🎧🍌 Elad Gil | Acting on Conviction, Future of AI, Biotech, Ambition, Speed, and Non-Obvious Startup Advice
🎧🍌 Elad Gil | Acting on Conviction, Future of AI, Biotech, Ambition, Speed, and Non-Obvious Startup Advice
Four AI trends to watch, why AI is underhyped, how Elad incubates companies, his investing strategy from early to late stage, AI-driven buyouts, and how to get conviction on unpopular investments
The episode is brought to you by Warp
If you’re a startup founder with employees in multiple states, you know managing state tax compliance is a pain.
Warp automates state tax compliance, with a user-friendly experience tailored to the unique needs of startups. They also make it easy for hundreds of startups to provide health benefits for employees, hire international contractors, and even claim R&D tax credits.
Don’t let payroll and compliance hold your startup back. Go from zero to payroll live in TWO DAYS, and receive a $1,000 gift card when you first run payroll.
To inquire about sponsoring future episodes, click here.
Elad Gil is the founder of Color and Mixer Labs, which he sold to Twitter, and an early investor in iconic companies like Airbnb and Stripe, plus upstarts like Perplexity and Anduril.
This is a wide ranging conversation that covers ambition, education, AI, and a broad range of advice for anyone building or investing in startups.
Timestamps to jump in:
3:46 Building cool monuments
9:12 Fixing education
16:38 Why AI is underhyped
19:02 Four trends to watch in AI
19:55 Why there aren’t large biotech companies
23:21 The current state of Elad Gill
24:33 How he incubates companies
26:32 Contemplating AI-driven buyouts
27:29 His investing strategy, from early to late stage
36:57 Why he remained solo for so long
40:19 Getting conviction in unpopular investments
42:53 What made Steve Jobs a good communicator
44:00 The importance of ambition and leadership
46:29 Why Elad puts so much weight in the market
47:45 The evolution of Google’s business model
49:17 How to monetize consumer products
50:06 Analyzing a potential startup market
51:24 How successful products eventually become distribution companies
56:30 Non-obvious startup advice
59:54 When its OK to give up
1:02:20 Advice on raising your first round
1:03:21 Picking board members
1:04:45 How to hire your first three employees
1:06:48 Avoiding bad hires
1:08:39 The importance of speed of execution
1:12:36 Why Elad’s adding to his team
1:14:31 Gardening
Referenced:
Elad’s Podcast, No Priors
Elad’s Book, high Growth Handbook
Find Elad on Twitter and LinkedIn.
Transcript
Find transcripts of all prior episodes here.
Turner Novak:
Elad, how's it going? Welcome to the show.
Elad Gil:
Oh, good. Thanks. Thanks for having me.
Turner Novak:
So I wanted to ask you, really interesting thing that I saw you mention online but haven't talked about much before. It's this thing that you're kind of getting involved with called Monumental. Can you just talk about that?
Elad Gil:
Oh, sure. And it's very early on.
I think basically maybe there's even a broader question. If you look at your life, you start to ask not just what do you want to accomplish in the next year or two, but what do you want to be around in 10, 20, 100 years in the future or after you die or whatever it may be. And then also, what are different ways to have societal impact?
Obviously, family is one part of this, right? And I think one thing that's almost under-discussed on the family side is bigger families or families with more kids, each of those kids tend to have more kids. And so, one could argue if you really want to compound genetically into the future, you should have lots of kids. And then if you have five children and each of them have five kids, that's very different from two having two if you just do the exponent or whatever.
But there's other things that you could potentially do societally besides family. And one thing that seems oddly lacking from modern society is building ambitious and inspiring monuments. And if you look over the history of humanity, you have everything from the Colossus of Rhodes or others sort of large scale structures, the Lighthouse at Alexandria and sort of the other wonders of the ancient world. The sort of things that inspired people in terms of scope and ambition and optimism and action. But more recently as well, if you look, the last real flurry of monuments probably happened 50 to 100 years ago.
And so for example, the Eiffel Tower was originally built as a centerpiece in entrance arch for I think the 1889 World Fair. And I think that was really meant to show off the industrial prowess and technological advancements of France in the 19th century. And so it's this big iron work engineering structure that I think was the largest manmade structure at the time. It's about 1,000-feet tall.
Similarly, the French actually shipped another monument to the US, which was a Statue of Liberty, and that was placed right by Ellis Island, which was the main entry point for immigration into the United States. And it really became this sort of message for freedom and opportunity and the symbol of hope for people who were immigrating or entering the country at the time.
And actually, the sonnet that's on the base of the Statue of Liberty is called the New Colossus by Emma Lazarus. And this is the famous, "Give me you're tired, you're poor, your huddled masses yearning to breathe free." That's the inscription on the plaque on the pedestal of the Statue of Liberty.
And so traditionally, there have been these very inspiring large scale structures and they almost reflect the ambition and optimism of a country, but they also inspire people who see them to do greater things. And we've completely stopped building those societally.
And so the question is, why have we stopped doing it and why don't we do more of them?
Turner Novak:
Yeah, I'm trying to think of maybe what was the most recent relevant one or what's the most comparable thing? I mean, maybe skyscrapers like notable sky, is that fair or...
