Stephen: This is your host, Stephen Sargeant, the Around The Coin podcast. We are going back to back when it comes to investment firms VCs. This is a great conversation with Seth Rosenberg. He's a partner at Greylock. They're investing in pre-seed, seed or series A companies that are powered by ai. They love FinTech, they love ai.

He talked a lot. Little bit about stable coins and payments. This is the episode that our audience loves to hear. We get great insights from the people that are investing in companies just like yours. We're hearing straight from a person that has been doing this for the last eight years. We go all around the board where AI and what, how do they invest in AI with all the AI rappers.

The emergence of technology. Where is AI really gonna place the human element and what jobs are going to take? It's not what you think. Definitely listen to this episode. Some interesting insights on AI might not just be going after the creatives, but the senior managers in every organization. And we talk a little bit about Greenlight, which is a major company in the AI slash.

Financial crime, compliance world, and exactly what other things Graylock is bringing to the table other than just investment, including recruitment, expertise, division, the grow to market strategies. This is an amazing episode. If you're a founder or if you're an investor, or if you're in AI or crypto or blockchain, you're gonna get a lot from this episode.

Let me know how you like it after you listen, talk soon.

Stephen: This is your host, Stephen Sargeant Around The Coin Podcast. We're going back to back with our investment focused podcast.

We have Seth Rosenberg, partner at Greylock, Seth, introduce yourself to the audience. Give us a little bit of background story about you, and then we're gonna jump right in to some, you know, the insights, what's trending in ai, crypto, and everything in between.

Seth: Yeah. Hey Stephen. Thanks for having me. Excited to be here. Excited for the conversation.

Yeah, I'm a partner at Greylock. Been at Greylock for about eight years. Background is in product. Originally grew up in Winnipeg, Manitoba. And I I recently opened up the New York office for Greylock, so I'm based here in New York.

Invest in both San Francisco and New York in kind of first check into amazing technical founders going after big markets.

Stephen: Now Winnipeg, Manitoba usually doesn't come up on, you know, the founder's bingo list. I'm a Canadian too. I'm based just outside of Toronto. What was it like, was there an entrepreneurial community, a tech community, or were you very young when you were Winnipeg before you started getting into like tech or investment?

Seth: So the cool thing about growing up in Winnipeg is you know, for whatever reason I've always been a very ambitious person and the. Outlier success that I was exposed to in Winnipeg were all entrepreneurs. Not technology entrepreneurs, but kind of classic Midwestern entrepreneurs, because in Winnipeg, for better or for worse, there were no hedge fund managers.

There were no CEOs of Fortune 500 companies, and so the. Ecosystem of people who are really, you know, leaders in the community were people who either started or were running multi-generational family businesses. And in some cases these were multi-billion dollar companies, right? So there were stories of, you know, one of my friends grandpa immigrated from Italy to be a truck driver to send money back to their family.

Then ended up building the largest private truck company in North America or another friend's grandpa, like bought a bankrupt TV station in North Dakota, ended up building Can West Global, which is a multi-billion dollar public company. And so for me, that was always the North Star in my career and I kind of realized a little bit later on that that technology is actually the entrepreneurial opportunity of our generation, and that's what got me into tech.

Stephen: That's super interesting. Are you big in haka? I know both of our teams, the leaves, Winnipeg Jets got early exits. Unfortunately I was at game seven. You know,

Seth: Oh, amazing.

Stephen: wish there was an email where you could send in like, Hey, I don't wanna complain. I just wanna see as. Or any kind of reimbursement, you know, process you guys have here.

But my kids did make it on tv. So that was the,

Seth: Oh, that's awesome. You gotta send me the clip.

Stephen: How about you? What are your thoughts on, you know, the Canadian aspect? Are you big in hockey there?

Seth: Yeah, it's, it is funny.

I'm on the board of a Canadian FinTech company called Pine. That is building kind of the wealth simple for home ownership in Canada. So like, you know, starting with low cost mortgages and expanding into kind of a full service search and discovery, everything to do with home ownership, they're digitizing it.

And so I'm on, on the board of, of Pine and during the playoffs a few weeks ago, we had a board meeting and I said if the Leafs and the Jets made it to the finals, I was gonna personally sponsor the whole team to fly out and we'll get a suite and we'll have a good time. Unfortunately, I felt confident making that bet because I knew

Stephen: Super

Seth: happen.

Stephen: Super confident. They probably made you sweat a little bit. I, I think, but, you know, that's interesting. I didn't know that's exactly what Pine did and you know, as a homeowner in Canada and looking around at the market, can you give any insights on, you know, I think Pine's probably stock raised since, you know what's happened in the market.

It's so hard to find affordable housing in Toronto, Vancouver. I'm not sure what it's like in Winnipeg. What are your thoughts on the market and fintechs like that coming into digitalize this space?

Seth: Yeah, well, I think FinTech and Canada is extremely underrated, right? Canada is GDP compared to the US is about a 10th. The market size for financial services in Canada is two x per capita, so it's about 20% the size of the US and I think it's kind of a Goldilocks market where it's not quite big enough for it to be the top priority for international expansion of like the global players, but it's still big enough for the market cap of RBC to be over a hundred billion dollars.

