look at across those three examples; there are other examples, too. The revenue numbers in terms of growth rate, acceleration, all the factors that you look at when you kind of look at the quality of revenue and so on - completely unchanged. The trend lines, if you snapshot the prior six months or nine months of the license change, and the post six months or nine months after the license change, or even a year or two - completely unchanged, trend line-wise. And the justification on the part of the board of those companies was “Look, we’re going to accelerate revenue, we’re going to have better conversion rates, we’re going to be able to better monetize”, and so on and so forth. And the data disagrees with those conclusions.

There’s a saying in science or in physics that when you have an opinion that data disagrees with it, your opinion is *bleep*. It’s wrong. So that’s the fact. People who made those conclusions about the necessity or the justification of changing the licenses are actually technically wrong, but they’re also morally wrong.

Yeah.

That’s the way I look at it.

And they torch their communities in the doing of it as well. I mean, they’re setting a lot of stuff on fire. Yeah.

Yeah. It’s really – no, they recover from it, because the thing is like… The reason they recover from the massive negative response, Jerod, is – and this is like pretty consistent across these companies… They have communities that are really huge, for sure. Millions of users, and all this. But at that point of the saturation curve kind of thing that I was saying, when they change the license - at that point they have customer bases that are really huge. They have thousands of customers, paying customers. And when the customers largely hear that about the license change – the customers actually really don’t care, because the customers are already paying for the product.

Sure.

They’re like, “Oh, the free thing is like slightly less permissive, but it doesn’t affect us. Okay. Nothingburger.” It’s like a nothingburger. So it doesn’t matter to the businesses. But to your point, the communities - it’s like a breach of trust.

It feels like you just got stabbed in the heart, really bad.

[54:01] So yeah, I’m very against that. We’ve actually put it in our investment documents for our larger investments, like 3 million plus kind of thing… We’ve actually done this in a couple of smaller investments too, but like 20 plus companies at this point where we have meaningful investments, we literally put – even if we’re doing a safe, which we’re not really that into safes anymore, like for equity documents… We put a clause – and all the founders love this, by the way. This is not something we have to like negotiate, and we don’t have any other clauses. There’s like literally one thing that we add in there on top of the standard stuff… That the companies need our permission as long as we are shareholders. We could get diluted down to one share. They need our permission – in the perpetuity of our having any shares in their company, they need our permission if they ever want to change the license of the core technology. And I think people have been following me for a number of years, I plan on doing this for probably 10, 20 plus more years of my life. We will absolutely not allow them to change the license. It’s that simple.

AGPL is a good middle ground. A lot of our companies are AGPL. If you want to change to like a basically non-compete competitive type of license, we’ll say no. It makes no sense to us.

Yeah.

So you wouldn’t be investing in LLaMA?

No, I wouldn’t. Well, largely because we have –

[laughs] I don’t think they need your money there, JJ.

Well, we have other investments. But I mean – yeah, the thing with LLaMA is it’s kind of… Yeah, it’s a brilliant, brilliant move by Yan and team, and… It’s really his brainchild, but Zuck has been taking a lot of credit. It’s going to go back, I think, in history as one of the most brilliant contributions to computing. And they can sue actually at this point more than a handful of companies for violating the user license, and things…

Really?

Yeah, because there’s quite a few clauses in the LLaMA license that they can prosecute people on. But they just relaxed the license terms. I would not be surprised if they just made the whole thing like full Apache, or full some traditional, conventional open source license in the next couple of releases. I would not be surprised. The reason is - and Zuck literally just said this on Bloomberg yesterday. I believe Emily Chang interviewed him. He said they have no need to monetize these models. They have a very strong advertising business, and they’re not selling, they’re not licensing software… This is not a core business-differentiating product, and they can give it away completely. And as a result of giving it away, they reap the same rewards as they did open sourcing PyTorch, and open sourcing Open Compute, which is indirectly saving them - actually, pretty directly - many billions of dollars on building out their compute infrastructure for serving applications that support billions of people, which is what they do.

And so it’s actually quite a beautiful, very aligned company to be embracing and investing so much in open source and open platforms. And you kind of think, like, “Why isn’t Google doing the same thing?” They also print money with advertising. Why isn’t X doing the same thing? I mean, Elon - maybe he doesn’t quite as much print money with advertising; he’s lost a lot of advertisers. But you know, the point still stands. If your business model is not in diametric opposition to your innovation model, you should be innovating like bananas. Like crazy. And your business model should be reasonably sound, and protected, and you have different aspects that compound the value of an advertising business, which is collecting and generating more user data, which is what they do… And unfortunately, that’s the trade. We use these free tools for free, we don’t have to pay for them, and we happen to part ways with our very precious personal data. That kind of sucks. I’m actually very anti that. I deleted my Facebook account many years ago. But it is a very aligned kind of company in terms of acknowledging “Look, we can invent and create crazy powerful technology, give it away, no problem, because our business model doesn’t need us to charge money for that technology, even if it’s costing us billions or tens of billions of dollars to produce”, because we get rewards and benefits elsewhere, that show up elsewhere on the P&L… Which is like compute efficiency, improvement, talent acquisition, whatever.