Elad Gil:
Yeah, there's large scale skyscrapers. And honestly, those imports were two things that countries used to use to anchor themselves, right? And that's why you see these giant structures in UAE and Singapore and other countries that have sort of these strong aspirations towards the future. There's a large scale statue that was built in India over the last few decades, I think two decades ago, a decade ago.
And then when I ask people what's the last monument that's been built in the US, a lot of people actually point to the sphere in Las Vegas, which I think is kind of interesting, right? They say, "Look, there's this large scale structure. It's different. It's inspiring, there's cool art on it." And so it's almost like a digital artifact. But really if you think about a society, we've been doing very little to actually inspire people or to create beauty and art in a public setting anymore.
And so the question is, why is that? Why are we not contributing to societal beauty, ambition, optimism, forward-thinking scale in a way that makes you feel small in a good way
And so this project, and I'm in the process of looking for somebody to help do this because I don't have the time myself to do a bunch of stuff right now, is to really start to ask, "Okay, what are sites where we should build some of these monuments potentially around the world? And then what are some that we should actually go and fund and construct and how can we inspire people going forward?"
Turner Novak:
Yeah, I'm trying to think of one in the US. Maybe Mount Rushmore would probably be one. Is that still technically under construction?
Elad Gil:
I don't know if it is or not, but I mean that was built a long time ago, right? Let me look it up now. Or not that long ago on a relative basis, but I think Mount Rushmore was in the '40s was when it was completed. I think it was 1941 or something was when they actually completed it. But that's a good example. Why don't we have more Mount Rushmores?
Turner Novak:
Yeah, and they do become cultural centerpieces generally. There's kind of, I don't know, like a rallying cry that they almost... Or symbolism for a certain cause.
Elad Gil:
Yeah, exactly. Like the Statue of Liberty, right? What a great symbol.
Turner Novak:
Yeah, it's like the symbol of the US basically, the whole country.
Elad Gil:
Yeah. Yeah. And again, it was a gift from France, but also it was the first thing that you'd see as an immigrant as you came into New York, into Ellis Island. And so, why aren't we doing more inspiration like that? Why aren't we doing more positive things at large scales?
And if you actually look at the cost of doing some of these things, a lot of it is it depends on where you're doing it? In some cases it's land value, in some cases the construction. But you could definitely do some of these things at scale in a really interesting way. So that's what this project is.
Turner Novak:
Cool. Yeah, I thought it was fascinating. I'm excited to follow on. I wish I could help, but not my skillset.
So another cool thing though that you're working on, talking about big positive change, you also mentioned you're starting to think through a new chain of K-12 schools that's kind of inspired by Ancient Greece. Another thing I thought super fascinating. Can you talk about that a little bit?
Elad Gil:
Oh, sure. And again, this is another one. This is even more of a concept than the prior one, although I looked pretty heavily into it. I talked to head of school recruiters and all the rest of it. And so this is another one where I would need somebody to drive it, and I'm very happy to sponsor that.
It's kind of in part inspired by this quote from Archimedes, which is, "If you give me a lever long enough and a fulcrum on which to place it, I shall move the world." Right? And so the question is how do you provide the skill set for kids so as they grow up, they want to best help move the world to a positive place? And what are the basic skills you want to give them?
And there's the obvious basic skills around reading and writing and math and history, critical thinking, basic financial fluency, probably physical fitness, diet, exercise.
But then there's two things that I think seem to be decreasingly available in some of the systems of schooling. One is just moral values and a basis for how to think about the world and basing it on older principles, not modern reinterpretations of morality, which I think are often flawed, but then also emotional resilience.
And so actually, I spoke to one of the people who were involved really early with this school called Proof School up in San Francisco, which is a math-centric school. I think it's 5th grade through high school. And it's basically for kids who love math. So it's small populations, really interesting school, really interesting curriculum. But from talking to one of the people who's involved with their early, they mentioned that the first year they're just teaching the kids resilience because math is hard. And so how do we teach them that things are hard and that's okay and you just keep going?
Turner Novak:
Yeah. How do you?
Elad Gil:
And it's the only time I've heard a school talk about this, of any school, is how do we build resilience? And so I think there's other aspects of things that just aren't emphasized that are actually really valuable.
And you can think of that across multiple dimensions of, "Okay, what does a modern education mean and what should it look like? But then also how can we be inspired through thousands of years of history across multiple cultures in order to really influence this?"
And so I wouldn't say it's just Ancient Greece, I just think it was inspired by a quote. And then I think there's lots of cultures you can draw on, but fundamentally, there's these really fundamental things that used to be taught that are no longer seem to be taught, or at least they're not as aggressively suggested.
I'll give you another example. If you look at all the fairy tales, the grim fairy tales, whatever it may be, a lot of them basically teach you that there are many people who are fundamentally bad people. So an example of that would be Hansel and Gretel and the Witch in the Woods, she coaxes you in with candy and then pushes you in the oven, right? Or little Red Riding Hood and the big bad wolf. And the wolf kills grandmother, dresses up as her, and then pretends to be grandma so that the wolf can then kill you, right?