Right. And just the market, like if you look at especially consumer FinTech. The pitch of most consumer FinTech companies in any geo is you start with X product and then you expand and become kind of the single, the one-stop shop for someone's financial life. Right? And Canada is a market where you can really do that 'cause they're there, there are real barriers to entry.

The, the level of service people get because of the oligopoly of Canadian banks is. Is weak. There's a lot to be done like around modernizing the financial life of a Canadian. And so we saw this with our investments in, in, in Wealthsimple, right? We invest in the last two rounds of Wealthsimple and they've done a phenomenal job of just being really customer oriented and also proving out this cross product thesis you know, from wealth to banking to tax to trading like wealth.

Simple is really built. Kind of a full service business. And we have the same thesis with, with Pine, right? Like it's, this is one of the number one things on any young Canadian's minds is home ownership and the cost of home ownership. It's, it's globally terrible, right? The, the state of the Canadian home, home market and that, that's the problem that Pine is going after is basically like owning the end-to-end system, making it much more efficient and therefore much cheaper for Canadians.

Stephen: To your point, the market is a ripe for disruption. You know, as a Canadian, my whole life, even recently up to last week, trying to send a wire transfer, the bank won't even gimme an email of the information I need to give to the client to send me the wire. They're like, I'll give it to you over the phone.

You know, whether it's having a bank account, a mortgage, or to any of your points. It's completely siloed. It's, it's not, and you know, the experience is, you know, you have better customer support probably, you know, at a local restaurant or a fast food joint because the customer service is almost non-existent within most of the majority of the financial, traditional financial sector.

I would love, you know, you talked a little bit about the Canadian market. I don't want to jump around, but you know, it's big news. Last week we heard Robinhood acquired. Wonder Fi, which had already acquired several different crypto entities with various licenses they were one of the biggest crypto service providers.

What are your thoughts around mergers and acquisitions, whether it's in Canada or the us? I know Lena Khan has left the CFDC and that's opened up a lot of the m and a season that we're seeing, whether it's Stripe, you know, acquiring bridge, a lot of conversations around circle. Can you give us, you know, your thesis on what's going on right now in the m and a space?

Seth: Yeah. So at Greylock, you know, we're investing like first check investors into kind of, potential exceptional founders. Going after large markets, and for the most, for the, for the most part, our, our thesis is around, you know, could this company be a public enduring company? And we're usually not investing you know, for an m and a outcome.

I do think kind of this regulatory environment is a little bit more. Open to m and a, which I think is just good and healthy for the overall startup ecosystem. But I would say most of the investing that we do, we hope for our companies to be the acquirers versus the acquired. And I think it's kind of the standard thing, which is like, it, it's, it would be a failure if kind of your m and a strategy is the core differentiator of what you're doing.

I think it's, it's very rare where you can have true transformative acquisitions.

Stephen: You know, looking at your, you know, previous career, you brought up the product side, you worked at Goldman Sachs, you were a product manager at Facebook. Any key takeaways from, you know, working with one of the biggest fan companies and these institutions from a decade ago that served you well as an investor today?

Seth: I mean there, there's a lot of different directions there. I think advice for kind of young people just starting their careers is. Finding places that are aggregates of extreme talent density is so important for your career. And Facebook and Meta at that time in the early 2000 tens especially, you know, some of the emerging products like Messenger was an amazing kind of draw for, for really talented builders.

And so. It's, you know, and now a decade plus later, I still work with a lot of those people. I actually, I just had breakfast with David Fisher this morning. Who's the CRO of, of Facebook. I backed Henry and Keith in tome, which is now light field. Were some of the you know, the most talented builders that I worked with, especially on zero to one at at Meta.

And you know, like more tangibly, like, I think, you know, one of my first investments at Greylock was in Roblox. And I think actually like spending four years in the weeds, building consumer networks, building consumer platforms, understanding what great looks like enables you when you look at a business like Roblox.

Which at that time was not an obvious, you know, massive success story and having the conviction to write one of Greylock's largest checks, first checks ever. We, we ended up running $50 million, which is large for Greylock, which is mostly an early stage fund. And so yeah, I, I think, it, it also leads to blind spots.

Like I, I do tend to be pretty focused on product and that's obviously only one input to building a successful business. And so that's something I've learned over time as an investor is to kind of balance out that bias. But at the end of the day, I think building delightful products is one important input into, into building a successful company.

Stephen: And it must be extremely hard. You know, you're doing pre-seed, seed series a investment. That's usually sometimes a founder and, and just their idea a lot of the times. How has your ethos evolved? You've been at Greylock since 2017. AI was still very much in its infa infancy stages. How has your ethos evolved over the last eight years?

Especially with the emergence of ai technology, especially in the last three years.

Seth: Yeah, I mean, you're right that like transformative companies are built by transformative people. Often there's some in like when you speak with a truly transformative founder, there's often some inevitability presence to what, to what they're doing. They have some unique insight and they have some insatiable kind of chip on their shoulder or perseverance to will that thing into existence.

And so I think that's one of the, like most common threads of founders that end up, kind of bending the world to their will. But almost equally important are the markets that they're going after. And so that's a role that we actually play when we're, you know, meeting with and partnering with people who are even pre idea, who are extremely talented and technical and they wanna start a company.

