# Why All AI Companies Are Under-Valued

The Future of Foundation Models: Scaling Laws, Generalised vs Specialised, Commoditised? · From Unable to Afford Rent to Raising $130M From Index and Peter Thiel with George Sivulka @ Hebbia

20VC · Jan 22, 2025 · 66 min · 13,724 words
Speakers: George Sivulka, Harry Stebbings
Source: https://www.996.fm/episodes/20vc--ep-9af99979/

## Cold open

**George Sivulka** [0:00]:

You can you can bucket great founders into into three backgrounds. I think probably the most common is that you had kind of a messed up childhood. The second most common would be, you're gay, and the third most common would be you were adopted. Look at, like, a list of of all time greats. Elon Musk kind of messed up childhood, Jeff Bezos, Steve Jobs adopted, Peter Thiel, Sam Altman, you know, like, publicly gay. And and I think that all of these early life experiences end up giving you some desire, some deeper passion to go out and prove yourself.

**Harry Stebbings** [0:30]:

This is 20 VC

## Intro

**Harry Stebbings** [0:32]:

with me, Harry Stebbings. And today, we have one of the most wild stories in AI, Hebbia. Three years ago, George Sivulka couldn't make his rent of $300 a month for a mattress on the floor. He snuck into Stanford dining room halls for meals after dropping out. He raised his first two rounds of financing with clothes hanging behind him on Zoom. Most recently, the company raised a whopping $130,000,000 and has investors, including Peter Thiel, Index, GV, and Andreessen. This is one of the most remarkable stories of the last few years.

## Sponsor read

**Harry Stebbings** [1:08]:

But before we dive in today, here are two fun facts about our newest brand sponsor, Kajabi. First, their customers just crossed a collective $8,000,000,000 in total revenue. Wow. Second, Kajabi's users keep a 100% of their earnings, with the average Kajabi creator bringing in over $30,000 per year. In case you didn't know, Kajabi is the leading creator commerce platform with an all in one suite of tools, including websites, email marketing, digital products, payment processing, and analytics for as low as $69 per month. Whether you are looking to build a private community, write a paid newsletter, or launch a course, Kajabi is the only platform that will enable you to build and grow your online business without taking a cut of your revenue. 20 VC listeners can try Kajabi for free for thirty days by going to kajabi.com/20vc. That's kajabi.com, kajabi.com/20vc. And after building your online empire with Kajabi, time to scale your global team with remote seamless hiring solutions. So every business is a global business in 2025. But how do you do payroll for your global business and team and comply with international labor laws? Well, remote handles payroll, benefits, taxes, stock options, and compliance to help companies of all sizes pay and manage full time and contract workers all over the world. Remote's global employment solutions keep your team, your finances, and your intellectual property secure. Remote never charges hidden fees, just best in class global employment solutions for a low flat rate. The world's top remote companies love Remote. GitLab, the world's largest all remote organization, trusts Remote to run their global team. Remote is funded by Index Ventures, Sequoia Capital, and the host of the greatest podcast ever, Harry Stebbings in '20 VC. Ready to learn more? Head over to remote.com/20vc, that's 20vc, and begin hiring within minutes. Enjoy 10% off your first three months by using the promo code two zero v c at checkout. Now that your team is up and running worldwide, make sure your finances work just as hard with Brex, the ultimate financial stack for start ups. Let's be honest, building something from the ground up is hard enough without dealing with clunky outdated banks that pile on fees and leave your cash idle. Brex was designed with founders in mind to make every dollar go further so you can focus on building. And here's what really stands out to me. Brex combines the best of checking, treasury, and FDIC insurance in one powerhouse account. You can send and receive money globally at lightning speed, earn yield from day one, and still access your funds whenever you need. Plus, with 20 x the standard protection through program banks, your cash is not just working harder, it's working safer too. It's no surprise that one in three venture backed startups in The US, with companies like Anthropic, Coinbase, and Robinhood, I mean, my god, these companies are incredible. Trust Brex to help them grow. If you wanna join the smartest startups on the planet, head over to brex.com/startups and see what they can do for you. You have now arrived at your destination.

## Conversation

**Harry Stebbings** [4:15]:

George, I am so excited for this, dude. I've been really looking forward to this one. I spoke to Kevin Hart, Sanguine, Corey. I found out all the shit there is to know, so thank you for joining us.

**George Sivulka** [4:24]:

Yeah. And, I mean, it sounds like you did a lot of research, so thank you for diving deep, I'm really excited to meet you as well.

**Harry Stebbings** [4:30]:

As a venture capitalist, it's amazing the amount of free time you have. Talk to me this is gonna be a fun show. Talk to me about your childhood. I spoke to Sanguine, and he was like, this was a really interesting part of me getting to know George. So talk to me about your childhood, and I'm leaving that deliberately open for you.

**George Sivulka** [4:46]:

It is fair. And I think the first time I met with Sangine, who's one of our investors at the Series B, it was like a thirty minute lunch that turned into almost two hours of us talking in-depth about, the dynamics that that I think made me have a chip on my shoulder. But in short, I was born in Staten Island, New York City, which, you know, you already have a chip on your shoulder from that. Grew up kind of around New York City in in New Jersey primarily, which a second chip. But my mom, is probably like a mafia child, born and raised Staten Island, and my dad is an immigrant from Slovakia who grew up under the Iron Curtain and then immigrated, really escaped to The United States. They both of them actually fully intended to be professional athletes. They had four children, of which only one was a boy. And so, you know, you can imagine their dismay when I was, you know, chasing butterflies on the soccer pitch, or, like, falling on my head many times, which I I have plenty of stories of me literally, like, falling over while trying to dribble a basketball. And I think, you know, my whole childhood, was I was really just a math kid. Like, not very out there. I wasn't really talkative, only really good at math. And my parents barely even knew what Stanford was. So growing up, you know, you kind of have this whole misalignment of of who I was and who I wanted to be, with who they wanted me to be. That gave me this drive and desire and passion to go out and prove myself in a way that was really tangible. Maybe not only to them, but, like, hopefully to to my own kids one day. Did you have friends? I had a lot of friends who were incredibly nerdy. So we went to a public school. Like, I was the type of kid that would would hack the school tablets to put Starcraft on everyone's computer. And then, you know, we'd all, like, not be paying attention in public school playing star you know? So there was a there was a a large enough contingent of kids that were also not athletes. Before when we were chatting,

**Harry Stebbings** [6:34]:

you said there are three archetypes. I don't know if you're willing to go into it. Yeah. I'm I'm I'm But of successful founders that you found as a trend. Can you talk to me about the different profiles?

**George Sivulka** [6:44]:

I'm I'm I'm happy to. I I always joke around and say that you can you can bucket great founders into into three backgrounds. I think probably the most common is that you had kind of a messed up childhood. The second most common would be you're gay, and the third most common would be you were adopted. Look at, like, a list of of of the all time greats. Elon Musk kind of messed up childhood, Jeff Bezos, Steve Jobs adopted, Peter Thiel, Sam Altman, you know, like, publicly gay. And and I think that all of these early life experiences end up giving you some desire, some, like, deeper passion to go out and prove yourself.

**Harry Stebbings** [7:20]:

I I actually very much agree with you. I always very much felt like a disappointment. My brother was always no. Mean, it seems like my brother was always incredibly talented and good looking and tall and smart, and I was kind of just pretty average and I was fat. And my dad didn't really hang out with me. He hung out with my brother, and so I always just felt like a disappointment. A mistake What that was, papa.

**Unknown** [7:40]:

What what does does your brother do now? That's

**Harry Stebbings** [7:43]:

He works for me upstairs. That's that's what I'm talking about. There we go. Did you feel like a disappointment?

**Unknown** [7:48]:

I

**George Sivulka** [7:48]:

think the answer is yes. I did. I felt, like, actually physically unable to do the things that I was wanted to do, or I I thought that I was good at things that weren't valued or weren't as important. All of my sisters are amazing athletes. They're all, like, know, six feet tall, and they're, you know, just incredible athletes. And and I was I was just knocked. And so it's kind of like the ugly duckling in many ways. Yeah.