That's the stories that used to be told. It was these big warning signs of there are some bad people out there who pretend to be very good and will choose moral high ground, but are really wolves hiding themselves.
Now, if you look for example on the Facebook portal, which I think is a wonderful product for video conferencing kids to their grandparents or whatever it may be, so I think it's a fantastic product, the Little Red Riding Hood story on that has now been upgraded, or updated I should say, to include a wolf that instead of killing grandma is just visiting and dressed up as her, and then it has a dance party with Little Red Riding Hood and they go shopping, and that's how it ends.
Turner Novak:
Hahaha what?!?
Elad Gil:
And you're like, "Is that really the morality of the current day and what does that mean for our kids?"
And so it's back to this, like, what are the values and types of thinking that used to be emphasized that are probably pretty universal and historical in how have we degraded some of those types of arch types and how have we weakened things that actually help people understand the world, build resilience, understand that not every person is a good person and how to deal with the world more fully. And so I think there's a lot of really interesting things you can think about from a curriculum perspective there.
Turner Novak:
Yeah. It reminds me of maybe you've seen this data, I think there's a map where it shows the distance that a kid used to play from home. 50 years ago, they'd be five miles away. 20 years ago when I was a kid, maybe it was like a mile radius.
And then today it's like, you can't leave the house. You need a parent with you at all times. A little bit concerning as a parent. I mean, we were both talking about how we have young kids. It's like, how do you balance that of protecting your kid obviously, but then teaching them resilience? How would you do it?
Elad Gil:
Yeah, it's interesting because I was talking with my wife about this earlier today, which is when I grew up, you just from, I don't know what it was, say 1:00 or 2:00 or 3:00 in the afternoon, whenever I get home, until my parents got home, it was just me and my sister on our own. We'd make food for ourselves. And this was from the age of six or seven. And so we would just take care of ourselves and we'd go out and play with the kids in the neighborhood. That'd be it. There wouldn't be all this other stuff around it.
And relatedly, if you kind of look at how people talk about themselves over time now, you often see people at older and older ages into their 20s talk about themselves as if they're still kids. And it used to be you'd have kids in your early 20s. You used to have children yourself.
And so I think a related thread, which Patrick Collison has brought up before is, where are all the people in their 20s who are doing amazing high responsibility, broad-based societal things? Or where are all the really young founders with big companies? Because if you go back 10 or 15 years, there's a lot of people in their 20s who are running Stripe and Figma and Facebook and all these things.
And the only one I can think of now is maybe Alex from Scale, right? But there aren't that many people in their 20s who've already built something significant. And so why is that? And that's true across all of society, right? I believe the person who oversaw the Hoover Dam was quite young. The first head of the NIH I think was late 20s.
Why did society turn over so much in terms of actually giving large scale responsibility to young people? Is that part of this sort of delayed adulthood that we're seeing experienced?
Turner Novak:
It kind of reminds me of founding fathers of the US. Some of them were early 20s, early 30s.
Elad Gil:
Yeah, no, seriously. And there's a lot of that throughout history. And maybe the place where it's expressing itself right now is an AI where there's a lot of really smart young people driving really important parts of that future.
And so maybe just where the smartest or most driven people are going has shifted over time, and that could be one explanation. But there's probably other explanations as well. And so I think it's kind of an interesting question that Patrick's raised over time.
Turner Novak:
Yeah. I think slightly different train of thinking, slightly different topic, but on the point of AI, I've heard you say before... Well, I guess maybe common knowledge at this point is AI is overhyped. I mean, we've gone through it all. There's been a lot of interest, but I've heard you say specifically, you actually think it's underhyped.
Elad Gil:
Yeah, I think that the reality is most of the adoption of AI is ahead of us. And if you look at ChatGPT launching 18 months ago now or whatever it was, is the starting gun for this new era. People waking up and realizing that this is interesting. And by people, I just mean broader-based companies, society, even the technology industry. I was investing in a lot of the gen-AI stuff maybe two, three years ago, and there just wasn't anybody else, or very few people were doing it at the time.
I led the first round in Perplexity, for example, before anybody was doing anything there. Or Character.ai, or some of these other companies that got involved with quite early. And then ChatGPT comes out and suddenly everybody's aware of this big technology shift.
But if you think about enterprise planning cycles, they're on 12, 24 months planning cycles in some cases for the biggest companies. There's been a lot of internal prototyping done. But the real wave of AI hasn't fully hit yet. And so I think if anything, we'll see a lot of acceleration of adoption in two years or whatever the right timeframe is for enterprises to really start using it. Large-scale societal use cases.
It's the old William Gibson saying that the futures here is just not equally distributed. And we have that going, but at the same time, we also have these step-ups and capability happening with each incremental large scale model, right?
And so certain companies, Harvey would be a good example on the legal side. They're a legal AI sort of vertical company. That product couldn't work on 3.5, on GPT 3.5, but it could work on 4. And so what markets or what use cases or capabilities does GPT-5 or 5 equivalents open up, or 6 or 7 or 8? And so there's almost like this GPT ladder of capabilities coming. And there's other things that build around reasoning and memory and other types of modules.
And so I think it's very exciting, but I think most of it hasn't happened yet.