Because an expertise that you can develop as an investor is having a more horizontal view of the market and understanding what customers need and how technologies can be transformed into products that can be transformed into valuable and durable businesses. And so that's kind of a thought partnership role that we play often at the kind of early and incubation stage,

Stephen: You know, you know, taking up on what you said based on that surface layer of ai, we're seeing a lot of AI wrappers. How do you decide like, okay, this, you know, AI transcription sounded monumental a couple years ago. It's like not a technology, everyone, and you know, it's like a, a CD burner. I remember when I bought an external CD burner for my computer.

I had a really great job at burning CDs. But eventually all the computers had a CD burner and you know, they didn't need my services as much anymore. Where do you see, how do you break past the AI wrapper to think like, this is going to be, this is gonna have longevity, this is gonna be transformational.

Is there a certain kind of like line that you're like, Hey, this is more of a wrapper, you know, maybe a million dollars for the first year, but after that it's gonna get dicey? How do you kind of determine where you're gonna put your chips?

Seth: Yeah. So first of all, I think people think too much about this. I think, I think most amazing companies, not all, but many amazing companies, especially great SaaS companies, like don't necessarily have some inherent defensibility right at the beginning, right? You need to build something that people want that has real value that people want to pay for, and you need to do it better than others.

And you know, if you look at some transformative companies, even in our portfolio like ramp, right? And you think about RAMP at the seed stage. Like they were issuing spend cards for startups to start, right? Like, theoretically, anyone could partner with Marketta and Stripe and do that. Like were they a wrapper on top of Marketta, right?

At the end of the day, like there's a million different decisions and inputs that go into building a transformative company in their case. You know, hiring the best team, getting the best investors having the right vision around saving time and money behind a massive market. Build, you know, building delightful products quickly and you build moats over time.

And so, you know, my partner Jerry Chen, like published a post where there was, where he was basically saying the new moats of the old moats, meaning the, the physics of business have not changed, be just because AI exists. And so you don't need some magical answer, in my opinion, of IP depth beyond the models.

We have this new superpower in the world, which is intelligence, you know, at the cost of an API call, and we're still just scratching the surface of the business models and the products that that can be enabled through that. New infrastructure that's available. And so I think it's the wrong thing to overfocus on this.

I I, I published a blog post two years ago that was talking about how everyone was looking for the picks and shovels of ai, and I wanted to dig for gold, meaning go after the applications. And that was when it was a little more contrarian, when everyone was obsessed about these. Chat GBT wrappers.

Whereas now I think when you have cursor and you have Deagon and you have Harvey and you have companies in our portfolio like Resolve and Greenlight and seven AI and abnormal security, it, it's, it's clear that there's a lot of depth you can build. And we have certain frameworks around this too, like if you want to get nerdy about it, right around like what's the depth of workflow what are the different, the disparate data systems that you're connecting into, what's the you know, reinforcing feedback loop that you can build an intelligent system on top of, beyond the underlying models.

So there's a variety of different frameworks you can apply, but I think the summary is that people think about it too much.

Stephen: That's really interesting and you brought up RAMP and you talked about, you know, their talent density. But there is a rise now in media and podcasts around like historical founders. I know ramp. Is a sponsor of the podcast founders, which I enjoy 'cause they talk about historical figures and some recent founders and talking about the tried and true, you know, ways to found a business and to operate your business.

Are there certain best practices that we can take from the Steve Jobs of the world? Are we, to your point, are we trying to get, you know, too creative with how we approach the world with all these new tactics and strategies?

Seth: I think amazing companies and organizations are reflective of the superpowers of the founders. And so, you know, mark Zuckerberg is. A very like, systems level thinker, right? And so he thinks about networks and he thinks about systems, and Facebook is best in the world at that, right? He's enabled an organization that can be very data driven that you know, is growth like Facebook's superpower in the world is.

Building and maintaining networks. It's not being like the first to be innovative on some consumer product. It's not like the last level of design expertise. It's it's systems and networks. And for example, when I was at Facebook, we launched Messenger as a standalone app, and that was not because that was a delightful experience that every consumer wanted to have to download a separate app.

It was because it was a network level decision, which is the only thing that matters. For using a messaging app is, if I message you on this app, do I think you'll respond quickly? And two things need to be true. One is you need to have the app, and two is you need to have notifications turned on. And so with the system of, with the constraints of the operating system on mobile, it needed in order to compete with iMessage, it needed to be a standalone app.

In order for everyone to force everyone to have notifications turned on. And that's a network effect decision, right? So that's Facebook. Steve Jobs was obviously very different, right? It was very design oriented and customer oriented. Google is very engineering oriented. RAMP is very, an Amazon's famously very customer oriented, right?

And so I think there's no one way to win. I think it's just the founder needs to have a superpower and then that ends up being reflected in the organization.

Stephen: You know, I was saying before this session that you know my background is in compliance. Your recent investment in Greenlight, and I believe it's not your first investment in Greenlight.