So I think Meta is a really great example of – I think, frankly, a pretty incredible steward of this open source kind of philosophy in the last few years. I mean, it’s been pretty surprising, sort of rebranding almost from the previous years of lots of drama and blow-ups that Mark Zuckerberg has had to deal with.

Yeah. And there’s a lot of speculation there too, because even with the pivot to Meta, the name, Metaverse, obviously, it’s short for Metaverse, and famously that’s been going not so great, pouring so much money into another tale in Silicon Valley, VR.

Right.

I was gonna try and pull in something, but I’m just gonna skip it for now. Maybe I’ll do it in post.

[59:03]

“If you guys want to feel worse, check this out.

Come join us for an epic gathering where we come together as one people, as visionaries. The most transformative technology event in the world. Hooli-Con.

See you all at Hooli-Con. I hope you’re excited… I know VR.

Yeah, VR.”

Can we talk about Zuck’s hair real quick, Adam?

Sure, please do.

I love his hair.

So we saw these recent round of publications he’s doing, and this tour he’s on, and he has very much this surfer dude vibe now. Adam and I were just talking before you hopped on, JJ, that it looks like a character from Fraggle Rock at this point, very much like overly sunned stoner dude.

Totally.

And I think that that’s very calculated. I think he’s a very calculated person. I think he’s trying to rebrand himself away from that robotic dude who was in front of Congress, and could barely drink a glass of water, and to something completely different. I think that it’s all part of this whole –

He’s wakeboarding with a flag, and… You know…

Yeah, wakeboarding with the American flag. I mean, he’s doing some stuff, and it’s just like – he’s trying to change his brand, for sure.

I think it is calculated, because he is an extremely intentional person, and very calculated. But at the same time, I also reserve the – you know, he’s kind of coming into his own comfort zone, and just sort of like appreciating new styles, and experimenting with different things… But I actually totally acknowledge too that that’s probably a very calculated thing. I think it’s a cool look, for what it’s worth. I think he’s a lot more grounded and seems much more–

He’s head and shoulders over how he used to portray himself, I think.

Yeah. I mean, he was doing M&A, or – M&A… Probably that too, but… MMA, for a bit. Wasn’t he going to like fight Elon, or something like that? Anyways, I’m not trying to [unintelligible 01:00:55.09] the drama. There was like some sort of MMA thing…

Yeah, that’s right.

And he was a pretty tough dude on the mat there. I can see him doing some cool stuff, at least what I saw.

No, he’s jacked. He’s practicing, for sure.

He doesn’t seem so jacked now, though.

I wonder who got him into MMA.

I don’t know…

Do you guys know who got him into M&A? I think it was like maybe Lex Friedman, or something… Because he did like a Lex Friedman interview like a few years ago.

Right.

Probably Rogan.

Maybe Joe Rogan, yeah.

Break: [01:01:19.29]

What I was trying to say there - and I’m glad you mentioned the hair, Jerod, because that is a funny point… Meme photo to LinkedIn and elsewhere coming soon, potentially. We’ll see. Maybe in good taste, maybe not.

Yes. Maybe.

But it was just this transition from – being in this position seems strange, and I’m curious why you think it may be happening, because they were a hardcore pivot away from obviously the brand of Facebook, which was the brand for so long… To the Metaverse, this idea of VR, that we’ll be doing more business there and more things there in our personal lives; to call the company Meta, to rebrand everything… And that was famously going bad, losing a lot of money. There was a lot of headlines about that. And now, thankfully, they’re able to give it away because their business model is so sound. And I’m just wondering, why? Is it just because they want to be generous to the world? What do you think motivated this seemingly generous thing… Obviously, it is generous to just give away –

You’re talking about LLaMA specifically?

Right. Their idea to be Meta, to go down this direction, to sort of – I mean, they have an amazing business still called Facebook, and advertising, stuff like that, and it’s going well. That’s their business model. But they’ve given away this very expensive to create innovation. It seems strange, even though it’s against their business.

It’s not strange at all. No, no, no. It’s not strange at all. It’s not strange at all. I tried to –

Explain why you think it’s not strange.

I tried to articulate it earlier. It’s extremely simple. And Zuck has been saying this publicly. So there’s two examples: OpenCompute and PyTorch. OpenCompute is the open standard, open source designs for top of rack switch, server, spec, chip design… All the data center infrastructure stuff that Meta does, okay? They open sourced that a number of years ago. My partner, Heather Meeker, helped them with the open sourcing. They have reaped billions of dollars in cost savings from open sourcing their designs for their data centers. Counterintuitively. Counterintuitively.

You would think, “Oh, wow. Why are they doing this?” They’re sharing their hyperscale designs with the world. If you all remember, Google was extremely proprietary and secretive with things like Borg, and Omega, and all of their networking Andromeda stuff… Like that was like top secret; a crazy small group of people had all the details of how that stuff worked. Maybe like 100 or 200 people at Google really knew all of the core inner workings of that. And again, this goes back to the view “The world is more efficient and better the more transparent you share cutting edge designs, cutting edge research.”

For a company specifically whose business model does not require monetizing a particular innovation that causes their business to be more efficient, they’re actually incentivized to be transparent and share that innovation with the world. They’re incentivized.