**Harry Stebbings** [8:11]:

I heard that you built lasers. You cold cooled NASA. Can you talk to me about these kind of very cool early influences in your life and how it shaped you?

**George Sivulka** [8:21]:

There's there's completely separate stories there, but I think How do you cold call NASA? This is actually a very good story. I wanted to be an astronaut. Like, that was my my number one goal, and I was I was hell bent on that. And so by the time I was, I think around 15 years old, I was going to high school in New York City, an all scholarship school where the alumni pay for everything. I was like I was kind of tracking academically really strong, and I wanted a NASA internship. And they were offering them to college undergrads or or graduate students. And so, obviously, I applied and got rejected five times. And then there was a snow day in February where my school was closed. I I commuted into the city. I actually showed up in front of their New York City office, the NASA Goddard Institute for space studies, and I I demanded that they let me in. And the front door security guard was like, you know, kid, get the heck out. Like, what are you doing? You don't have an appointment? Like Yeah. I was like, I printed my resume out on the nicest paper. I'm wearing a suit. You've gotta let me up. And he kicked me to the curb. And so I actually sat outside. It's like A Hundred 10th Street in Manhattan, and it was snowing. It was like so, so, so cold. And I, I didn't know what to do. I started crying. And I actually called my mother because I was like, well, I'm gonna come home. And she's a salesperson. She works in medical sales. She picked up the phone and said, listen. No. You're not going anywhere. You sit your ass down, and you call every single number that you can get into the building. And so I sat on the curb, and I cold called every number on Google, you know, from from my from my old phone. And finally, someone picked up. It was one of the only people in the office that day, and they came down, met me in the lobby, and I basically pitched them on myself for two hours. Gave me an interview. I interviewed botched the interview because I didn't know anything about, like, kind of, like, linear algebra. I didn't know anything about physics. But I I memorized all of the titles of the posters on this professor's wall. Came back the next day. So showed up again cold and told them basically everything that I could possibly know about his specific research, he was impressed enough to let me work for him for free. And then they paid me the next year, and then I published internationally recognized research the next year. And and by that time, I think that was impressive enough to to let Stanford let me in, which which was

**Harry Stebbings** [10:27]:

a which

**George Sivulka** [10:28]:

was a a life changing moment.

**Harry Stebbings** [10:29]:

That is incredible. Yeah. That is also an incredibly heart wrenching moment thinking of a a little boy on the street crying. The advice of when to give up versus when to persist and fucking relentless. Me and you are both young. We've been taught you win by persistence and going for it. When is that true, when is it not?

**George Sivulka** [10:49]:

I think I have an unhealthy obsession with driving really hard. Yeah. I think I you just can never give up. Like, I just don't think that's an option. You can look at every company ever and and some get to a $100,000,000 in in revenue in whatever, like, some span of time, which they probably their marketing team has hacked. And some, you know, end up taking really, really long periods of time. The only thing that actually changes is the rate at which you get there. And so sometimes things go in your favor, sometimes they don't. But if you're so persistent that you just continue, like, you can you can bring a lemonade stand to a $100,000,000 in ARR. Like, there's nothing that's actually stopped. You can brute force your way as a founder. You screw product mark. You can literally brute force anything in the world. You just have to have that ship. You have to continue to to just pound away at at whatever is is is in your way.

**Harry Stebbings** [11:33]:

Yeah. Stanford was a big one for you, I imagine. It was a really big personal validation to get in. Correct? Yes. Yes. How did it feel when you got in?

**George Sivulka** [11:41]:

I was on to the next one. You know? Was like I was like, okay. You know, that that's done. And, like, the next day, was like, okay. Well, how do I become the youngest PhD student in my school's history? It's like, yeah, it's not even a moment. I I think, you know, I was excited for a moment, but, you know, it faded very, very quickly. I spoke to Corey

**Harry Stebbings** [11:57]:

before the show. Someone who's known you since you were 18? Probably even earlier. Take me to the founding of Hebbia then. We're at Stanford. We're doing incredibly well. We are the wonder child. How does Hebbia come to be in that situation?

**George Sivulka** [12:10]:

I'm one of the youngest PhD students in the history of my school, and I actually believe that I was working on you know, at the time, one one of the areas of research that was most interesting to me was meta learning, this idea of teaching machines to learn to learn. And June 2020, Sam Altman, OpenAI came out with the GPT three. And if you remember the title of that paper, it was large language models are multitask or meta learners. And I'm I'm sitting in in my lab one day. I'm playing around with this new technology, And I'm like, wow. They just stole the most important thing I could work out right right under from my hands. And I said, well, if I can't build the most important technology, how can I build the most important product? And I think those are two very separate things. Obviously, at the time, GPT three was not a product. And I don't even think chat GPT is a really good product. It's like a calculator. It's got the technology in there encapsulated in very simple form, but it's not a product that, like, in Excel that lets you just build whatever you'd like with it. That's very human first. And Stanford always pounds into your head the idea, hey. You've gotta start a company where there's a lot of pain. And I had a lot of my students or a lot a lot of my friends would go into investment banking or private equity if they were really lucky, and, they would come back and and basically be the least happy versions of themselves. They've, like, lost 50 they just hated their lives. It seemed like there was more pain in financial services around processing unstructured data than anything I had ever seen. And I was like, well, there's a great company to be had here. Let's let's give it a shot.

**Harry Stebbings** [13:33]:

So we're sitting in that lab. We're like, hey. There's a great company to be had here. Let's give it a shot. What now? Because I heard I mean, and I saw pictures of this wonderful bedroom. Unable to make right. You made me feel like such a diva when I saw that bedroom. Unable to make $300 rent sneaking into Stanford dining halls for meals when you weren't studying there. Yep. George. I had no comment. Off the off the record. Raised two rounds of financing with clothes hanging behind him on Zoom. Take me to the next step post that I'm gonna do this in the lab.

**George Sivulka** [14:09]:

So I was on a PhD salary. You know, you're you're making, what, $38,000 a year. I think 42 if at the at the time, if you if you had the Stanford graduate fellowship, which I had. And I said I was gonna go on leave. And I actually originally went on leave and said told my adviser, I'll be back in a year. You know, this coronavirus, just, like, give me some time. And I didn't have anywhere to go. Like, there was, like, a logical next step, and I wanted to work on this company. So I asked my friends who were renting out a house in East Palo Alto to let me rent a room, the cheapest room they could possibly find. And they were all fully booked, and it was, like, I think, over $1,000 of rent. And they said I I think it was actually, you know, 500 or $600, not $300 to give my my broke self some credit here for not being able to afford the rent. But they said you could rent out the master bedroom closet. And so I bought, I brought in, like, a mattress from the dorms, and I had a folding table from Home Depot nearby. And I would basically rotate whether the mattress on was on the floor or the folding table was on the floor, and that was I just sat there and worked all day, sixteen, eighteen hours a day. Go to sleep, wake up, do it again. No weekends. Know, it's just kind of like I turned into almost a monk where I was just obsessively building heavy. I was training models at the time. So I'd wake up in the middle the night, check on them, you know, continue to use my GPU because I didn't wanna spend any money.

**Harry Stebbings** [15:21]:

Is there a period where more work is not effective? Like, when I think about sixteen to eighteen hour days in that environment, dude, I'm masochistic to the extreme where it's unhealthy and alcoholic, bulimic, tortured child. I mean, fuck. I'm like Lindsay Lohan adventure. But when I think of even, like, me in that, I would not function well there. I need fresh air, exercise. I just have to.