Turner Novak:
And you think the models will keep getting better. I think there's a little bit of a debate that's come out as some people say we're almost peaking. Do you agree with that?
Elad Gil:
Yeah, I don't know anybody in the AI research community who believes that.
Turner Novak:
Okay.
Elad Gil:
Literally, I don't know anybody who's at one of these major efforts that believes that, that it's peaking in terms of the capabilities. Everybody's like it's early days. It still seems like there's a lot that you can just do with scale, yeah.
Turner Novak:
So what are you kind of expecting next then like GPT-5. Or whatever the next big leg up of model power, what are you kind of expecting to happen?
Elad Gil:
Yeah, so I don't know about 5 specifically. I think there's three or four general trends that are definitely coming.
One is that the increase in capabilities and understanding and ability to logically deduce certain aspects of the world or write code or do other things, that's one.
Two is different forms of multimodality. So I actually have been really impressed with the Meta AI launch in terms of generating images, animating them with the click of a button. There's some really neat stuff in there.
A third would be stronger reasoning and chaining of actions. So more agentic flows, right? And so how do you actually have things that start doing more and more sophisticated actions on your behalf? Long context window would be another one. So Google has a million token context window. I think Magic had 5 million already.
And then what happens when you have tens of millions of tokens in a context window? And what are all the things you can do with that? And that seems to be really important for things like biology models.
Turner Novak:
I was going to say, speaking of biology, you got a PhD in biology. That was kind of originally how you started your career. Are you still closely following biology? Are you doing many investments there? Can you talk through that?
Elad Gil:
So my academic career is a mix of I have degrees in mathematics and then in biology. And then to your point, I have a PhD in biology.
It's kind of interesting because if you take a big step back and you look at the entire biopharma market cap, public companies, private companies, big pharma, small biotech, all of it, it adds up to Apple and Microsoft roughly.
Turner Novak:
Okay.
Elad Gil:
So two companies in tech equate to the entire industry in terms of how they're being valued. And that's companies, if you look at US healthcare, it's about 20% of GDP, and then biopharma saves 20% of that. So you're saying 4 to 5% of GDP is equivalent to two companies in tech. If you throw in NVIDIA, then those three companies are 50% more valuable than the entire industry.
And if you look at biotech in general, you can ask when was the last time a $50 billion plus company was founded? And if you do that in tech, there's lots that you can think of. You can think of Stripe and you can think of Snowflake and you can think of Shopify. You can think of all these companies that are roughly there or more.
If you look at biotech, I think the highest most recent market cap is Moderna, which is about $40 billion and that's because of Covid, right? It wouldn't be there at all if it wasn't for our pandemic response. Vertex is about $100 billion and Regeneron is worth tens of billions or a hundred billion as well, but those were started in the '80s.
So it's been 30 to 40 years since the last wave of very big biotech companies were started in terms of biopharma. There's other things in tooling like Illumina and things like that, which was started in the '90s I think.
So then you ask, why is that? And obviously, it takes time to develop drugs, but is there something more systemically distorted about the biopharma industry relative to technology that prevents the ability to actually build large scale companies on a relative basis? And you can start getting into industry structure and regulation and founder models and all the various aspects of it.
Turner Novak:
So you are not investing in biotech or biology specifically right now? Is that a fair thing to say?
Elad Gil:
I'm not doing much. Sometimes I'll invest in healthcare infrastructure companies. So, more like software to deal with the plumbing of the healthcare industry. I've over time looked at some of the biomodel companies. I'm a small investor in something called Inceptive, for example, which was started by Jakob Uszkoreit, who I guess I probably mispronounced his last name, but who was one of the main authors in the Transformer Paper and is just a brilliant computer scientist. But I don't do much in biotech.
Turner Novak:
Okay.
Elad Gil:
Actually, I should qualify that by saying I am involved with a couple different anti-aging biotech companies.
And so for example, one of them is called BioAge, and I invested in them at this point, I don't know how long ago, seven, eight years ago. They have some really interesting drugs now for muscle wasting and related pathways and sort of anti-aging related biotech. And so they started off as a more computational and sort of AI-centric company. And then they morphed over time into a drug development company. And they have some really interesting drugs in their pipeline now, specifically around aging, which I think is really exciting and important work.
Turner Novak:
Yeah, I agree. I mean, it's definitely an interesting topic. And that was actually a question from Ayush at Warp, one of your portfolio companies. He was curious why you stopped investing in biotech or why you don't do it as much.
I guess, level setting the conversation, what is just the current state of Elad Gil? Like, you started some companies, you help founders, you're investing, you have a couple people that work with you. Can you just talk through the current state of everything going on your end?
Elad Gil:
I guess the common themes from me are basically a mix of some form of technology related optimism. Biotechnology is most of the time a force for good, helping to pull hundreds of millions of people out of poverty, making people's lives better, easier, et cetera, growing the economy and global GDP and things like that.
And so the question is, if I think it's an important sort of motive force for the world, what are different ways to engage with that?
The way I think about my activities broadly is sort of a mix of incubations, funds and investments and operating companies for the technology of future. How do you build towards that important future state?