A couple of other examples would be Tesla basically saying “Look, all of our patents are belonging to you.” Microsoft also did that, by the way. This is also kind of shocking. People don’t quite recall… Microsoft opened and said “We’re giving away all of the patents in our entire–” I think 70,000 patents in their portfolio. Pretty recently. That’s an even bigger head-scratcher, to be honest, because Microsoft’s monetization model is licensing their software. And they said, “Look, we’re not going to go after anyone that infringes on our patents.” That was a few years ago as well.

The second project that Meta released that created huge derivative results and benefits for them was PyTorch. React also, by the way. React is just this enormous ecosystem at this point. So PyTorch and React, I think we’ll just focus on those two.

PyTorch allows them to both manage, shape, and influence the trajectory of the way deep learning and neural network development, hyper-optimization, and shipping work - these are artifacts called neural networks - because it’s the industry standard; completely demolished and destroyed TensorFlow, if you remember TensorFlow. PyTorch is the standard. It’s not even a comparison at this point. And then what React did was allowed them to have a similar type of influence and trajectory over the talent base of people building front-end web experiences, which are largely part of the entire portfolio of products that Meta releases. And then – okay, so again, also hugely derivative, positive talent acquisition efficiency benefits for their business, probably measuring in the billions of dollars.

[01:10:24.00] And so how much money did they actually spend on open sourcing and releasing these things? Maybe a couple hundred million dollars for both of those projects. Maybe a hundred million or so for React, I would estimate, and probably a couple or a few hundred million for PyTorch at this point.

Okay, now your question on LLaMA. Now we’re talking about not just a human thing, where you have Sumit, and Yann, and 100 or 200 people… There’s a decent payroll there. But we’re talking tens of billions of dollars in hardware that they’re buying, building… Buying and building. And they’re transparent about all these numbers. Like, they’re saying like what the numbers are; it’s tens of billions of dollars per year. Probably something in the order of 30 to 40 billion this year, and next year maybe even more. So to your question, “How the hell is this justifiable?”

How does it make sense?

It’s just as justifiable as the two other examples, but in a slightly different way, which is that not only do they get to shape the standard trajectory of how these LLMs are built - because LLaMA is an LLM development framework effectively, and it’s coming with these different base sizes of models that people then fine-tune and quantize and implement and deploy in all kinds of different ways… But what also happens is they are running different versions of LLaMA, both in the ones they’re scaling internally before they release them, which would be like another dot release for LLaMA, maybe LLaMA 3.2 or LLaMA 3.3, which would maybe be trillion-parameter models. We don’t know. I mean, I would guess they would probably wait until 4.0 before they ship a trillion-parameter LLaMA, which is probably not more than nine months away from that, maybe a year at most. They’re running their internal applications, which are all getting transformed, by these transformers, which I think the transformer architecture is like extremely inefficient, but it’s pretty powerful and amazing… And they’re completely reshaping the way their core products, that already print money to the tune of $150 billion a year in advertising, can deliver even more compelling, contextual, personalized experiences to the users. I guarantee you that is going to make a very material impact on the efficiency, productivity, and growth of their business.

And so there is an indirect benefit from saying “Here’s LLaMA.” Both it gives them a huge advantage in terms of the speed of innovation, that increases the rate of improvement of this thing… Because there’s a huge developer ecosystem around LLaMA. If you compare it to some of the top open source projects, it’s probably on the trajectory to be within the top 10 of all open source communities within the next couple of years. It’s on that trajectory.

What kind of engineering improvement and innovation acceleration efficiencies do you get out of that? It measures in the billions of dollars. And then if you’re a company with a portfolio of products, like Instagram, Facebook, WhatsApp, and other things, that are all deeply, deeply getting infused with AI to transform the experience of these applications, and you want the absolute best industry-standard model to be running those, that you just so happened to create yourself, open sourcing is literally the only way. They can’t go and say “Here’s proprietary LLaMA, that you can run anywhere. It’s compact, but it’s totally proprietary. Closed weights. Trust us. Just trust us.” And maybe Zuck still looks like a stoner hippie, but he’s like saying “Look, bro, trust me.” Is that going to go down well? I really don’t think that would go down well.

I don’t think anybody trusts him as is.

Probably people still don’t trust the guy.

Yeah.

I still don’t fully trust the guy.

That’s why I asked the question, how does this make sense? I get the rationale that you’ve proposed here…

I mean, it makes sense because it’s verifiable. It’s verifiable. I mean, their paper – I encourage you guys to read the LLaMA 3 paper. It just came out. It’s very detailed.

[01:14:03.10] I get it. I’m not arguing that. My question was from a position of disbelief, I suppose, or unawareness.

Are you a believer now? Did he make you a believer, Adam?

Well, here’s what makes me a believer, to some degree… Is I am an Instagram user by way of using the application, not so much posting. And I will say, my experience with –

You’re a lurker. You lurk.

Yeah, I’m a lurker. I will say that my experience on Instagram with advertising is by far my favorite experience with advertising.

It’s pretty good.