Which is creating AI agents for financial crimes to sell to fintechs and crypto exchanges, et cetera. Can you talk about this recent development and how you approach, you know, a company, especially in an industry that's so highly regulated,

Seth: Yeah, so, you know, we backed we led the last two rounds in Greenlight, so we led the seed round and then we doubled down and led the series A. And this was founded by Will Lawrence and Alex Gin. We went through Y Combinator, we met them during Y Combinator and backed them. And it's, in my opinion, a perfect example of, you know, everyone talks about AgTech ai.

There's very few real use cases of it being deployed in enterprises and working at scale. And Greenlight is one of those use cases. And it comes down to going after a market that was. That had no incumbent software player, right? Which is L one, L one compliance analysts, right? Most of that spend is either in-house or outsourced labor, and then going after our mission critical workflow, right?

So when you sell to CIOs or any business leaders in an organization and you're just selling OPEX reduction, people don't always care. That's a thing that's very important as you're starting companies, right? So for example, like if you're a large enterprise and you have a 50 person accounting team and you have some accounting software that can bring that 50 person accounting team to 35, no one cares, right?

Even if it.

Stephen: it 'cause they're just used to dealing with the pain and they think it's too much pain to like get rid of those shifting people and implement the software and get their IT department on board. Is that what it is?

Seth: Exactly. People have limited time and attention in so organizations, and so if you're not the top one priority or the top three priorities, then you just don't matter. And like making the back office finance function slightly more efficient is just not the top one. Priority is not the top three priority.

So with compliance it's very different as you know, for, especially for financial services companies. First of all, it's a major expense, right? So when you're talking about companies like Citi and large banks, you're talking about hundred hundreds of millions of dollars of, of just labor spend, right?

That's one, but that's not it. Two is you're spending hundreds of millions of dollars to annoy your customers and still not be compliant. So people are spending hundreds of millions of dollars. There's $5 billion a year in, in financial crimes fines, right? Because there's a backlog even with, you know, tens of thousands, or let's say a large bank would have like 5,000 plus L one compliance analysts.

Even that is not enough to handle the alerts in real time and do it properly. And and you're knowing your customers, right? You're, you're losing revenue. It's a slow process to onboard. There's a bunch of false positives, you know, this isn't directly related to a Greenlight product today, but, you know, I'm a pretty good banking customer.

Pretty low risk person. I go to Canada all the time 'cause I am Canadian. Every single time I go to Canada and try to withdraw money, I can't. Right, because it's like a fraud alert, right? So like the system's totally broken in ways that really matter for a business. And so it's a very well-suited kind of category for ai because, ai, not O AI can just do, not only can it do it cheaper, but it can do it better. And so the other nice thing about Greenlight is they have an entry point in a very like high trust category. They have an entry point that can build trust and then expand. So they have a land and expand motion that works very well.

So they start with sanctions, alerts, handling which is usually pulling from public databases. Has very high volume. It's a huge pain point. It's very manual. It's a really, like, it's a thing that they can nail that's pretty easy to say yes to. And then once, once they build trust, then they can expand from there.

Stephen: And that makes a lot of sense. And to your point about, you know, compliance being pretty much a cost center, these companies are still spending millions of dollars outsourcing 'cause they don't have enough in-house. Talent. So they spend millions outsourcing. It's not only, you know, an annoyance to customers, it's annoyance to the other business departments.

You know, the, the, the personal relationship managers, the people that are trying to bring in customers and the revenue drivers of the business. It's even more annoying to them, and I think AI can be more consistent. If you have 12 different people from around the world all working in your compliance department, you're probably gonna get various.

Results when it comes to following suspicious activity reports and reporting, and, you know, the regulators are huge on consistency.

Seth: Exactly.

Stephen: thoughts about regulators actually accepting these reporting that are based on ai? Is it all about showing exactly what the AI did? Is that what Greenlight has to do for these financial institutions?

FinTech crypto is it show the thought process of how the AI came up to these results? Because I think that's probably the most challenging part, or the bottleneck is will regulators accept their results? If it's not 100% completed by a human.

Seth: Yeah, exactly. I mean, from the regulators we've spoken with, they're extremely open to you know, understanding technology and, and, and understanding its benefits. Right? It's, it's similar to a lot of different categories of ai, right? Autonomous vehicles where. People are always more complacent with the status quo.

And so in, in order to beat the status quo, you need to be so much better.

But so for example, like we're fine that 40,000 people die a year in car accidents, right? Because it's the status quo. You know, we're less comfortable if we save 20,000 people a year, right? Let's say, you know, but, but. If an autonomous vehicle even leads to a single death because it's something different, we're, we're more worried about it.

But I think in compliance, like I get, there's actually less resistance to evolving behind this, beyond the status quo. I think to your point, like people understand that the status quo is broken. Humans are not perfect and are not a hundred percent accurate, that the audibility and the consistency has actually improved significantly as well as the coverage and the and the latency, right?

And so there's so many benefits of using these systems.

And then you also have the built-in kind of graduation rails of, there's L ones and there's L twos, right? So Greenlight can still automate all of the manual investigation and. Report creation and still have a human in the loop kind of, you know, stamp it.

And so it's, it is actually very easy to start with low risk categories and then also to keep humans in the loop while still increasing throughput.