**George Sivulka** [15:46]:

Yeah. Ultimately, I probably went too hard. My hindsight's twenty twenty. I think I think I definitely left nothing on the table to a point where it was detrimental to my health. But at the same time, I think that that was, like, a crucible where it it helped form me. It's very hard to be a founder, and those are the moments where you're just, like, you're eating microwave meals every single day and, like, you're just losing weight and you're trying to will something into existence. I actually was trying to pitch one of my former bosses at a professional services firm, and he looked at me on the Zoom call and he almost cried. And he was like, just come work here. Like, what are you doing to yourself? Like, just don't like, come back. We'll give you, like, a proper salary. Like, you don't have to do this. There's so many low points like that. And and I think, you know, I just kept on on chewing through it and and, yeah, raising money in the closet. I I actually got on with for we raised a pre seed from from Peter Thiel and and and Floodgate and then our seed from Mike Volpe at Index. And Mike was like, hey. We're gonna do a partner call just to as a formality with a few partners just to close out. Is Volpe. This is Volpe. Yeah. And so Volpe is like, there'll be a few partners on this call. What round is this for? This is a seed. So it's a follow on to the pre seed in, like, November of, I think, 2020. And I get on the Zoom call, and I've I've literally got, like, clothes hanging behind me. And they're, like, all of a sudden, Mike shows up and then four other partners and then 80 partners and, like, the the Zoom screen tessellates with, you know, hundreds of faces, and I'm, like, horrified at the fact and Mike's like, look. He's living in a closet. And everyone's like, ah, great founder. And I was so embarrassed, and then and then I pitched my company. Honestly, hilarious.

**Harry Stebbings** [17:19]:

Yeah. I'd say I I just wanna unpack that element there because I said I I can't remember who it was who told me I had to ask, but they said I had to ask about driving to Peter Thiel's house. Well,

**George Sivulka** [17:28]:

so so I was saying two months prior, again, I think I was at Stanford or or just about to leave Stanford, and one of my friends had interned at Founders Fund. And he's like, hey. I hear you're raising financing. You should talk to Peter. And I was like, well, I'm not gonna say no to that. And so he introduces me on an email thread with me and Peter. And I'm like, hey, Peter. Would love to do a lunch or dinner. I'm not a morning person. So I was like, Peter would love to do a a lunch or dinner anytime soon. And Peter was like, I can do a breakfast. I was like, I really wanna do a lunch or dinner. You know? Like, can we do a brunch? You know? He's like, I'm gonna do a breakfast. And so I said, well, that's that's fine. And he gave me a slot on a Saturday. And so got in my car. That was an old, like, beat up 2006 Audi convertible that I had, like, fixed up from Craigslist and bought for $4,000 at three in the morning and drank a bunch of coffee, like 18 cups of coffee, like a five hour energy, like, all the disgusting stuff, and drove from three to eight to his house to go and pitch this guy. And he I think he showed up forty five minutes to an hour late. He's, like, just waking up, and I'm I'm wired. I'm, like, sitting in my chair, you know, like, ready to go. And it was supposed to be a thirty or forty five minute breakfast. I was like, it's kind of already shot. But we ended up talking for, I think, like, four or five hours about not only the company and all of the the the flaws that I had in my business model, but then also math and and and deep esoteric philosophy and, like, just the world. And he said, you know, I'm not investing at the time because it was coronavirus and a variety of other factors, but I'd love to put in a check. I got out of this conversation thinking I had just made a friend or was seen from someone who's incredibly, incredibly, incredibly bright. And, I'm I'm leaving his house, yeah, I felt like I was inducted into the Illuminati. I was like, whole know, drop top, the sun's shining. I was playing Kanye West. I drove out as as as, like, my first offer from a venture investor. How much did he invest? I think the the total round was, a million dollars. So it's like it's like nothing. Yeah. But What do you think makes Peter so incredible? There's two things. He is incredibly ontologically smart. And so he he can build this worldview or this perspective of the of the world where, like, he actually just, like, knows how to pattern match to a variety of other things. But then he's also, I think, phenomenologically smart, which is the idea of, like, he understands processes and how humans behave really well. And so he's always thinking like, hey. You know, ex ante or or, like, you know, if if I'm looking at something that's about to unfold, could I have predicted this ahead of time? And he always just asked himself that question. So he's built up a really rich perspective of of the the fallacies that human society has, memetic behavior that that people kind of go out and and and copy each other with, etcetera and etcetera.

**Harry Stebbings** [20:03]:

So we then have money from Peter, and we have the Peter Thiel stamp of approval. Yeah. Does that open every door in the valley?

**George Sivulka** [20:08]:

I I mean, I think we were in late discussions with a lot of investors, and then everyone else was like, yeah. Let's, like, you know, let's pile on in here. It's, you know, it's some of the best money that you can get, and that was that was a game changer for us.

**Harry Stebbings** [20:19]:

So then we closed with Maples and Floodgate. Yeah. It

**George Sivulka** [20:21]:

was Anne Anne at Floodgate. Okay.

**Harry Stebbings** [20:23]:

Anne and Floodgate and and Thiel. And then we get back to work. Mhmm. And we've got a million or so.

**George Sivulka** [20:28]:

We've got a million. And then two months later, Mike actually hears about Hebbia from his daughter who was a Stanford student. And I think I've seen the product, was friends, and and then Mike actually comes in and is like, hey. This is completely different than Elastic or all these other search technologies that I've seen and invested. Obviously, he's on the board of Elastic today, so he's like, well, let's just go add some fuel to the fire. How much did he invest then? He, I think, invested, like, an additional 2 or $2,500,000 at the time. So where did the 130 come from? Well, as as years later. You know? This was years later? So this is all in 2020. We we end up building a product studio, which is the first to productionize RAG, retrieval augmented generation, also in 2020. We build the first semantic search engine. We actually go out and and search Can you just help us understand what is RAG? RAG is an acronym that stands for retrieval augmented generation. If you look at large language models today, they're really good at at maybe thinking if you give it the right context, but they hardly ever have the right context. And so RAG was the first real attempt to give them the data to answer questions correctly. Okay. And so we are building on RAG to start? We were one of the first people to productionize the idea of putting a search engine behind an LLM. So you'd ask a question, and then instead of it just replying from its memory, it would actually go and do a search and then reply with that context. So in an enterprise where you have a lot of offline data, we were really the first people to hook up that offline data to large language models to answer that question.

**Harry Stebbings** [22:01]:

So you're one of the first. And we're we're seeing that now in action, and it's working?

**George Sivulka** [22:05]:

That's a bit of a plot twist over here. We I actually don't think RAG works at all. It's one of the most used AI architectures in the world, pioneered at at Hebbia in a very meaningful way. I think every enterprise is experimenting with it, but it has a lot of different failures where a lot of the time the questions that people ask these systems aren't ever explicitly in the data. They're never explicitly stated. They're actually about the data. So for example, you know, if you're asking an AI system, is this company a good investment? Which is actually a very common thing that people ask Hebbia over marketing materials. Maybe it'll say in a pitch deck, yeah, this company is a great investment as a something that the CEO says or like a recording, etcetera. But what you actually want from that system isn't to search in the data, It's an answer about the data. Hey. What's the customer concentration? What's the strength of the management team? What are x, y, or z criteria that are are fundamental to our specific investing process? And that's a process. That's not that's not ever explicitly stated. Actually, the marketing materials are often like a load of crap. You have to actually distill what's true out of them. That's what Hebbia does. So it's not actually finding something that exists already. It's taking all the things that exist already and starting to answer questions about that information.

**Harry Stebbings** [23:16]:

Take me to that transition then because we were building on RAG, and we're like, great. We're gonna productionize this, and then we move off. Yeah. And you realize that actually it's bullshit, and it's not as good. Yes. Take me to that realization.

**George Sivulka** [23:28]:

Yeah. So we actually deploy at some of the largest finance firms in the world. We we ended up going from 0 to a million dollars of revenue sometime in 2021 or 2022 and raise our Series A also from from Index, from ICON Index, which is $30,000,000. And we start to see that all these customers okay. Now they know what, you know, ChatGPT is. They they know what LLMs are. Hebbia has this, like, really mature enterprise product in the market, and we're by far the first to actually get there. And we just looked at all the queries that people were asking. And the questions that people were asking weren't ever, okay. Find me the quote or find me the, you know, command f questions. They were actually more, okay. Read all the documents and then tell me all the times they mention AI or what our exposure to Silicon Valley Bank is during the regional banking crisis. And so all of these questions it was actually almost 90% of the questions that people were asking these systems weren't answerable by a search through the documents, but rather they had to be work done on top of the documents and then answered.

**Harry Stebbings** [24:24]:

My question is, like, what the fuck happens to the rest of the landscape if you're like, no. RAG is not actually the right approach, and they're all loving RAG. Mhmm. Couldn't be hotter right now.