It knows what I want. I’m always happy to even say – I have a folder called Sponsored, that I save ads in, so I can go back and check them out.

Wow. That’s pretty impressive.

I think it’s so cool, really. And they’re so relevant to me.

That’s a great example. That’s a great example. It improves the advertising contextual relevancy.

And so I trust Instagram more because I enjoy my experience there, even if I’m not posting. I’m enjoying friends and family content. And while I’m there, I might be on there for - Jerod, I’ve told you this before, I’m an up-and-coming chef, who knows… I’m there studying things, I’m paying attention to recipes… They’re advertising things to me… I’m not going and buying them…

He’s bookmarking sponsors…

…but I’m definitely dropping it into my saved folder called Sponsored, or Cool Products, or things like that. They get me, and I dig that. And I’m cool with that.

Okay. I want to say a couple things that balance out my massive bull case that I just laid out there. I personally think the transformer is extremely inefficient, and like a brute force instrument for creating really powerful AI models. And I think this whole “Scale is all you need” narrative, and “All we need to do is build a half a million or a 10-million GPU compute cluster and we’re going to get to AGI”, I think all of that is utter horses**t. And I think there’s very diminishing returns that we’re going to reach, and I think the architecture is going to be viewed – the transformer thing is going to be viewed in a few years as today we view Fortran or COBOL in comparison to Rust, or to modern languages and modern compilers.

So with that view, which is a very technical view - and I can elaborate more on that if you want, but it would probably take a lot longer - I think that what Zuck is doing is actually totally insane. Okay? Because he’s basically mirroring what the big labs are doing, except he’s like sitting on a hundred billion dollars of cash, and going “Ha, ha, ha! You guys are trying to raise a hundred billion from the Middle East. I have a hundred billion right now. And also, I can do very effective liquidity transactions and leverage on my trillion-dollar company.”

By the way, I think Zuck is probably the most powerful technology founder in the world, above Musk, because he has a trillion-dollar company today, that he solely controls with his super-shares. Not even Larry and Sergey from Google actually kind of approximate or come close to that, because Google’s two, two and a half trillion or something, and it’s two guys. And one’s kind of like unknown, checked out where he is, and the other guy is sort of like bouncing around at AGI House and trying to help Google figure out how to build competitors to LLaMA. Which is Sergey.

So Zuck actually is playing the same game that Sam is playing, except he’s like this proprietary stuff. It’s going to create a totalitarian communist future. Let’s build in an open way, which is awesome. But I think the way in which he’s going about doing it is actually really flawed. And so we could see, in a couple of years, even though I’m still fairly optimistic about what you actually just said, Adam, like “God damn, my advertising experience is so much better. It’s like so relevant. It knows what I want. The thing I was kind of thinking about over lunch or whatever, there’s an ad for it. And I was kind of thinking about it… My God, I’m going to go buy that thing later”, or whatever. Like, that’s the best possible advertising. It’s not invasive, It’s not creepy… It’s like really, really specifically relevant to like what you want. That’s AI.

AI is making that happen. There aren’t programs that developers are writing. These are neural nets that are continuously learning based on all kinds of new inputs coming into the model, giving you a better recommendation for something.

[01:18:09.00] So even if Meta loses 50 to 100 billion on super-sloppy, crazy, inefficient transformer-based neural nets that are doing language model token generation, they could still recover those costs by producing a better advertising experience.

However, if they don’t actually have that much of a quantum differential on the cost that they’re going to be putting in the ground, which is probably to the tune of $100 billion, and the improvement in their advertising business, in reflection, they will actually look dumb the moment a transformer architecture replacement comes to light and shows the world that actually “Look, tweaks and optimizations can get you 5 or 10 or 20 times better performance on the transformer.” But there’s this new thing that’s actually 500 times to a 1,000 times better. At inference and at training time maybe it’s 100 times better. You cannot compete with that. Everyone will switch, and it’s going to be like “Wow, we actually only need 200 or 300 H100s to produce some unbelievably powerful system. We do not need 50,000 H100s”, let alone B200s, which are also still fairly optimized around this transformer thing. And so the AI bubble house of cards moment at that point will basically reveal that - whatever, the emperor has no clothes, and NVIDIA stock price will correct rapidly. I’m not recommending to short NVIDIA, by the way.

Why couldn’t they be the ones to also create this other innovation you’re talking about?

Startups, universities, labs… It’ll come from a lot of different places. Unfortunately, I think the inertia at Meta and a lot of the big companies is such that they’re not willing to run the kinds of – I mean, Yann LeCun would say otherwise. And I talked to Yann a bunch, and I think he does actually know about the very specific technical views I have on this kind of stuff… You make a good point. I mean, theoretically, Yann should actually have some of his people building alternatives to this. And in fact - look, actually, Yann has been very critical and negative on what he calls autoregressive LLMs and this transformer architecture being super-inefficient. But at the same time, look, that’s producing really powerful and transformative capabilities for them. And so they’re going to continue investing. As far as what we’re hearing from Zuck, the dude is totally comfortable eviscerating tens of billions of dollars in cash.

Yes, he is.