Stephen: I don't wanna stay on green light for too long. I know you have many companies in your portfolio, but I did see a post by Will Lawrence talking about it wasn't just the investment from Greylock, it was also the expertise that you had. It was the support and advice when they went to go to market strategy or recruitment.

If you're talking to founders and why they should choose Greylock versus any other VCs or investment firms. What are some of the things that you think you add? Because I, I've heard very few founders talk that way about their investors. Usually just say the name and thanks for the money. You went really deep into all the bonus and perks of dealing with your firm.

Seth: Yeah, thanks for saying that.

I mean, look at Greylock. Our strategy is to invest in the very best founders going after the most interesting market opportunities. But to do that, it means the very best founders in the world need to wanna work with us and our strategy. You know, while many venture funds are taking the kind of institutionalization approach where like, we're gonna build an investment bank, we're gonna have hundreds of people, we're gonna have hundreds of, of portfolio companies, that is okay for kind of, the assembly line of, of startups.

But I think if you are you know, if you have the choice of anyone to work with. We aspire to be the top choice of the best founders. And then the question is, how do you earn that? Right? And I think one is just being really deep in the domains and being deep in the individual businesses. And the only way to do that is to keep our portfolio very concentrated.

So we'll make only one to two, maybe three investments per partner per year. And so a founder has their entire career staked on a single company. We should at least have a good amount of our careers staked on their company so that you're in it, you know, from a structural perspective through the ups and the downs, and you have the bandwidth to be up to speed on what they need and how you can be an effective partner for them.

And then in terms of where we invest in our operations team, our, our recruiting team is larger than our investing team. So we find that founders want three things. One is a thought partner. Through the ups and downs of building a business. Two is talent and three is customers. Those are the only thing that, the only things that matter.

And so that's what we've invested in. So we have a talent team that helps bring in product edge design as well as executives. And then we have a customer development team that has warm relationships with Fortune 100, fortune 500 companies. And again, it, it, it, it goes back to our structure, which is.

You know, in, in order for us to be kind of a trusted curator for discerning buyers at enterprises, it would be impossible to do that if we had 500 companies that we were selling them. Versus, hey, like Greylock, you guys see every company in the market or a, a large majority of the companies in the market and you're only picking one or two.

So we're gonna take that as a signal, and then that has a reinforcing loop that's helpful for our companies.

Stephen: You know, it's, where are we in the market right now when it comes to VC and investments? I, I feel like there's two phases, just like the job market. There's either, it's so hot. All the, you know, the job candidates are picking and choosing where they want to go. Similar to founders when there's a lot of money in the system and then there's like, Hey, the job market's not that great, and the VC market's not that great.

And, you know, founders are trying to find the best deal. And you know, right now we're hearing a lot that the terms are not favorable, the founders. Where do you see the market as it exists today? Have we gotten rid of some of the froth over the last couple of years that seem to plague the market? What are your thoughts?

Seth: I mean, for hot AI companies and repeat founders, the market is as hot as ever. And I've been in venture for eight years now and that's always been the case. There's basically like for categories that are in favor. For companies that have outlier performance and for founders who, you know either are clearly outliers or potential outliers based on their background or expertise, or just some, you know, innate skill of, you know, building of, of attracting people or capital or, or, or building exceptional products.

Like for those people, the market is always really frothy. Now that doesn't mean that if like you happen to have a seed round that's not hot or a series A round that's not hot, that you can't be an outlier. Like that also happens very frequently. Like if you look at Chime as an example of a company that's about to go public, right?

Their seed and their series A and even their series B, we're not that competi. Right. Well, simple is a good example in our portfolio where a lot of their early stage rounds were not done by like, you know, the biggest brand name venture capital funds in the us. So the, I'm not, I'm not saying that like, the best companies are always hot at every round.

But I would say that the market is because of, again, what we do at Greylock, which is we're looking for. know, a very, very small set of companies that could be potentially category defining, multi a hundred billion dollar, you know, public companies, those rounds are often very competitive.

Stephen: Is there, you know, a certain characteristic for either a founder or a company where you don't think it's gonna make that mark, or even maybe companies that you have invested that haven't reached that full potential of the vision. Is there a certain thing that happened? Is it external market conditions?

Is it like the inability to adapt as a founder? Is it like, hey, they're just, you know, they frivolously spend, or they weren't reducing their expenses enough to have a longer runway? Is there certain areas where you're like, oh, we know exactly why this thing hit exactly where we thought it would.

Seth: I can't give you one specific thing. There's a thousand ways to, to fail. I think one thing that, that. Young product oriented founders often underestimate is the, is the quality of the market. And I think spending time up front, like we, we kind of, one playbook we have at Greylock is kind of go, go slow to go fast because this is kind of obvious, but if you're, I, I would say most amazing founders have a sense of urgency, right?

And so, a natural tendency for first time founders is to. Build with extremely high velocity in the totally wrong direction. And it's obvious, but if you're moving very quickly in the wrong direction, then you're actually slowing yourself down. And so, one thing we do with Greylock Edge, which is kind of our.