**George Sivulka** [24:34]:

I I don't think they're loving RAG.

**Harry Stebbings** [24:36]:

You don't think they are?

**George Sivulka** [24:37]:

I don't think they're loving RAG. No. What makes you say that? I think, like, 90% of enterprise AI right now is almost like this vaporware. We swear it works. Like, look at this amazing demo where we ask, what does the CEO say about the investment? And the minute that they actually go to, you know, try to use it in a real world example completely just fails. And so I actually think that the majority of AI usage, a lot of these usage statistics are all kind of, one of my favorite phrases is. And one of things that Hebbia really tries to put forth in the market is is to say, hey. Change will take time, but we have a system that that is actually starting to drive real measurable value over very specific defined use cases. And our tagline's always, hey. Stop experimenting with AI, which everyone's experimenting. They're all really excited about it. Start driving value, like, getting value out of it.

**Harry Stebbings** [25:26]:

To what extent is this RPA versus a janty?

**George Sivulka** [25:29]:

I actually am not a big believer in RPA. I think RPA is a is almost not an AI application in the in the new sense of AI. It's like AI in the old ten years ago sense of AI, where RPA is effectively, like, very simple computation. But some of the things that people are asking Hebbia are over 800 page credit agreements or 230 page, you know, SIMS, confidential information memorandums, this marketing material. They're not actually asking for things that like copying numbers. They're saying, hey. Tell me what are inconsistencies in this document. Tell me where there's an event of default that we can trigger. There's almost this like open endedness or this new

**Harry Stebbings** [26:08]:

level of computation that people can do. I think Daniel Dines, who we had on the show, comes out on Wednesday, he said it very well. And he said like, listen, RPA is low skilled, low level cognitive processes and agencies, high skilled Yep. Ambiguous Great. Decisions. Yes. I think that's a nice phrasing. I think it's it's incredibly clear. Yeah.

**George Sivulka** [26:26]:

And I think that we very much are capturing the the agent, the the high level ambiguous decision making and trying to trace it all the way down back to individual citations or individual characters that led the model to that decision.

**Harry Stebbings** [26:39]:

I thought about actually Satya's statement the other day that he made, which is the notion that business apps that exist today will all just collapse into agents. Do you agree with that? And will apps be the predecessor to agents?

**George Sivulka** [26:51]:

I actually don't don't agree with that at all. I think he's completely wrong. I think it depends on how you define business apps, but I actually think that if the new business apps are platforms, you'll you'll actually start to see those platforms really take hold. Like, Hebbia is a platform. It lets you build whatever agent that you'd like. And so here's a bit of a mind fuck. When when building Hebbia or when building all of these foundational primitives for how people use AI over the last four and a half years, Hebbia has always asked ourselves, what are the apps that AGI would want to use? Or what are the apps that agents would want to use themselves? I e, what are the tools? Because these AI applications are really good at using tools that we could build that would assist LLMs or these really smart foundation models, whatever they are in the future, to get to an answer more quickly. It's quite interesting. You know, Hebbia Matrix orchestrates lots of smaller LLM calls. It's actually scaling at inference, I. E. It's running massive amounts of compute at the orchestration layer. And we think that, you know, an AGI system would prefer to use Hebbia Matrix to diligence a company or to look through thousands of documents versus to read them all by hand and they're in a really long context window.

**Harry Stebbings** [27:58]:

So is the future of business apps not business apps, but business platforms? We've got platforms, agents, or apps. Yeah. What is the future?

**George Sivulka** [28:08]:

I ultimately think that it will be a mix of of all three. History doesn't repeat itself, but it often rhymes. You know, sixty years ago or even longer, the foundational unit of compute, I. E. Doing a calculation on a computer, was effectively introduced to to the enterprise. There were plenty of people that were tallying things or bookkeeping in actual books, and their jobs changed. And there were apps for bookkeeping, and then there were platforms like Excel that let people build better apps for bookkeeping, and then Excel was unraveled again into better better apps for bookkeeping. And I think that, there's opportunity not only in verticals, but there's also opportunity in the entire industry in in in terms of building platforms, in terms of building cooperatives, in terms of even building new types of, quote, unquote, agent employees. And I think that that opportunity is the exact same size. It's if if there was a 100,000,000,000,000 of dollars of value that was created in the stock market from the introduction of the computer or the fundamental unit of compute, I actually think a $100,000,000,000,000 of value will be created, in the next sixty years from the the introduction of inference or or of AI

**Harry Stebbings** [29:09]:

compute, which is Will will that be additional value, or will that be value that denigrates from the existing value of alternatives? I believe it will be additional value.

**George Sivulka** [29:17]:

And maybe I'm too techno optimist, but I actually think the S and P 500 is is completely

**Harry Stebbings** [29:21]:

$100,000,000,000,000 of additional. I can't quite get my head around that. How does that even exist? Adding in that $100,000,000,000,000 of value, is that just because we will see GDP and productivity grow so much that takes place?

**George Sivulka** [29:32]:

I I ultimately genuinely believe that more than 50% of the GDP will be contributed by what you can call agenetic applications in the next few decades. Yeah. I actually think it'll happen faster than the next few decades. Do you? Yeah. Because, again, Daniel.

**Harry Stebbings** [29:48]:

Daniel Hot on take. Yeah. Daniel on the show was like, listen. Put me up against Daniel. Yeah. What? He's my neighbor, so he's literally next door. So we can arrange that. But he said that we consistently underestimate how long it takes for enterprises to adopt new technologies, get comfortable with data security, to get comfortable with processes. Is that right, or do you think actually we are past the tipping point?

**George Sivulka** [30:11]:

It's a good point if you're cutting cost, which I think Daniel and UiPath are are one of the best examples of of using AI to to make companies more efficient. Like, if you look at finance and how fast Excel went to 90% market penetration in finance, 1985 to 1986, literally eighteen months to twenty four months, Excel took over all of finance. Everyone switched from using a calculator, the HP 12 c, to to using Excel. And if you look at how fast, finance actually ended up, you know, using credit card data to value public companies ahead of their earnings, That was, again, a two year period more recently. And so finance is the slowest moving, most lethargic, you know, leviath. It's the worst possible customer base to go after unless you're providing outsized alpha or real value. In which case, the minute that there's something real, finance moves faster than any other industry. And so I'm actually making a bit of a bet by going into finance and and starting to take a go out and try to get to my own. Do you know what I worry about? I worry that we lose the

**Harry Stebbings** [31:08]:

education process. What I mean by that is, like, a lot of, like, GPs or managing partners or you name the title in it, senior firm, they've been through the shit of analyzing companies, staying late, understanding what makes a great business, all of these things. Yeah. And then we'll just say, well, don't worry about that shit. Happily will do it. And so we have this no graduation pathway for the next generation, and so we have decision makers who don't have that graduation.

**George Sivulka** [31:33]:

Yeah. I'm less worried about that. I think, ultimately, one of the best things about being, you know, having the years of experience is is actually having the depth of knowledge about investing. So for example, if I'm a junior trying to price an asset, I haven't seen that many other companies that look like this company. And so I might say, hey. I think it should be priced at x, y, or z. And then someone in the IC IC meeting will be like, hey. No. I've seen 20 other companies in my forty years of career that look exactly like this, and they all went nowhere. And I'm I'm leaning on my prior experience. With Hebbia, now juniors, who are really smart themselves, can say, okay. You might have remembered 20 deals, but I'm looking through every deal in our company's history in a giant matrix that anyone has ever seen. And and I actually can tell you quantifiably that when a company is performing here, it's it's ninetieth percentile across all this investing criteria. We should actually pay 90% premium to market. And I'm actually using more deals than you've ever seen because I know your name is on this many IC memos. And that type of structured thinking or that type of of additional information that you can now give juniors in their career actually think makes better investors. I I I don't I don't think that that takes away Do you

**Harry Stebbings** [32:44]:

really think it will be a tool for usage, not a tool for replacement?

**George Sivulka** [32:48]:

I genuinely believe it makes humans better. I genuinely believe

**Harry Stebbings** [32:53]:

years' time, do you not think that is a different story?