And in order to build better models than Anthropic and OpenAI… Which, by the way, are struggling, really struggling to raise the capital and support the capital needed to compete with LLaMA. So what people were saying six, nine months ago about like LLaMA is going to take forever to catch up… LLaMA is basically parity, and will flip and maybe way accelerate. The open models may be way ahead of all the closed models by the end of the year, potentially. We don’t really know what is GPT5 going to look like, what is Claude going to look like…

Right.

Anthropic’s been beating Open AI now for two months on the benchmarks and performance. So Anthropic is like actually way ahead… But even Anthropic doesn’t have tens of billions in equivalent compute to scale. Like, Meta is literally doing this. They have hundreds of thousands of H100’s-equivalent from the last report that Soumith put out, who’s the PyTorch creator, and kind of leading a lot of this stuff. And they’re probably going to have half a million or a million more… And they’re literally just getting started with like scaling all this infrastructure.

So I take two sides to it. One is I think it’s justified and it makes a lot of sense, in the same way that releasing OpenCompute and React and PyTorch made a lot of sense, to deliver massive derivative benefits to their business, that were measurable. Really huge benefits. ROI - very easy justification.

On the side of LLaMA though, you have tens of billions of dollars now you’re spending. It’s not a few hundred million dollars or maybe a billion dollars. It’s like tens of billions. And so I still see some returns there, and you could see maybe a multiple of that if you look at the enterprise value of the business.

[01:22:09.15] But I don’t see the transformer approach they’re taking as the most efficient and effective way of actually continuing to innovate and scale the capabilities of the models. I think we have much better techniques. I mean, you have to really understand - what is the transformer doing? The transformer is taking in tens of trillions to pretty soon hundreds of trillions of what are called tokens; each token is a couple words, and split apart, and so a token is like a few bits of information, basically… Some tens of bits of information. And then you have hundreds of trillions of these things that you’re compressing and training into a system.

The human brain uses 10 to 30 watts of energy to compress what you could approximate as probably tens of exabytes of data by the time we’re wobbling around and able to like string together the first few coherent words. The amount of energy getting consumed by these transformer-based neural nets is gargantuan. We’re talking about tens of megawatts. It’s crazy.

The human brain is many, many, many, many orders of magnitude more energy-efficient than the state of the art neural nets that are getting produced today. And I do think the human brain resembles something probably roughly approximating a biological neural network of some kind, with electrical signals between our synapses, and all that kind of stuff. I think that’s kind of like the Red architecture. But to do that in silicon, at our current approaches is as if we were doing vacuum tubes in the ’20s and ’30s. That is our current state of the art sophistication in building neural nets. We are in vacuum tube days.

Yeah.

We’re trying to do like crazy advanced linear algebra, and the ways in which the mechanisms we’re using to do it are like vacuum tubes. That’s the sophistication of the current transformers.

And the reason that’s important is obvious, to some degree… It’s energy consumption, right?

Yeah.

Like, you want to consume less because energy consumption equals dollar bills, or whatever currency you spend.

Not only that. Power is a huge component of data center design and cost structure, for sure. The other part is the chips. Literally, buying and assembling and interconnecting with high-speed interconnect the chips themselves - holy s**t. Extremely expensive, too. Profit margins on the NVIDIA chips is like 80%, 90% last I checked, gross, and net, 50%-60% net profit on those chips that NVIDIA is selling.

Love it.

These are software margins. These are literally software margins. We are not in an efficient market. I don’t think efficient markets exist, but competition causes things to come down quite a bit. So I do think – going back to Meta, I think that there’s just going to be innovation and things that come to light that will in retrospect make their current execution and eviscerating huge amounts of capital to create LLaMA, which I think is a net good and amazing for the world; even if it ultimately looks a little silly for their business, it’ll look really brilliant in retrospect. But if you just look at it selfishly from the standpoint of the P&L of Meta, I think what they’re doing is actually pretty, pretty wasteful. That’s just generally my view. But I still defend it, I still defend it, because the general contribution and the phenomena that LLaMA has created in this tech industry is kind of hard to measure. I mean, it’s pretty incredible.

It’ll certainly be exciting if and when it flips the proprietary models here. Like you said, nine months to a year. I mean, that will be quite a moment. Leaving Zuck aside, but still talking business models and incentives… You’re a fan of open core, right?

Yeah. Open core - yeah, for sure. I think it means different things to different people.

So I have mixed feelings on open core, and I think that it can sometimes misalign incentives for the creator, for the founder, for the software maker, because they are constantly battling the question of “Where does this feature go? Does it go in the open core or the proprietary wrapper?” And I feel like that’s always a struggle, and people sometimes make the wrong decision, they make the right decision…

[01:26:14.29] True.

…and I think that can be problematic. I’m sure you’ve helped people through those decisions, or a lot of your investments probably are open core companies, right?

For sure. Yes, they absolutely are. I’ll kind of just summarize briefly what I think open core is. I think open core – it’s not a licensed model. It’s not a replacement for open source. It’s not antagonistic with open source. It is a business model and a business architecture, if you will, for commercial open source companies.