Quote unquote incubation program, but basically a structured program that we do to partner with highly technical repeat founders or potential great founders who are still exploring exactly what they want to do is we are thought partners with them in that negative one to zero phase, right? Not zero to one, but negative one to zero, which is defining exactly what you want to build and how attractive a market is because.

Look, I think there are amazing companies, and honestly, my favorite companies to back are, you know, a specific insight that comes from a founder that would not come from from like analyzing a market. But I would say like one of the most common failure modes I've seen for like exceptional founders is not focusing enough on the characteristics of the market upfront.

Stephen: And, you know, is there any, you know, there's this term that goes around a lot of the business podcasts, about like a one graph or a one page market where it's like one graph. Whether it's a hockey stick or show, some kind of insight or some regulation has changed and they're like, oh, this will completely blast off based on what's happening in the market.

Is there anything that you see out there right now where it's like, I could see why people are investing in it because of this like one graph or one pager that they've seen kind of being pushed around.

Seth: I like that. I, I always love those prompts of like, you know, the, the one chart. I can't think of anything off the top of my head. You know, I, I think it. I can think of several, like, for like very specific thesis areas, but, but I can't think of, of, of a single one. One thing that that is top of mind for me right now is, is the contrarian, the potential contrarian nature of what AI is gonna be good at.

And so I thought there was like a really interesting narrative in 2023 when chat g PT first came out. That, you know, the assumption was that AI was gonna replace quantitative work first, right? Like accounting and tax, et cetera. But the reality of like the first version of a AI, of these large models was that it actually replaced creative work.

First And the initial assumption with AI is that like writing poetry and writing books and writing music. Creating art was gonna be the last thing that gets automate with ai. But it turns out it was the first thing, which was really interesting because it was, the precision required was lower, right?

And I think the potential parallel narrative today is people think that it's the junior level workers that are gonna be automated first, but it's actually possible that it's the senior level. Work that gets automated just as quickly or in some cases more quickly because the junior level work requires more precision.

Whereas senior level work often requires just like inputting a bunch of context from a bunch of different sources, synthesizing it, and then making decisions that use reasoning but not necessarily requiring extreme precision.

Stephen: I think you make such an interesting point there. You know, I've been at several conferences over the last two months in crypto and a lot of the crypto exchanges, you know, and I won't name any, but they're. You know, it was based on that thesis. It's the senior managers are creating reports and the person's like, Hey, if I need a report and I'm waiting a week for a report, you know, I can press a button in AI and get that report within minutes.

And to your point, we're seeing a lot of those jobs being erased because those are the ones that are taking the inputs, analyzing large vats of data, and AI is just doing a, a better job at that overall. Do you see any moves? I think in the VC world, we've seen a move back to hardware. Which we haven't seen maybe in the last two or three decades.

You might know better than me, but we've seen a move back in the hardware around infrastructure to do with AI data centers, obviously the chips. Is that an area that you're at all interested in, or have you seen this and you're kind of just staying in your software as a service lane?

Seth: It is really interesting, right? When you think of how to build defensibility with AI companies now that, you know, creating software is now becoming fully democratized and the cost is, is approaching zero, you know, and you think through how much defensibility can be at the software layer and the at the application layer.

Versus solving problems that go deeper into the physical world or selling into areas that require more specialized relationships, like into government, for example. Right. I do find a lot of those things interesting. And then also just the pro proliferation of robotics and autonomous vehicles.

So I, I think. I do find a lot of that very intellectually interesting. I will say at Greylock, we're primarily right now focused on software. Right? There's still, I think, a massive amount of value to be built, both replacing many of the horizontal software categories like ITSM, the CRM right, which are investment in light field is going after that.

HRIS, even the ERP, I think all of these horizontal software categories. Are potentially up for grabs again. And then you think about all of these agentic verticals that are replacing like huge areas of labor spend. You know, that's a, a large area of expertise. And then I, I also think consumer networks are gonna we're gonna have another wave of amazing consumer companies over the next few years.

You know, I think. The first year and a half, two years of ai, consumer companies were very similar to the first couple years of mobile consumer companies, which are basically like the toys versus the businesses. So it's like, you know, the first camera filter app was like maybe a 99 cent app or a free app with a bunch of ads, and it had some cool filters right on your iPhone, but the business was obviously Instagram, which created the in the network.

And I think very similarly, like Mid Journey, right? Where it's a discord channel to like generate AI images, reach hundreds of millions of revenue. But I still think it's in the toy category, right? But I think, you know, now the models have gotten better and people have figured out the, the physics of these businesses where in the next couple years, I'd be surprised if there wasn't a new consumer network that emerges.

Stephen: You talked a little bit about creativity, and that's what AI's attacking first. Do you think we're at a tipping point yet where people are like, oh, that was created with ai, it looks cool, versus like, oh, another AI graphic of somebody you know, as a Barbie doll or in a rapper. Do you think we've reached a level, like when I see an AI comment on LinkedIn, I'm not warm and fuzzy about it.

'cause I feel like the person took a shortcut and I think still people are at that point. Have we gotten to the point where it's like, oh, everyone's using ai. You can't really tell the difference, or you don't care that it's a difference. You're like, well, if that person's smart, they should be using AI to write their comments versus themselves.