And so on the business side, there's stuff like that that I'm doing. And then on the non-business side, there's these broader societal questions of how do you think about societal inspiration and how do you encourage people to work on important things and build important things? How do you identify young talent that could help a boost, could make a big difference for in terms of their future impact, things like that?
Turner Novak:
So you mentioned incubations. Do you do a lot of those? I think I can think of at least one that's maybe a little bit more public, but how common is that and how do you do those?
Elad Gil:
Yeah, I basically traditionally have done something like that every year or two.
Right now the most recent thing I've done, which I'm super excited about is a company called Braintrust. Where there's this great founder, Ankur Goyal who was at SingleStore as VP of Eng. And then he started a company called Impira that Figma acquired, and he was driving parts of the AI platform at Figma and then he left to start Braintrust.
And that's one where he and I just started iterating on what are pieces of infrastructure that we thought people needed for this new sort of AI or LLM world around evals and other products. And originally it was just going to be a side project and an open source thing and just done for fun. So it wasn't meant to be a company. And then it kind of morphed into one as we talked to different people and they just expressed the interest or need for the product.
And so now it's being used by everybody from Airtable, to Notion, to Replit, to Instacart, a lot of the early adopters that are actually have product-centric launches of AI who are actually using it. And so I think it's kind of emerging as an important product in that infrastructure for AI space and eval for AI space. So that's the most recent example.
I think there's a set of things I'm starting to work on now that are incubation-centric again. And if you look at a lot of the AI world, one of the things that I find really striking is how much opportunity there is and how oddly clustered all the startups are in a small number of opportunities and then there's just white space.
And so you'll have seven companies all jumping on the same identical thing and then there'll be all these opportunities that nobody's doing anything on. And you end up with this sort of model where I have a very small team working with me now and we've gone through various areas of services, for example, and mapped out what are all the parts of legal tech that could use a startup or all the parts of accounting or sales motions or whatever it is. And then you're like, "Okay, do I invest in a company? Do I incubate? Or in some cases, do you do a buyout?"
Turner Novak:
You would do that?
Elad Gil:
Potentially. Yeah, I mean I've already committed to back to AI driven buyouts. And if you think about it, there may be industries where AI has such a radical impact, a cost structure that you have a real advantage as an operator of that business or maybe you can just expand the footprint of things that you're doing.
It doesn't always have to be cost reduction. It could be market expansion or market share expansion. And if that industry is a slow adopter of technology, actually building a company to sell into that industry may not actually be useful because it may just get adopted too slowly. And so maybe what you do is you just buy some assets and roll them up and go after it.
And so I think there's lots of different opportunities in AI. And that's back to there's so much to do right now. And for some reason, there's so few people still showing up for so many of these opportunities that I'm increasingly taking this sort of multi-pronged approach of invest, build or buyout.
Turner Novak:
So you mentioned doing a buyout. I know you do some angel stuff, early investing. I think you're known for recently you started leading some later stage rounds, decent sized checks.
What's kind of the gamut of what you do when you're investing in a company in terms of how you get involved, check size, what stage, all that stuff?
Elad Gil:
Most of the investing I do by number is early stage. And most of what I do by dollar is late stage.
From an early stage perspective, at this point, my preference is to lead or co-lead rounds whenever possible. Because I just don't have that much bandwidth, and I'd rather just be deeply involved with a fewer number of things than be slightly involved with lots of things. And so there's a number of pre-seed or seed rounds that I've led in different companies.
Braintrust, I led their pre-seed. Perplexity, I led their pre-seed, etc. But there's a variety... At this point, I think I've led early stage rounds for half a dozen or so, or maybe a little bit more than that AI companies.
And then on the mid to late stage side, in some cases, a major participant. In some cases, I lead a round. And so I've co-led rounds or led rounds in things like Anduril, I led their Series D, although I got involved with the company at the seed. I led or co-led two rounds for Applied Intuition now. They're really emerging as one of the main companies in sort of vertical SaaS for automotive. I led or co-led the series B in Harvey, which is the legal AI company.
So there's a variety of companies at different stages where I've now been the primary lead on the round or in some cases I've been willing to split it with someone.
Turner Novak:
Yeah. And that started a long time ago you wrote your first angel check. Do you remember at what point in your life your career that was at?
Elad Gil:
I wrote my first angel check around 2009 or so, I think.
And then, I've actually been leading rounds quietly in the background, or been doing largely stage things in the background. It just hasn't been very public. And then I think more recently, founders have asked that I become more front and center on some of these things.
And so sometimes I'd be the biggest check in the round and not even be announced as part of it. And I think that's now shifted in terms of perceptually people wanted me to be a bit more front and set in terms of some of the things that I'm doing.
Turner Novak:
So were you trying to be a little bit more under the radar? Is that why? You just didn't want a lot of distractions?
Elad Gil:
Yeah, I didn't see the point of it. I used to say they call it private equity for a reason. It's private. And so in general, my bias has been, there's certain things that I've just kept really quiet. Where I've taken pretty large positions and important companies over time and just haven't talked much about it.
Turner Novak:
But the founders actually now want you to start talking. Do you know why that is?
Elad Gil:
I'm guessing it's just I've been lucky over time in terms of, maybe being involved with some important companies. And so I think that's a cause people to feel like sometimes my involvement is a positive signal to other parts of the market, maybe.