Red Hat is open core. When I say this to Red Hatters or people who are like the free software people, they lose their s**t. The reason is they go “No, no. Red Hat does not have any proprietary software. It’s all open source. You’re crazy. Open core means something really toxic and bad. Look at all these blog posts.” The reality is every business is open core that pursues this commercial open source approach, because your core technology is permissively licensed in open source, but the crust, the wrapping, the frosting kind of around it is where you have some thing that you’re charging for, which you can call intellectual property… In some cases it can actually be open source, but you’re charging for it because only you are uniquely able to deliver that to the customer, and solve the customer problem. And so that’s my view of open core.

Yeah, open core can – so what you said around picking the features and like the trade-offs and the tensions… You know, this is also apparent in every single business ever created since the beginning of capitalism. The difference is that you just don’t see the differential. What you see as the customer is just this. I’m holding up an iPhone. You just see the final product. You just see the perfectly packaged thing that the company that created the product assembled, compiled, and is selling to you. You didn’t see the massive avalanche long list of features and trade-offs that they had to think about on “Do we leave this out? Do we put this in? What about the next version? Do we add that? Do we not?” It’s called information asymmetry. It’s a concept in computer science called – it’s actually an information theory. Claude Shannon style information theory. We basically have – one person can see some of the picture, the other person can see all of the picture.

Sure.

When you have that dynamic in any type of system, there is inherent tension, frustration, and kind of confusion because one person knows some of the stuff, the other person knows all the stuff. So that in open core companies or commercial open source companies, we have like the free open source project with – very transparently, you can see literally everything. And then you have the commercial product where you can only see the things that the company decided to implement in the commercial product, which are usually extensions and additions to the open source thing.

Sometimes customers get frustrated with this, but oftentimes it’s more developers. What I’ve noticed is community people are like “Look, I’m not going to be a customer, but I really want this paid feature to be in the thing.” The best practice that we have advised our companies to sort of adopt is expectation setting at a very high, rigorous level, and very clear levels of communication. I think if you have those two things, you really manage to mitigate all of the potential issues that come from this frustration, and this kind of like chaos in the community. And so far, I haven’t really seen any major issues.

Now, I will say, in the past 10, 15 years, maybe 20 years, close to that now, there have been examples of companies that pursue this open core approach, but they view specifically the way in which they’re building the open source project as a diminutive, crappy version, that’s really limited, and is specific to this particular type of thing. I’ve never invested in a company like that, I don’t support those companies. I think that is actually really duplicitous and not a good way to build a business… Because what it does is it subjects the core open source technology to a perception that it is lesser than, and insignificant, and is only meant for a couple of trial scenarios. Honestly, kind of like indistinguishable from like freemium, where the minimum thing is like sort of you can do a couple of things, but you can’t do everything, and then if you really want the good stuff, you’ve got to pay for the other –

[01:30:27.04] It’s like a teaser for the product.

Yeah, exactly. That’s not a good dynamic. In our portfolio, dozens of companies, you can look at – really all of them, I think, adhere to this. The open core, the core open source thing is completely unlimited. You can do whatever you want with it. Yes, it might be missing some features, but you can certainly add a few of those features if you want. But it’s not like caged in for a specific type of use case, or a particular type of technical domain. You can do literally what you want with it. They tend to be more geared towards individual users, or a single computer or a single workstation. They tend to be less geared towards “Here’s all the features that you need for an enterprise, or an organization, or a team, or a group of people, or a large group of people”, because those features tend to be more complicated. You have more coordination complexity, there’s more business logic that you have to maintain… The codebase becomes exponentially more complex. And so when you have a high degree of complexity in the logic of an application, it justifies charging for that logic, because it’s hard to build that thing. In the same way that you guys, I’m sure, are happy paying customers of Riverside, there’s probably a free version of Riverside for solo/indie…

You keep outing us here. Twice this call.

Well, I think it’s a good company.

I’m just kidding.

I’ve heard really good things about it.

We’ve been happy customers for years now, so…

We like Riverside.

I’ve heard very good things about Riverside. By the way, one thing I forget and I neglect to say a lot when people hear me or whatever on the internet, they think I’m some weird guy - I’m not an open source zealot fanatic. I don’t need the whole world to be open source. I have an iPhone, I’m a happy Apple customer, I use a lot of proprietary technology in my life… I just so happen to have a view that when you create systems, processes and companies that accelerate and expand the universe of open source in a sustainable way, that is going to be a net good for the future, and make the world better for my children, our children, and the future of humanity. That’s really important. We’re not going to be here that long. World’s probably going to stay around much longer than most people think, and we need to think positively about this.

You mentioned Nadia… One of the things that I think is very counterintuitive as well is sustaining open source, and open source sustainability, and open source funding - all of these arguments are nonsensical, in my opinion. They make no sense. I disagree with all of them. There is no such thing as the open source sustainability dilemma, and open source needs more funding, and all of this. I disagree with all of those. None of them make sense. Yes, there are a lot of maintainers and people who do not have jobs, where they’re not funded enough… But I think people fail to understand - and this is also related to open core expectation setting. People fail to understand that open source - not open core, or any of these companies, which is also kind of true - is an opt-in system. You decide, as a human, to spend your time doing this thing. And if your technology becomes crazy, super-successful, you have massive amounts of users and people waking you up at 2 AM demanding that you do something, you owe them absolutely nothing. You should not respond. You should have no feeling of obligation, because you just woke up and decided to create something beautiful for the world. That’s great.