**George Sivulka** [32:55]:

I think that it will change the way that people do work, but I genuinely believe that it'll actually increase the AUM of the firms that use it. I think it'll actually drive more employment. I I think there will be some jobs that change. Hey. There's no more bookkeepers that do, you know, tabulations in in spreadsheets on two sheets of paper. But What changes most and what stays the same? The cognitive tasks that are lower in cognition, like more back office, kind of middle office, maybe even some of the more junior front office tasks, I think will start to be start to move into, okay. How can we manage AI juniors rather than, you know, actually do this ourselves by hand? But I don't think that, you know, just as Excel didn't end up taking away jobs from people or just change people, you know, to having to learn Excel, the exact same thing will happen with AI.

**Harry Stebbings** [33:45]:

So you don't think that we will see team sizes reduce as a result of agent integration into enterprise?

**George Sivulka** [33:51]:

You know, there's there's all these stories and, you know, you have, like, that's positioning for investors that they're firing half their staff, and and and no one really wants and I think a lot BS. I think it's BS. Yeah. There might be some reality to it, but I think that it's it's an amazing marketing story. And so anytime that I ever hear something that's put out as a marketing story, I almost negate it in my head to actually think about, like, what the implications are. When you're saying something and screaming it from the rooftops, that almost always means that inter internally you're freaking out about something. I think I look at that really loud behavior, and I think that behavior itself really negates the content. That's maybe my my positioning on on this sort of stuff. How do you feel about competition? What are your lessons

**Harry Stebbings** [34:29]:

on competition? There are several players now in the Hebbia slipstream. How do you feel about that?

**George Sivulka** [34:35]:

I think that if a $100,000,000,000,000 of economic value will be created by AI and agentic applications, that there will be so much room and so much opportunity for a ton of different players. I don't think that when Excel came out and then Marc released Salesforce and then, you know, people created TurboTax, you know, all these unravelings of Excel actually were were produced later, but that made Excel any less valuable. I actually think it made Excel more valuable. I view Hebbia as this platform as something that will actually get better the more people get inspired by it and build increasingly verticalized applications. What models do you use? You sit on top of what? We are completely model agnostic. We use all of the major model providers, some of our own models. But ultimately, the foundational difference that Hebbia is capitalizing on right now is fundamentally new and and very important difference, which is I actually think on the order of creating RAG and and creating agents and decomposition, is this idea of us in the last year or so having pioneered scaling at inference. Talk to me about this. Right now, you actually so OpenAI is starting to do this with o one, where they'll have a model recursively think about a question over and over and over again before it produces an answer. And so instead of training a larger model, they're using effectively a similarly sized model and just telling it to run multiple cycles, I. Compute more before answering. Hebbia is actually pioneered something different where a year or almost eighteen months ago, we said, hey. We can't wait for these models to catch up. What we'll do is infer simple single question. Let's actually run hundreds or even thousands of sub models of the best models in the world to compute over every single document to answer the same question. And so ultimately, if you can't train larger and larger models fast enough, you could take whatever state of the art or cutting edge and run it more times to get more compute, I. More computational power, better decision making for the same user right now. And so this is an idea that we pioneered. It doesn't matter if you're using Claude three five or if you're using o one itself, I. Scaling at inference, at the orchestration layer with something that is scaled at inference with the training layer, but you get way better results. And it's, it's a it's a way to to drive to more accuracy.

**Harry Stebbings** [36:47]:

Find provides the best results. We had Des Trainer on from Intercom recently. He spoke about the movement away from OpenAI to Anthropic.

**George Sivulka** [36:53]:

We've seen that for certain types of documents, like the dense legalese or or or more colloquial documents, Anthropic works better. But for other types of documents, like o one or OpenAI four o works better. And it's always trade offs between accuracy and speed and all kinds of things. Actually, a lot of the time when we're decomposing a task, we'll use mixes of OpenAI, Anthropic, even Gemini.

**Harry Stebbings** [37:15]:

Do you think we live in a world moving forwards of many models that are specialized in different things? As you said, some do legal, some do whatever we wanna talk about, and that's the world we live in, or there's generalist monolith models which really kind of own the whole stack.

**George Sivulka** [37:30]:

This makes me think of the story of Bloomberg, which has the best financial services training set of all time. And they trained a GPT 3.5 class model. It was called the Bloomberg GPT. And they released an archive paper, and everyone on LinkedIn was like, wow. Bloomberg is, you know, cutting edge, and they're gonna steal finance AI. Why did they not? They did have all they've got the best data in finance. Well, so then GPT four was released, I think, like, a few weeks later. I don't know exactly know the right timeline, but it just destroyed Bloomberg GPT at every single finance task. And so you saw the idea of post training or or or or kind of like this refined verticalized model creation just always would lose to scaling laws. And maybe we're at the end of scaling laws at training, but I actually think, you know, Hebbia and and now OpenAI and a variety of other companies are starting to pioneer the idea of scaling laws and inference. And I actually think that it will like, nothing that that other players can do to fine tune models will ever catch up.

**Harry Stebbings** [38:25]:

I

**George Sivulka** [38:25]:

need

**Harry Stebbings** [38:26]:

to break that down. Sorry. So everyone's like, oh, are we at the end of scaling laws? Aren't we? Yeah. You know, read Benioff and Daniel Dines are like, yes, we are. Yes. We are at the upper end of LLM. Yep. Reid Hoffman is like, no. There's so much more room to run. Can you just break down for me the difference between scaling laws at inference and scaling laws at training?

**George Sivulka** [38:50]:

Yeah. I think, you know, there's it's a bit of a marketing distinction. Right? But, ultimately, the idea is that the way that we got here over the last five, seven years of of training models has been let's build a bigger and bigger model, and let's give it more and more data, more and more clean data. And then maybe we'll do some RLHF or some, you know, reinforcement leaning to reinforcement training to to fine tune it after pretraining. And that worked great to get us here, but we're running up against the amount of good data that exists in the world. We're running up against Are

**Harry Stebbings** [39:20]:

we? Because people push back on this and say there's so much data that we haven't used yet, whether it's video data that can't be translated, whether it's synthetic data. Like, we are not at all exhausted in terms of data supply.

**George Sivulka** [39:32]:

You know, I think that we're starting to run up against the constraints of it. That's a that's a gut feel. I'm not, you know, I'm not looking at particularly in data collection myself, but I I I think we're starting to run up against the limits of of really good data that we can. What's then the problem? So so, ultimately, that might mean that, hey. We're training larger and larger models. XAI, again, just created the largest GPU cluster of all time, and they're gonna try to train larger and larger models. But regardless of how the scaling laws for training larger models or parameter count and accuracy or performance carry out, I'm starting to believe that you could still get better compute, not by building a larger engine, to use a metaphor, but by actually putting a bunch of smaller engines together. The Hebbia, by by orchestrating large amounts of inference to answer one single question, ends up kind of building like a Tesla, where Tesla is made up a bunch of smaller engines or a bunch of smaller electromechanical motors, that make a lot of torque and a really a really amazing larger engine. Does it not make it incredibly capital inefficient? You know, I think the one thing that people in my position, will always tell you is that will go to zero. The cost of intelligence will go to I mean, I think that since Hebbia started, the cost of of of inference over a fixed number of parameters has decreased in by, like, seven orders of magnitude in four years. And so I genuinely believe that scaling compute is like a no brainer. And, yes, we run more large language model calls than anyone might even say would ever be necessary, but we have the best accuracy in the business. We can answer much more complex problems. We're driving real value for enterprises. And I I actually think that every single quarter, like, our margin goes, oh, we're not spending money fast enough.

**Harry Stebbings** [41:14]:

You mentioned x AI's GPU cluster. Yes. What they've been able to do in such a short amount of time is miraculous. Yeah. What do you think that tells us about the layer itself?

**George Sivulka** [41:24]:

Ultimately, the model layer, and I think this is not a hot take anymore. I've been saying it for a few years, but I think it'll become commoditized. I think that a lot of value will accrue at the hardware layer. We we could talk about what that means for NVIDIA, especially as NVIDIA has a stranglehold on training, but not as much stranglehold on inference. And so you might actually see other chipmakers, you know, actually start to their their chips start to be used in a more meaningful way because CUDA is what all ML scientists were trained on in their PhDs, but then, you know, inference doesn't matter, kind of what you're using. And I think it'll be the infrastructure layer and then actually the application or agent layer that will accrue the most value.