Turner Novak:
Yeah, I've heard. I think it was Marc Andreessen said this, where a big chunk of the job of an early stage investor is just basically lending credibility to founders. Like, somebody who's just getting started and no one's heard of them, but "Oh, they took money from Elad Gil. He thought they were good enough to take a bet. Maybe I should take the meeting or maybe I should consider working there."
So I think it's a valid reasoning.
Elad Gil:
Yeah, it's interesting because I've always helped people close candidates, but that'd be a good example where who's helping to close a candidate matters in terms of helping hire or do other things.
One of the more emotional moments since I've had over the last two, three years as an investor. There's a company I funded and as I led their seed round and we went and we grabbed some dinner and a drink to celebrate it.
At the end of the dinner we were outside, and I was about to leave and one of the co-founders said something along the lines of, "My parents thought I was crazy for going and doing this. And then you invested and they looked you up and then they felt like everything was going to be okay. And we come from a really modest background. This really made a huge difference for me and my family. So thank you so much. I don't know if it'll work or not, but you backing me means the world."
And I was like, "Oh my God." That was kind of a special moment, and so...
Turner Novak:
Yeah, I've definitely heard that before. I mean, one prime example, I think a lot of people will make fun of certain credentials, like lists and stuff that people will get on. I won't name anything.
But for some people, they might be an immigrant and they're moving to the US and they suddenly decide, "I'm going to start this startup thing" and they get off the track of maybe the more tried and true doing investment banking or being a doctor and they do this crazy startup. But then they get on this list or they're mentioned in the Wall Street Journal or something. To some people, that might seem like nothing.
But to their parents, to their family, to their friends, that's a way bigger deal than anyone can appreciate. So I think it's a very underrated aspect of all this that a lot of people don't necessarily think about when you're in the weeds on the day-to-day.
So then, at what point did you actually start taking on outside capital? So you did your first angel investment. Did it go so well you just kept recycling or was it a certain point where you're like, "Okay"?
Elad Gil:
I mean, these things take so much time, yeah. I mean, basically I invested roughly the majority of my liquid net worth into startups and I just ran out of money.
As I kept helping companies over time, they kept offering me to invest in leader in later rounds. And so that led me to start raising SPVs or these single purpose or special purpose vehicles where you're just investing a fund in a one company.
And then some of the folks who worked with me on those said, "Well, why don't you just do a fund?" And so then I started raising funds after that.
Turner Novak:
Okay. So it was a organic process?
Elad Gil:
It was very organic, yeah. And at this point, to your point, I can do a mix of small personal checks. I can lead rounds. I can do SPVs. I can do sort of the whole gamut of things.
So anything from leading pre-seed on through to leading a late stage multi-hundred million dollar round.
Turner Novak:
And then you've actually, like you mentioned, started hiring a team to work with you. How big is the team at this point?
Elad Gil:
The team is really small. And so I always had a chief of staff or somebody else helping me. Often, that would be somebody who would join me for a year or two and then go a company. And so I was kind of viewed as a temporary thing.
That's shifted now where if there's people who want to stick around in the long-term, great. If I can just help people over the course of their career instead, great. I view most relationships as just long-term relationships.
Right now I have three people helping me. One of whom is on a one-year rotation to go start a company.
Turner Novak:
Okay, what does that mean?
Elad Gil:
It basically means he's joining me for a year to help out with the idea that I hope he will go start something out of the experience.
Turner Novak:
Okay. So kind of an EIR is maybe the most comparable...
Elad Gil:
We're kind of calling it like developer in residence or builder in residence, you know? His background is really cool. He was an undergrad in CS at Stanford. He dropped out. He was heavily involved with a bunch of stuff in the EIC and then started working with me. And then the idea is he'd go and start something next.
In general, what I've focused on is mainly bringing on people with strongly technical or operating backgrounds. And then I've also added a really strong person who used to be at Dragoneer and Vista before that, to help with more of the traditional financial centric investing stuff. And I'll probably add one or two people on that side and one or two people on the early side.
And on the early side I'll focus on operators again. And so the other person who works with me was a manager at Ramp, and was undergrad and master's in CS from MIT. And she's exceptional engineer. And so a lot of what we do is things like, "Let's go try out all the different AI infrastructure projects and benchmark them. Or let's go and try out these sets of dev tools and use them to build something internally and see how that goes."
So it's this really interesting mix of builders and operators with some finance expertise smattered into it too, which I think is the right mix for technology.
Turner Novak:
Yeah, I agree. I mean, you need all the pieces. I mean, you think about how do you start a company, how do you find a company. Usually someone who understands the technical aspect of whatever you're doing. There's usually someone that can sell it. There's someone who can do all the boring internal stuff that no one else wants to do, but is super important. So you need a well-rounded team.
Elad Gil:
Yeah. And you also want to be able to help the companies in different ways. And I think for later stage companies, sometimes it's like, "Hey, I have somebody who can help you with M&A modeling," or "Hey, I have somebody who can help you with certain types of financial planning or whatever it is."
And so again, I view them as complimentary skill sets between the operators and people who can help on the more sort of investment centric side.