I feel like we're in a battle of like the creative art and the ease of automation with ai. Any thoughts around that, even maybe from your own personal usage of it?

Seth: Yeah, well you've probably seen that meme where is like a guy writes like a single line sentence email and then gets an AI to make a like a three paragraph email, and then the guy on the receiving end. Then uses AI to resummarize it back to one sentence. So it's just kind of like wasting time. But I, I mean, what I'm excited about is like moving beyond like sentence completion and moving towards, of course, like doing work.

Right. And I think that's where the real transformation comes in, and that's where we're spending a lot of time, right. Is kind of defining an AI agent as three things, right? Which is. Reasoning tool use and perception, right? So understanding an environment, understanding you know, how to plan and reason to execute work and then actually having access to the tools.

And I think there's still a huge wave of enabling infrastructure that needs to be built around identity security payments in order to truly enable a, like AI to, to execute work. Then and then once all that, and then there's kind of reinforcing loop as that infrastructure gets built, then more powerful applications can be built.

So I, I'm less excited by the like, image gen and like, you know, the sentence completion and like the, like email spamming bots. And I think there's, there's, I, there's much more sophisticated businesses that are being built now of of actually understanding an end-to-end business process and executing it.

Stephen: And have you looked at companies differently? 'cause I, I'm assuming 2017 company comes to you, still very young, doesn't have a large amount of funds coming to you, obviously to build up their teams and expand. Now you're like, Hey, this is a three person team. We could probably use a Gentech AI to cover a lot of the heavy workload in these certain areas.

Plus, you know, they're not, they're not gonna have to have a huge marketing spend. We could probably leverage a little bit of ai and the, obviously you probably have more technology expertise within your organization that can support these companies. Even if going into these companies didn't know how to use a Gentech AI to its fullest potential, has that changed the way you have approached companies, seeing how you can leverage AI to maybe build up these companies to get to that point that you're assuming they could get to quicker?

Seth: Yeah, I think every company needs to be AI native, including us at Greylock. Right. And I think if you're not, then you're just gonna lose. And I do reject the narrative a little bit that like people say, like companies are gonna need less capital to grow. I think like often capital needs. A function of like ROIC, like return on invested capital, like are, are there high ROI places to deploy capital.

And I think for many of these businesses, the ROIC is phenomenal right now just because there's a new superpower in the world that has an incredible amount of value. And so if the ROIC is is higher than it's ever been, then you're actually gonna require more capital. Not necessarily require more capital, but.

But have productive use for more capital. Now I do think it, it distorts the stage that's appropriate to invest, right? Because you can get more traction on less which, which re relates to ROIC, right? It just means that for any given dollar, they're able to get more a RR out of it, which means that I think there's more value in investing early because, you know, investing at higher prices at later stages should ideally come with a significant de-risking in certain categories. But the challenge is, like for an AI SaaS company to get revenue is not de-risking the key question, which is durability. And so, often all that happens at later rounds is that the price is higher, but the fundamental risk has not changed.

And so. At Greylock, we've always been very early stage, but that's something we're doubling down on even further, just given the market dynamics.

Stephen: You had an interesting conversation. The video, you know, was a live in person session where you talked a little bit about AI and FinTech and you know, maybe can you share some of these, you know, interesting why you feel that there's so much promise in AI and FinTech.

You included such things as like the ability to process unstructured data, which we talked a little bit about that, and even like making small improvements, having these huge financial rewards.

What are some of your key concepts for some of the people that are in payments and FinTech that are listening to this episode?

Seth: Yeah, so I think financial services is one of the most interesting categories for ai, right? It's 25% of the economy. You have huge amounts of unstructured data, right? So you have 10 Ks, 10 QS receipts, bank statements, right? All trapped in PDFs and text-based documents, right? Think about any loan underwriting, any, you know, business payment process, right?

And then you have a bunch of manual workflows wrapped around this unstructured data. And then you have an industry where the economic value of making even a slightly better decision based on a better understanding of data is enormous, right? Whether it's making a better investment decision, a better underwriting decision for insurance, a better underwriting decision for risk and fraud, better underwriting decision for for a loan, right?

And so you have this market that $16 trillion of aggregate market cap. Where there's a massive amount of unstructured data that's trapped in PDFs, and then you have this massive economic value for understanding that data a little bit better and making a little bit better decisions. And so I think it's just very, very well suited for ai.

One thing that's interesting is if you're an investor or an operator in FinTech, it can feel very crowded, right? You could feel like there's 20 different companies in every category. But what makes me really optimistic. Is if you as a consumer try to do anything right, like literally just try to do anything.

Try to get a mortgage, try to get a loan, try to get financial advice, try to do your taxes, try to send money internationally, try to do anything. It all is terrible, which means that there's a massive amount of opportunity. And the other thing that's really interesting, and, and by the way, that applies to both consumers and businesses, right?

Like, go look at your finance department, you know, ask them any question and, and of, of like the state of the business as of this morning. See if they can answer it, right? Like, it's kind of amazing how, how broken the entire system is. And what's also amazing is two of the most interesting technology waves that are happening today, I think are extremely well suited for financial services specifically, right?