**Harry Stebbings** [41:59]:

Ultimately Why does it not follow the same vein as cloud where cloud is commoditized? Yep. But Azure, Google Cloud, AWS, I mean, completely commoditized, to be honest, cloud. But it's great business for them. I think

**George Sivulka** [42:13]:

it might. There's probably fewer players and more entrenched players in cloud. And, ultimately, you you know, I think those those players honestly kind of have, like, an an OPEC oligopoly where that, you know, they can control pricing. I just think that, ultimately, cloud is actually more complex than training larger and larger models to And

**Harry Stebbings** [42:30]:

the cloud providers are basically using models as a loss leader happily to build stronger moats in their cloud businesses. And you see this with Anthropic and Amazon. You see this with Microsoft and OpenAI.

**George Sivulka** [42:42]:

Absolutely. Whoever has the best models will continue to attract the right amount of investment. The the different thing about clouds too, though, is that the cost of switching is is much higher. So to so to refine my earlier point, right, like, I can switch models readily. Like, I think there's even entire businesses now. There will be an entire industry of being able to switch models from OpenAI to Anthropic when OpenAI goes down. But to switch clouds is like, you know, for any, like, substantially sized startup, like a 10,000,000 to $20,000,000 investment just to switch. It's almost always never worth it. It's much, much, much sickier. Whereas here, it's a very simple API key. You know? It's it's very simple to switch models. And so I think that that's also a differentiator.

**Harry Stebbings** [43:25]:

OpenAI at one sixty, Anthropic at 40 Yep. Or xAI at 50. Which one do you buy?

**George Sivulka** [43:30]:

I think xAI is the most undervalued company and a really spicy take. I actually think xAI might overtake OpenAI and Anthropic into value over the next, you know, twelve to twenty four months, which is crazy, but I think they're all undervalued. Your thinking? I think Elon is is very well positioned in the geopolitical sense. I think Elon can run a more efficient business than and not have to deal with as much administrative bloat or as much friction from from employees. How important is geopolitics

**Harry Stebbings** [43:58]:

in winning this game?

**George Sivulka** [44:00]:

I think geopolitics is is actually very important. I think that governments will be some of the largest users of AI, especially with, like, some of the the recent things that the the the new administration in The United States has been talking about with with increasing government efficiency. I think that ultimately energy is a a very big bottleneck. You know, it's a very common thing in Silicon Valley to talk about, hey. We need nuclear reactors to to flatten the duck curve so that, you know, we can, you you know, we can continue to drive to, larger and larger data centers, etcetera, etcetera, etcetera. And and those are ultimately geopolitical resources. And so I think all of these things end up being very important. Then and then Elon's just operationally so talented. Right? So I think that ultimately, if this becomes commoditized and whoever can really operationalize, model creation and and and serving models the fastest, I think what I think might start to

**Harry Stebbings** [44:46]:

So you think xAI and you would invest in them?

**George Sivulka** [44:48]:

I would, but ultimately, I think all of them are undervalued. I genuinely believe all AI companies and the S and P 500 are all undervalued, which is a very hot take. If we're about to create a $100,000,000,000,000 of value, think this is a real tangible technological shift. It's a massive unlock on the order of what computing did for the entire economy over the last sixty to eighty years. I think this will do for the next sixty to eighty years. I think all these companies are massively undervalued, including the non AI companies. Unpack the last bit, including the non AI companies. I genuinely believe that computers made legacy businesses better if you use them correctly. And so it's a massive disrupting force. But if you can ride the wave of change, AI agents in this new fundamental paradigm, that is a massive unlock for human life. Do think there

**Harry Stebbings** [45:34]:

is a slight difference? Everyone talks about kind of different technological transitions. When you look at, you know, the agricultural transition or the kind of agricultural dependency on human labor and movement to machinery, computers and workforces, these were at least ten year transitionary periods. Yep. At least. This is like, hey, we use AI tools now because we just bought them today. Yep. The transition period instant. Yes. Much

**George Sivulka** [45:57]:

faster. Mhmm.

**Harry Stebbings** [45:58]:

Does that not change the enterprise value accumulation and whether they're good or bad for businesses? Because it's like instantly your business will die if you don't have it or not.

**George Sivulka** [46:07]:

Yeah. I actually always liken technological revolutions, you know, to to what Hebbia is doing right now where, you know, people invented or we discovered the technology of fire, and then someone invented the torch, you know, I don't know how many years later. Know, we invented the engine, then someone invented the car or the wheel and then the chariot. And so this idea of encapsulating and building a useful product on top of a technology change is actually the thing that takes more time. And I think that Hebbia has built if Excel was that product for compute, I actually think Hebbia has built that product for AI. And I think that when you have a good product, that transition will be very, very, very quick. Right now, we have these chatbots or these, you know, surface level search engines that that give you facetious, you know, surface level value. Yeah. It'll help your kids cheat on their homework, but to drive to whether or not something's a good investment is a much, much more rich problem. Is chat the right interface for

**Harry Stebbings** [46:56]:

many of these applications?

**George Sivulka** [46:57]:

I I ultimately do not think so. I think I think that chat was always a useful feature. It's a it's an it's a useful interface. It's like a single cell in Excel. It's like asking if the t I 84 was the right interface for computers or the terminal was the right interface for computers. We have not even started to explore the opportunities for interfaces. I actually think that you think they are? You know, I think that Hebbia is is the Bell Labs, and I conceive of ourselves as the Bell Labs of defining AI interfaces. Right? I I think that, you know, RAG was one of them. I this idea you could find things in the data really fast. Decomposition in agents are another. This idea of scaling an inference with our matrix product is another. You can look at a lot of the other things where agents are controlling four screens at once, and you're actually looking at someone use a computer or computer use where AI models are moving cursors are others. Almost all of them have

**Harry Stebbings** [47:46]:

Ultimately, if agents are efficient, does interface not become irrelevant?

**George Sivulka** [47:50]:

I I actually think that the better agents are, the more work that they do, the more important it will be that they are easily understood by humans. The idea would be, okay, let's say we have a bunch of of employees, 10,000 employees or 10,000 AI agents dropped at a company. They're all experts at doing something. That ends up not becoming a a problem of giving them the right tasks, but actually it becomes a management problem. Right? There's this whole infrastructure orchestration layer, the thing I always come back to, of of making these things work together. And that's actually going to be the challenge. And that's that's going to require a very human first, ultimately, a product, and that's what we're trying to build. Do you think Elon will be successful with Doge? I think it will be his greatest challenge. There's a lot of self protecting mechanisms in the largest organization in the world, which is the kind of the US government by by spend, by head yeah. It's just it's just this massive, unruly organization. It's not it's not gonna be as simple as Twitter.

**Harry Stebbings** [48:46]:

Are you more excited in the post Trump?

**George Sivulka** [48:48]:

I think the thing that I care most about in the world is that we, as an industry, have very clear guardrails that we can follow and and understand to build the best possible tools to get our tools out to the economy to to make sure that everyone transitions in the best possible way. So I'm

**Harry Stebbings** [49:04]:

ultimately regardless thrive on a better financial system, and we're seeing now a financial system in The US from afar that would seem to be thriving. Objectively Yeah. It would appear that Trump is good for business.

**George Sivulka** [49:16]:

I I won't make a comment here. I I I think that there's a lot funny.

**Harry Stebbings** [49:23]:

Think it went very viral before, the election because they said it's so interesting. There's 99% of CEOs come on the show, and they either shut up or they say they vote for Camilla, and then it ends and then it ends. And they're like, by the way, I'm so Trump. I am so Trump, but it's fascinating. Yeah. For sure. For sure. So I totally understand they're not answering. Yeah. You are not alone. It's okay. But the one question I wanna ask, you mentioned NVIDIA before. Sure. That's a really big question around their ability to to sustain that monopoly. You've seen Google. You've seen Matter. You've seen Amazon all wanna move into the chip play. Yep. How do you think about NVIDIA's ability to sustain their pretty unwavering monopoly so far?