Turner Novak:
Makes sense. And you mentioned you're going to add to the team or you're thinking about it. Are you going to post a job posting? Is it mostly your network inbound? How is that probably going to happen if you had to fast-forward a year or so? What would you be looking for?
Elad Gil:
Yeah, I think it's going to be a mix. There's no urgency to do it. I just think I both want leverage, but I also want to provide this flexible career path where people could join me forever. But also I'm assuming a number of people will turn over time and go do other things, and that's great.
I mean, it's just like a company, right? I kind of view it more as a company than anything else. And this is back to that theme of its incubations, it's investing, in some cases maybe we buy operating assets and run them. And so I think core investing is something that I'm going to keep doing and that I'm excited to do, but at the same time, I like building things. And so I think on an ongoing basis we'll have a mix of things going on.
Turner Novak:
Yeah. I also want to give a shout-out to this. I don't want to forget. You have a podcast that, I don't know if you consider that building, but you have a podcast called No Priors. That's pretty good that people should check out. We'll throw a link in the show notes.
And then this is a question for my friend David Lee. He was at SV Angel for a few fund cycles. He was wondering why you remained solo for so long. You probably could have joined different fund. Why remain independent for so long?
Elad Gil:
Yeah, much of my investing career was done while I was also running companies or working at companies.
And so I sold my first company to Twitter and that gave me a little bit more of a financial cushion to do more investing. And so while I was working at Twitter, I was also doing angel investing or in some cases some of these SPVs and things. And then I left and started a company called Color where I was CEO for a couple years. And so a lot of what I was doing initially was done on the side.
And I think operator investors often both have a very different perspective on the world because they know what works and doesn't work and what's important and what's being adopted and where their real needs and problems, but also they can be really useful to people starting companies in terms of advice and access and other things like that. And so a lot of my career has been running and starting companies. And so investing was initially one component of what I was doing.
Turner Novak:
And then in terms of investing, I'm sure that... Because I mean, you kind of mentioned it before, pretty early investor in a couple companies that are maybe not consensus. Does it help being independent to just have your own conviction? You don't have to worry about convincing a bunch of people that it's the right decision. Is that fair to say?
Elad Gil:
Yeah, I think that what I've observed is that you need to have the ability to do things that may be unpopular. And there's different forms of things being unpopular, right?
So when I first invested in Anduril, it was a very controversial company. Donald Trump was getting elected. There was concerns around the border and ICE and all these other things. It was always a company that engaged in a bipartisan mission, right? National defense has always been bipartisan. But for some reason it was perceived a very specific way.
And so that was a company where after I invested people sent me angry texts and were very upset, one well-known founder asked me how I could do such a fascist thing or fascist investment and things like that. You have two things you could do at that point.
You could say, "Oh my gosh, you're right, and I'm so sorry." Or you could say, "No, I have conviction and this is important and this is important for the world." And the same people now, the same people then who were saying all these awful things because of Ukraine are like, "Oh my gosh, this is such a great company." And you're like, "Well, you were denigrating the company a few years ago and suddenly you're all for it. What changed? And why don't you have conviction?" That'd be one example.
I remember when I invested in Stripe, because I started investing in Stripe I think at the Series A when there was eight people. And then I kept investing in multiple rounds after that. I was a reasonable size participant in their $9 billion round. And at the time, that was very controversial. It's such a high price and how could you do that and what's the real return left in Stripe and all this stuff. And so that's a different type of conviction, right?
And then things like getting the Generative AI early, it was just driven by this belief that it's important technology. And when GPT-3 came out, it was a big step function.
And so Anduril was a conviction in the importance of a market. Stripe was a conviction in a company and its momentum. And then investing in gen AI early was conviction in a technology wave.
Turner Novak:
If I'm somebody who I'm trying to figure out how do I get my own conviction on something, I'm not really sure how to phrase this question, but maybe you kind of get what I'm asking. What would you recommend someone do back up and say, "Okay, here's how I get my own conviction on something or even convince other people." Are there strategies for doing that?
Elad Gil:
I think that if something isn't popular, it's going to be hard to convince other people. And so you kind of have to take that off the table unless you’re a founder, and then you always have to convince other people, right? The best companies are always non-obvious at the time because if they're obvious, everybody would be doing them and then you wouldn't have to convince anybody of anything.
And so often it's more how do you double check the assumptions that are being made either broadly societally or by yourself? And how do you just question things? How do you make sure you're asking the right questions and revisiting assumptions? That's really, I think, the core of it.
And then are there things that are just always important but that people kind of forget? I feel this is back to our original conversation on what are kids taught and the big bad wolf is suddenly a good character, right? And you're like, "Well, that goes against all of human history, perhaps that's the wrong lesson to teach."
Turner Novak:
Yeah, that's fair.
And I think of Anduril, I'm just trying to remember what it was like in 2017, I think that's when the company started. I can't remember the dates, but I mean it was very unsexy to do. It was evil almost.
I remember we saw recently the Google protests, of Google kind of kicked the people out. But I remember there were times Google employees would protest working with the Pentagon, I think. I don't know if you remember this.