Now, when we say “open source sustainability problem”, I think it is better framed as an open source expectation setting problem. People are not creating boundary conditions for their time and their energy. And by the way, in terms of open core companies setting expectations, open source maintainers should do the exact same thing.

[01:33:56.18] I agree.

I’ve created a hugely successful project… Wow. Amazing. Guess what? Here’s what I’m willing to do, here’s what I am doing, and here’s what you should expect of me in the future. Anything that deviates from that, I will not respond. That is completely not only reasonable, but very admirable. And those people should be celebrated. There are so few of them. So few of them. And instead, what people do is they mistake and misunderstand, partially because I think of a lot of weird, twisted framings that Richard Stallman actually proposed early on, because he had a much more ideological view of this, which I think is not helpful…

In a pragmatic sense, when you create an open source technology, you are opting into creating a permissionless innovation thing of some kind. And you shouldn’t feel obligated to the world of your users and your people impacted by your technology that you should work for them and do anything for them. And in that type of world, there would be fewer instances of burnout, frustration, pain, suffering on the part of open source maintainers. And so for us, investing in open source companies is a way of creating a capitalist system and structure around sustaining an open source community, while also creating a business model for it, which expands the universe for open source and creates the right incentive system. But what it also does is it says “Look, you should be really, really good at setting expectations, maintainer… Because you’re going to work with a professional investor who also has other investors invest alongside after that fund.” Many of the best investors in the world, in fact. “Those people will help you and advise you.” And as an open source maintainer, learning how to create a company gives you the skills of learning how to create and set expectations for many complex stakeholders, employees, customers, and investors, to name a few, that you didn’t have before, and it gives you the ability to get really good at setting boundaries on what you are willing to do, what you will do, and what you are actively doing to the people that support you.

And if you translate that, and you kind of reverse that into the way you run open source communities and projects, the world would be so much more efficient. I think a lot of forks happen because people are like “Look, I disagree with the maintainer” or “They’re not doing this thing, or not doing that thing.” Those things wouldn’t have happened if you said “Look, you can always fork the project. That’s the fundamental of open source. But this is why I’m doing what I’m doing, and this is kind of the expectation you should have.” So I think expectation setting is really fundamental and super-critical for people to kind of learn. It’s hard. It’s a very difficult skill to learn, but it’s very, very important.

Just marinating on this… We’ve talked about clarity and expectation a lot. I know I have, Jerod… And I don’t disagree with that premise, because - yeah, if you’re clear with what your boundaries are, what your constraints are, and somebody tries to violate those things, then they’re not your friend, or they’re not for you. Maybe they’re trying to be helpful, but you don’t owe them anything, because you’ve set your boundary pretty clear.

I think you owe them a non-response. That’s an action. Inaction is an action. Sometimes inaction is the best action, actually. Not doing a thing.

It’s hard to tell from what you said if you feel like funding an open source is flawed. I couldn’t recall which exactly –

No, I don’t think funding in open source is a problem.

Right. As an investor it doesn’t make sense, because there’s no capitalistic engine there to return the capital, right? That does make sense. But you can do it as generosity, or altruistically, for example…

Sure. I’m not against philanthropy or donations. I just think that they don’t actually massively move the world forward. And I think that public good funding is important, government funding into universities, that’s important, and obviously, tuition, to a certain extent, maybe has its use… But I actually think – I’m kind of going on a rant here, but I think the gatekeeping for knowledge and education in modern academic world is very flawed and broken.

[01:38:09.12] And those institutions are basically hedge funds that are sitting on mountains of capital that they don’t really know what to do with, and investing it in all kinds of asset classes, and stuff… So I tend to think donations and philanthropy into open source in general are fine, but I don’t think that they actually serve to solve the fundamental problem, which is that you need a sustainable – I don’t think philanthropic donations are actually sustainable. Unless we lived in a world where it was absolutely required that at the end of your life or towards the very end/tail of your life, last decade or last 15 years or something, you absolutely had to give away most of your wealth philanthropically. In that case, maybe. But that is not what all of the wealthiest people in the world do. Most of them keep it in their companies to compound for future generations.

I would say some of the time there’s large philanthropic programs, but the reality is those philanthropic dollars wouldn’t have been made without the engine of capitalism that came before it. So I still have a hard time really thinking about “Okay, is that the solution?” It’s not. We need more capitalism. More capitalism is a solution, actually, in fact. More capitalism is necessary.

What about a scenario where a developer is creating something they find is valuable, and they put a call out to the community, they’re making it open source… And maybe this is part of the clarity and expectations point that you made before, so that might be true… But they’ve said “I can only do this if I’m funded”, which is open source sustainability. That’s a sustainability effort, right? “I can only do X if y’all care about what I’m doing.” And I’m from Texas, so I say “y’all.” “If y’all care enough about what I’m doing to support my effort.” And they go on GitHub Sponsors and they sponsor that person. Is that open source sustainability to you?