**George Sivulka** [50:05]:

So so NVIDIA has, you know, I think the the the best moats aren't technological moats. They're not data moats. They're actually people moats. People and networks have the most friction to change. One of the things that NVIDIA does best is is is the fact that they made this early bet on machine learning. They created CUDA, which is the way that, as I mentioned before, almost everyone learns how to train models. Like, they learn how to, you know, how to interface with NVIDIA chips for training. And as you're starting to see, maybe that that prediction that I made earlier, the shift away from from training to inference as a fundamental, like, almost macro shift in how people deploy AI, actually I think that will destabilize slightly the dominance of NVIDIA chips. You can start to actually use AMD chips or even custom architectures, which which all the major model providers are also currently exploring to do inference. So you have your academics and your researchers, you know, training large models on NVIDIA chips, but the minute they deploy them, they can deploy them, on cheaper infrastructure. And that and that actually, I think it will be a big change. So I I'm actually still bullish on NVIDIA, but I'm even more bullish on other chipmakers and and custom ASICs to do inference because I think there will be a larger shift to inference moving forward.

**Harry Stebbings** [51:16]:

Is that other chipmaker paradigm existing incumbents, Google, Meta, Amazon, you name it, or is it a new generation Cerebras style?

**George Sivulka** [51:25]:

Probably be large tech providers and and AMD. I I don't know about Intel. Right? I would probably bet on them. There's definitely an opportunity in the market, but chips are hard.

**Harry Stebbings** [51:36]:

Before we do a quick fire, do just wanna kind of resurface back up to the agent layer? Sure. For sure. Are we out of the experimental budget phase?

**George Sivulka** [51:43]:

I think that 90% of the market is still in experimental budget phase, but we're starting to see early promises of actual value. And my entire business is focused on on just those repeatable use cases.

**Harry Stebbings** [51:55]:

Everyone thinks they're a master of agents and agentic workflows. What do they think they know that they actually don't know?

**George Sivulka** [52:02]:

Like, I think ultimately the people in the enterprise that are most excited about AI and, like, positioning it so strongly are CTOs and information technology people. And maybe the thing that we've that Hebbia has always said is that the CTO or the IT folks are actually the people that know the least about the business. The people that that actually understand how to use AI in a business context are those that are closest to the business. And so we're jumping the gun a little bit with the CTOs trying to build the CRM before it's been invented. And, you know, you actually need business people to build the CRM and excel first in kind of order of operations. And so there's a lot of unbundling of AI applications or CTOs trying to go out and build, you know, a very specific, you know, vertical application. But I actually think that building this platform, Hebbia Matrix, is the thing that will unlock users' ability to discover what they can use AI agents for.

**Harry Stebbings** [52:53]:

What will be the pricing mechanism for the future of agents?

**George Sivulka** [52:57]:

It's a good question. There's, like, four canonical price there's, like, consumption based pricing, and there's per seat pricing. There's, like, hey. Rent a salary. So pay a salary for an employee, which is seems a little bit ridiculous, but will be less so. And then and then maybe there's, like, flat pricing. And and I think it it ultimately depends on how you're driving value. Because Hebbia is building human centric AI, the human layer to how you orchestrate an AI agent staff, that that scaling at at inference, we do per seat because it's ultimately always back to the human. I think you'll see all of these new business models and pricing mechanisms. Do you do per

**Harry Stebbings** [53:29]:

seat because it's back to the human or just because it's what they know as a buying mechanism?

**George Sivulka** [53:33]:

I I actually think that we are human first. We're business user first to to the point where CTOs like to pay for consumption or API, etcetera. And, like, you know, business users like to pay per seat because it's how they how they map back to value. But also we want to incentivize change. Tech is not the hard part of all of this. It's hard. But the hardest part of AI change management, no matter what company you are, are people and, like, actually getting people to use the software. When you charge for consumption or API pricing, you're disincentivizing the change. You're saying, okay. Well, I'm gonna penalize you in a monetary way for every time you use an AI application. What the heck? Versus here's a per seat fee. It might be expensive, but use it more. You could you could run more LLM calls on on Hebbia effectively for free than any other platform if you if you actually are driving real chain, and that's what I love to see.

**Harry Stebbings** [54:22]:

Are you ready for a spicy round?

**George Sivulka** [54:24]:

Let's give me the spicy round. We got the tissues out here too. We got the tissues and tissues.

**Harry Stebbings** [54:27]:

Right? This is the in in case you need them to hide behind. So this is a spice round. So this is questions from friends of yours.

**George Sivulka** [54:35]:

Okay. We got we got some changing colors up here. Love it.

**Harry Stebbings** [54:38]:

Yeah. Yeah. I know. It's a full game show. There we go. It's like a fucking David Gutter concert. Perfect. Number one question, would you sell for $2,000,000,000 today? Would I sell for no. What was the single best VC meeting?

**George Sivulka** [54:52]:

It's somewhere between, you know, Peter talking to me about anything but the business and deeply academic things and and Mike taking me on a walk around the the Woodside Horse Park,

**Harry Stebbings** [55:01]:

which is Do you trust Sam Altman? No. Who asked that question? Yeah. Okay. I I don't reveal my sources. But listen. I wanna do a quick fire round. I say a short statement. You give me your immediate thoughts. That sound okay? Sounds good. Let's do it. And what do you believe that most around you disbelieve?

**George Sivulka** [55:19]:

Oh, I have a crazy one. I I believe that UFOs are real. I think a little bit a little bit more on the nose right now, but I I actually believe there's fundamentally different propulsion technology, and that and I think the US government has access to it. Wow. It's like It's very scary. I I have a lot of spicy takes as a specialist. Bring that

**Harry Stebbings** [55:38]:

out. What trait are you slightly ashamed of but has contributed to your success?

**George Sivulka** [55:43]:

I don't think I'm ashamed of it per se, but one thing that I I I always hid was the fact that I'm I'm deeply religious in an industry that's, like, very atheistic or agnostic. It was, like, something that I that was very personal to me, And I think it's been massively contributing to How has it contributed? I think that ultimately when you're doing hard things or when you're you're you're chewing the glass or, you know, working all like, those really late hours, Believing in something larger than yourself or believing in what you do as a vocation or something that's deeply purposeful and deeply meaningful is actually it's additional fuel. It helps you in a way that is, I think, good for the soul. It it it really charges you up. And and Do you pray? I do. I pray for an hour every morning. What? Yeah. I wake up. I sit on a meditation cushion. And I used to meditate. I thought I think meditation's also great. Praying and then putting something out into the universe or or or, you know, actually having a dialogue with whatever you believe, I actually think is even more powerful. It's it's it's almost Do talk out loud? You know, I live by myself sometimes. So but but but, sometimes it's it's all it's all on my I think I think it's incredibly good for the human mind. I think it's it's almost an antivirus for the human mind. And For an hour? For an hour. Yeah. Yeah. I well, it's you know, people meditate. Why is it so weird to pray? I think it's No. It's even better. I didn't know what I would say. When you when you dive into the human psyche and and you're not looking at your phone and and you're lot of the time, it's also a really great channel to think. I think a lot of a lot of the best ideas that I've had at Hebbia have have come from from moments of silence. Yeah. Gosh.

**Harry Stebbings** [57:25]:

I get up at, like, 08:20. My first meeting is at 08:30. There's, an espresso ready for me. I'm like, oh, fuck. Where are my shorts? Oh, god. Mom texted. Jesus. Anyway, I'm I'm here. I'm alive. Hi. So we have different morning routines. What's the gym routine? You're a fit dude.

**George Sivulka** [57:44]:

I try to try to work out every day. I I actually end up mostly channeling the the startup pressures and and anger and anxiety into heavier and heavier things and lifting heavier and heavier things. So it's it's nothing that's, like, in particular.

**Harry Stebbings** [57:58]:

Is Silicon Valley back as the center of all things?