So one is blockchain, right? Which is basically has the potential to, to re-architect the entire backend of finance and make it truly digital first and truly programmable, right? And the first mainstream use case we're seeing of this is payments. Right. But but we're still at the very, very early innings of just using stable coins for cross-border payments, not actually using smart contracts to execute real world contracts, et cetera.

So the backend is slowly being reinvented to be truly digital native, and then the front end has the potential to be fully automated with ai. And so I think the role of financial services in the economy should be an enabling layer, not a primary layer. So the fact that it's 25% of the economy is, is ridiculous in many ways.

And so I think fi like the best version of financial services enables the economy to grow much more quickly. So it ends up being a smaller piece of a much larger pie.

Stephen: I think that's stable conversations. We'd have to do a whole other podcast around stable coins and the regulations around it. I would love to know in the final few minutes what is on the roadmap for, you know, the company in the next, you know, eight months, the next few quarters. What are you looking at?

Is there anything that's peaked your interest from either a category level or you know, regulatory level? I know you mentioned, you know, crypto, blockchain. We're seeing a lot of momentum in the us. What are some of your thoughts going into the end of this year?

Seth: Yeah, so one thing I've learned over time, of being an investor is like the best. Things are really simple, right? And so the first question I ask myself is, where are the biggest companies that you can build in FinTech? And my answer is payments, consumer and software. Those are the three categories that matter, right?

So payments you have companies like Stripe and Ramp and visa and MasterCard. Consumer. You have JP Morgan, right? That's a big part of their business. But if you look at like the, you know, the new entrant, you look, you look at new bank, you look at Revolut, you look at Coinbase, right? You look at Wealthsimple these are tens of billions of dollars of market cap each, right?

And then software, right? And the story with software is AI agents to replace every part of services within financial services. So Greenlight is one example of that within compliance. Consumer, you know, a, a big trend is stable coins, right? So we're, we, we're investors in Vance, which is doing cross border payments for non-resident Indians, which is the biggest market in, in remits.

And that's just one example of, I think there we're gonna see a whole new wave of consumer, both with stable coins and when AI models become good enough to be consumer grade and then payments, and again, within payments, stable coins is a big theme.

Stephen: I would love to note, you know, is there any, maybe your own set of portfolio companies, is there anything that you play with on the weekends that you're maybe nerdier friends are playing with on the weekends that you're like, oh, this is probably gonna become mainstream. Whether it's a game an AI agent app, or anything that you're, that you're seeing that has a, is not really business oriented, but it's really fun and you can see it going mainstream.

Like, kind of like how slime for children and Roblox for children.

Seth: Yeah, it's a good question. I mean, it's probably not that unique, but I love using like versel to just like create apps and that's been like a mind blowing experience for me. You know, I was a PM so I, I've, you know, some. Some technical background, but I, I'm definitely not a software engineer. So being able, like over the weekends to over the weekends to just like create apps for myself, like I created a budgeting app for my family.

I created a just workflow app for, for my job. And it does make me think of like, anytime like creation is democratized there's often an opportunity for a marketplace or a network. So that is something I'm, I'm curious about. I think the other like kind of like, I don't know exactly like what the right business is is here, but like I've been really interested in this concept of like AI agents with real world networks behind them.

So if you imagine like a career coach agent that's also recruiter agent, and if I'm talking to my career coach and I'm saying like. Like imagine I'm a PM at Facebook and I'm like, how do I have this difficult conversation? Like how do I have more exposure to X, Y, and Z, whatever. So they learn about me and then occasionally they're just like, Hey Seth, by the way, do you want to meet you know, Reid Hoffman at Greylock?

'cause they're looking for a product person to join their investment team, right? That would be a magical experience. And so again, I think we're too much in the single player mode of, of AI right now and I'm excited to like, move into the network mode.

Stephen: You know, as someone that spends a lot of time on LinkedIn, I feel Boardy has covered a lot. You know, it's a. AI agent you can talk to on the phone, explain it. He has a network that he has of 20,000 people that are also calling him from, from

Seth: it's a good example. Good Canadian.

Stephen: directly, which is kind of interesting.

You know, as a founder, you're like, Hey, like I'm, you don't have time to look through your network of who has a connection here or there. You can just kind of pick up the phone, speak to him for five minutes, get a nice Ozzy accent responding to you. Where's the best place to find you? I found some great YouTube videos of you and others just sitting down and chopping it up with other founders, which I found very insightful.

Where's the best place for people to find you?

Seth: Yeah, thanks. I think Twitter or LinkedIn, I'm just Seth G. Rosenberg on Twitter and yeah, also the Greylock website. It's pretty easy to find me and get in touch. So yeah, if, if, if you're kind of a. If you, if you're an early stage or future early stage founder, building an AI in FinTech would obviously love to meet.

Stephen: I think you have some, you know, great portfolio companies. I even did a video for my audience on LinkedIn saying, I looked at your job portfolio board and it has some really interesting jobs in compliance and why not with a company that's, you know, a firm that's investing in, you know, what they feel are gonna be the next unicorns of the industry.

What a great place to find a curated group of jobs. So I, I did a video for our audience and thank you so much today, Seth, for joining us.

Seth: Yeah. Thanks Stephen. This has been fun.

‍