No, I think that that’s a person asking for help and getting some help. I don’t think that’s open source sustainability. I think that’s a person basically saying “Look, I don’t have the ability and the wherewithal to do this. It sounds like everyone wants it, and I have the unique capability to do it.” By the way, if it’s that viable to the world and so many companies want it –

Well, these aren’t companies in this case. It was just people loving what the –

Content as individuals.

Yeah. In this case it was like not a company.

I think if it’s just individuals, then yeah. To answer your question directly, I would not call that open source sustainability. I would call that someone who has a following of people that believe in them, offering help and willing to chip in and put some money in a hat, and help that person. I don’t think of that – so the word sustainability means something specific to me.

That’s what I was trying to – because you said it pretty stark in terms of your response to Jerod’s question… And so I guess I’m trying to get to the lens of what you mean by sustainability.

To concretize my view of what I think sustainability is - which, again, human language is so horrible and brute force. Everyone has different definitions of different words and stuff. My view of sustainability is that you have a system, it’s like a system, and you have continuously more input going into the system relative to the outputs going out of the system. So the outputs will produce and generate more inputs into the system. Okay? That is sustainability to me. I mean, it’s kind of the same as saying you have a profitable business. More revenue than spending. This is really basic stuff. Many other examples that we can use.

These projects are not many businesses, what you’re saying. You don’t think that these open source projects should transcend from communities to many businesses that aren’t really businesses. Not necessarily disagreeing with that either, by the way. So the trend line, I think, for capitalism right now is it’s becoming more and more pervasive. Fractal. And there are better tools and representations of value - one of which is crypto, by the way - that will allow people to participate in capitalism more easily, in a more fractional way in the future. You can come up with some buzzword for all that. I don’t know what that would be. I do not think philanthropy, donations, chipping something into a hat is sustainable. Nope. That is not sustainable, because it is not a system that creates more inputs relative to the outputs. It’s not.

[01:42:17.21] So your example - the way I would describe your example is this person is respected by a community, they raised their hand, and the community responded and said “Look, you asked for this. We’re going to chip in. There’s what you asked for.” Great. That is not sustainable, though. It’s like a one-off activity. How many more times could that individual make such a request? I mean, I guess it would depend on the value of what they’ve done before… Things are going to get, I think, pretty weird in the future as these AI systems can produce very valuable technologies… And so how do you actually sustain the notion of value, and then connect a currency to that? I think that we’re careening - not gradually, we are careening towards a world of completely repricing and reshaping the way value is represented, both in terms of basic standard goods and services in the economy, like bananas and bread, to digital products, and also the price of advertising, which is like the predominant business model of the internet still today.

So I try to remain humble about – like, I have no freaking clue how these things are going to get solved… But I deeply believe capitalism is a very sustainable thing. And you want to create a continuous stream of revenue and income, and have your cost structure and your spend be less than that. To me, that’s called sustainability. There are many other – I’m just like a business person, sort of, so I kind of use the business example… There’s many other examples that you can use… Like in physics, and systems that have inputs and outputs; the same principle applies. To me, that’s how I think about sustainability.

Well, even in business, when it comes to capitalism for example, money exchanges hands when problems are solved. That’s capitalism.

Yup.

That’s why the bigger the problem that you solve, the more money that can exchange the hands. Oor the value, however you want to term – currency, bartering is how we began, then it turned into things you can actually exchange… Those things became true currency backed by local governments, and nation states etc.

Yup.

Yeah. If you’re solving a problem for somebody, value gets exchanged. Typically in currency.

Yeah, typically. Typically. But I mean, it’s kind of interesting, the degree to which you can solve problems with five people has dramatically changed relative to what you could have done 10 years or 20 years ago with five people.

Well, JJ, you’ve given us more time than you said you would…

Yeah, I appreciate it, man.

I’m really happy to chat, guys. It’s been fun. It’s been a lot of fun.

It has been fun. I appreciate the viewpoints you’ve shared. My initial question was just to provide a map to the developers that listen to this podcast… Like, “Is open the way?” Is open the way, whether it’s capitalistically – like you had said, the crust, the frosting, capitalism frosting around open source… Is it the way? And I think you’ve shared a lot of points that suggest that it might be.

The last thing I’ll say is capitalism is philanthropy. Capitalism is philanthropy.

Tweet that.

Because you are creating something that changes the world, makes the world better, and it just so happens to have a sustainability mechanism built into it. Whereas philanthropy might change the world, make the world better, do all this stuff, but it relies on capitalism to exist. Which one’s better?

So philanthropy is a derivative of capitalism, obviously, right? You probably couldn’t have – unless there was inherent wealth, or whatever else.

Correct. But the fundamental truth though is that capitalism done right, done well, is actually a great form of philanthropy. I argue the ultimate and ideal form of philanthropy is actually capitalism. So that’s a super-controversial thing to say, for sure, but I’ll just leave it at that.

Well, let’s leave it there. Thanks for saying that here on the show, and thanks for sharing –

I’ll see you again in five years, guys.

Yeah, man.

Hopefully sooner than that. Hopefully sooner than that. But JJ, always a pleasure. Thank you so much. Good to have you on.

Thanks, guys, for your time.