**George Sivulka** [58:02]:

Hebbia, I think there you know, there there was a podcast that actually recently came out where everyone's like, if you're gonna build an AI company, you've gotta build it in Silicon Valley. But, you know, there is one company in New York that is doing a really amazing thing. That company is Hebbia, and it seems like, you know, they're they're actually doing something interesting. And I I do think we are the exception rather than the rule, unfortunately. So I'm a big believer in Silicon Valley.

**Harry Stebbings** [58:23]:

Why? Why are you the exception?

**George Sivulka** [58:25]:

I think that we are a Silicon Valley company in terms of our style of work, in terms of how hard we work, in terms of of how we actually pursue new technology and and and invest in technology. And we started in Silicon Valley, and we have almost only Silicon Valley investors.

**Harry Stebbings** [58:39]:

What have you changed your mind on in the last twelve months?

**George Sivulka** [58:42]:

Longer than twelve months, probably, like, eighteen months ago, it was the scaling and inference thing, like, belief in, like, a new set of scaling laws and that they would be really, really, really important.

**Harry Stebbings** [58:51]:

ServiceNow Yeah. Salesforce or UiPath. Okay. Shag, marry, or kill.

**Unknown** [59:01]:

I wouldn't I wouldn't shag any of them. I I don't think that traditional Shag is like short term excitement. Oh, I I mean, I I don't

**George Sivulka** [59:11]:

think that In case you needed the context. Yeah. I know exactly I know exactly what you're getting at here. I really you know, I'd probably kill them all. I don't think that traditional enterprise b to b applications are sexy. We're an enterprise AI company.

**Harry Stebbings** [59:26]:

Are you a buyer of Salesforce? We are. We are. But everyone says Salesforce is fucked in this next generation. You think

**George Sivulka** [59:32]:

they are or not? I don't think they are. I think that, I think that Salesforce, like, has has built, again, like, a very, very, very sticky network effect with people, and people are the the the shifting function at the end of the day. It's it's not a technology problem. Claude can build a Salesforce. I think Klarna, again, had another Fugazi story about building like, not going off Salesforce because, you know, Claude had built them a CRM. And I just think that the the switching costs, the network effect of changing human beings' habits is too high. Salesforce is is one of those, like, monopolies in that they have so much stickiness, habitual stickiness.

**Harry Stebbings** [60:06]:

You can buy one company in the public markets that will be most benefited by the next wave of AI. Which company do you buy? That's a deep question.

**George Sivulka** [60:14]:

I would probably buy NVIDIA. It's a it's a lame answer or or AMD rather. I think AMD because I believe that they will benefit from the the shift to to inference scaling more than in an outsized way.

**Harry Stebbings** [60:25]:

You can be CEO of any other company for a day. Which company?

**George Sivulka** [60:29]:

Not a company. I'd I'd love to be mayor of New York, believe it or not. I just think that's, like, a fascinating job. I think it would be really, really interesting and would love to make some change there.

**Harry Stebbings** [60:38]:

What question are you never asked by investors, by angels, advisers, employees, journalists that you think you should be asked?

**George Sivulka** [60:47]:

I think that one of the most interesting questions is where does creativity stem from, or or where do you get inspiration from, or kind of like how do you come up with new ideas? Like, I don't I don't believe that people come up with new ideas by brainstorming or in like, I just think that's, again, Fugazi, Fugazi. But I think that ultimately, that question of where creativity comes from I'm I'm I'm also a very big painter. I'm a very big so I I do large scale, like, 10 foot plus oil canvas, oil painting.

**Harry Stebbings** [61:14]:

I heard about this.

**George Sivulka** [61:15]:

Yeah. Where did that come from? I just I I think Are you a poet as well? I I I love to write. I'm probably not as good a poet, but I I actually think that other creative outlets are are really, really good for

**Harry Stebbings** [61:28]:

coming. Me perfect kid

**Unknown** [61:30]:

at school.

**Harry Stebbings** [61:31]:

They're like pain If you want me to try to write it, then I

**Unknown** [61:33]:

will fall on my head. I I

**Harry Stebbings** [61:34]:

told you that story at the start of I mean, if you find what do you find about

**George Sivulka** [61:39]:

painting good for you? I think it's one of those activities where you can channel emotion or intuition or, like, latent thoughts that are somewhere in your subconscious and connect things in a really meaningful way. And so, you know, in a world where there's all this stimulus or you're always kind of thinking or churning through something or all this distraction, you know, you're standing in front of a canvas for, like, ten hours with some nicotine and and and and you're just lost in this art. I think great artists will tell you that they don't even know where paintings come from. It just, you know, is this channeling something. It's one of the best places to think. It's it just gives you connections. It brings up these parts of your subconscious, these connections that that I think you you you can't really access without being creative. Whether you're you're making music or or writing or painting. I actually think that's one of the best ways to process anything.

**Harry Stebbings** [62:26]:

Final one. Do you feel that your parents are proud of you now? I think so. Yeah. I think so.

**George Sivulka** [62:31]:

I think that, you know, they've they've heard about it. There was one moment where I think my my father's boss ended up calling him and he's like, you know, your son's kicking ass. And I was like, well, was a very happy moment for me. That's a special moment. Yeah. The chip remains though. It's not going anywhere.

**Harry Stebbings** [62:47]:

George, I so appreciate you being so open, and I so appreciate the conversation. You've been fantastic to have on.

**George Sivulka** [62:52]:

Yeah. I've I've loved it and appreciate all the research that you've done and and all all the crazy lines of questioning. Thank you, Harry. I appreciate it a lot.

**Harry Stebbings** [63:00]:

I have to say that was such a fun show to do, and I was so so grateful to George who flew over from New York for that episode. It was so much better in person. If you wanna watch it, you can find it on YouTube by searching for 20 VC. That's two zero VC. But before we leave you today,

## Sponsor read

**Harry Stebbings** [63:16]:

here are two fun facts about our newest brand sponsor, Kajabi. First, their customers just crossed a collective $8,000,000,000 in total revenue. Wow. Second, Kajabi's users keep a 100% of their earnings with the average Kajabi creator bringing in over $30,000 per year. In case you didn't know, Kajabi is the leading creator commerce platform with an all in one suite of tools, including websites, email marketing, digital products, payment processing, and analytics for as low as $69 per month. Whether you are looking to build a private community, write a paid newsletter, or launch a course, Kajabi is the only platform that will enable you to build and grow your online business without taking a cut of your revenue. 20 VC listeners can try Kajabi for free for thirty days by going to kajabi.com/20vc. That's kajabi.com, kajabi.com/20vc. And after building your online empire with Kajabi, it's time to scale your global team with remote seamless hiring solutions. So every business is a global business in 2025. But how do you do payroll for your global business and team and comply with international labor laws? Well, remote handles payroll, benefits, taxes, stock options, and compliance to help companies of all sizes pay and manage full time and contract workers all over the world. Remote's global employment solutions keep your team, your finances, and your intellectual property secure. Remote never charges hidden fees, just best in class global employment solutions for a low flat rate. The world's top remote companies love remote. GitLab, the world's largest all remote organization, trusts remote to run their global team. Remote is funded by Index Ventures, Sequoia Capital, and the host of the greatest podcast ever, Harry Stebbings in twenty v c. Ready to learn more? Head over to remote.com/20vc, that's 20vc, and begin hiring within minutes. Enjoy 10% off your first three months by using the promo code two zero v c at checkout. Now that your team is up and running worldwide, make sure your finances work just as hard with Brex, the ultimate financial stack for startups. Let's be honest, building something from the ground up is hard enough without dealing with clunky outdated banks that pile on fees and leave your cash idle. Brex was designed with founders in mind to make every dollar go further so you can focus on building. And here's what really stands out to me. Brex combines the best of checking, treasury, and FDIC insurance in one powerhouse account. You can send and receive money globally at lightning speed, earn yield from day one, and still access your funds whenever you need. Plus, with 20 x the standard protection through program banks, your cash is not just working harder, it's working safer too. It's no surprise that one in three venture backed start in The US with companies like Anthropic, Coinbase, and Robinhood. I mean, my god, these companies are incredible. Trust Brex to help them grow. If you wanna join the smartest startups on the planet, head over to brex.com/startups and see what they can do for you. As always, I so appreciate all your support, and stay tuned for an incredible 20 sales with Carlos Dela Torre on Friday.