Elad Gil:
Yeah, no, no, seriously. That's the reason I got interested defense tech. It's because Project Maven got canceled by Google, which was the project where they were working with the defense world. I said, "What a great opportunity for a startup. If the big tech companies are going to abandon it out of some weird employee activism thing, if they're going to capitulate, then suddenly there's opportunity for startups."
And so that's when I started looking around. So I had the opposite reaction. I'm like, "What are these people thinking? Of course, defense is important. Of course there's going to be wars again. Of course there's adversaries and people who don't think the way that we do in will and try to impose their will on the world like Putin is doing right now in Ukraine. Of course that's going to happen."
It's not, "If it's going to happen." It's, "When?" And it's back to, what are we teaching people? What frameworks of the world are we teaching people? Are those actually very negative in the long run because it's causing a misunderstanding of how the world works. We're divorcing reality from what we think is important, and reality always comes back.
Turner Novak:
Yeah, that's right. So a good way to think about this would be just observe what's happening in the world and figure out places where you just think the assumptions are wrong or it's out of whack. It's gotten disconnected from reality in some sense, and then figure out what the opportunity is on the investment side or on the side to start a company.
Elad Gil:
It's not even necessarily out of whack. It's just like, what is reality? How do you see reality more clearly?
friend of mine used to work with Steve Jobs at Pixar. He said that his biggest takeaway on Steve Jobs wasn't that he had a reality distortion field. He said the thing that jobs is really good at was listening more clearly than anybody else, so really that clarity of hearing what somebody's really meaning under their words and then communicating more crisply and concisely and directly than anyone else.
And to me, that feels a little bit like seeing and communicating reality. And so maybe it wasn't a reality distortion field because it all happens. So you're not distorting reality. It was possible. It was the ability to see what's possible and then communicate that to other people.
Turner Novak:
Which is a hard skill.
Elad Gil:
Oh, it's very hard. Otherwise, everyone would be doing it, right? We'd be a thousand years ahead technologically or whatever. But I think that's really important.
And so sometimes you also meet people starting companies who have that clarity of thinking and that clarity of communication and articulation of what's important and why in the future. They're willing to change or adapt relative to those viewpoints because they're rechecking reality. And I think that's incredibly powerful as a scale, both as a founder as well as an investor.
Turner Novak:
Yeah. In terms of then backing founders, the types of founders that you want to partner with and work with, that sounds like one good trait, any other traits that you specifically look for?
Elad Gil:
One thing that I think was kind of underweighted for a long time that I've been reminded is really important is just the degree of ambition. I'm going to call it aggressiveness. It's the wrong way to phrase it. But that drive and that willingness to just go hard against something.
And it's not just hard work, it's just that intensity. I think that's been a little bit forgotten. Sometimes there are people who come across as more mellow or chilled out, but they're very intense and they're very driven and they want to win.
And so maybe it's that hunger to win, I think, got really underweighted for a long time in the technology world, and it actually matters a lot in addition to being smart and finding product market fit and learning quickly and all the rest of it. Sometimes that's phrases resilience, but I think resilience is a little bit different.
Turner Novak:
Yeah, because resilient, you could just keep losing and you're terrible, but you don't give up versus ambition to win, aggressiveness, intensity. Hopefully, I rephrase those all right. Those are a little underrated.
Elad Gil:
Yeah, I think it's been underweighted for a period because it became really unpopular. Travis, who I thought was really good, left Uber and other people were kind of viewed as, "Oh, too aggressive." And so it's almost like people were trained to modulate their style down or to back people who are just really nice and you're like, "Yeah, it's good if people are nice, but they should also be driven."
Turner Novak:
That's true. I guess you got to find the happy medium between being nice-ish, polite-ish, but also getting shit done.
Elad Gil:
Yeah. And also, I think there's some very nice people, but they're very motivational.
It's back to the coach. Like, do you want a coach who's just really nice to you and hand delivers the kombucha to you? Or do you want the coach who wants the best for you and is going to push you hard on it? And so it's kind of that thing.
And I think the other thing that kind of happened in a lot of companies is a lot of CEOs confused popularity and leadership. And I think people are increasingly realizing, "I need to be a leader. I don't need to be popular."
Turner Novak:
Yeah, I can definitely see that.
There's one company, a public company that I invested in that, I know just based on conversation with people that work there kind of post-"-it-not-working", was issues other than the actual product and making a company became more important. And I mean, that's a terrible recipe if you're an investor in the company.
One interesting thing that I agree with you on, and I kind of want to hear you talk about, so most investors will emphasize the founders and the team. That's the most important thing. T that's obviously important, but you have a heavy weighting on market as well. Can you kind of explain why you think about it that way versus the more traditional founder centric?
Elad Gil:
Yeah, and I should say I've started two companies myself, so I think founders are incredibly important, right? Otherwise, I wouldn't think my involvement with things I've started are important.
But when I look at what causes a company to win or a team to win, it's just product market fit. And so it's, what is the product you're building and what market? And is there a natural momentum to it? Is there a clear why now that would suggest that this thing should exist now when it perhaps couldn't have existed before? And is a market actually open and ready to adopt something?
And so I think product market and product market fit matters a lot more than anything else. And I've seen really amazing teams get crushed by terrible markets. And I've seen pretty bad teams do very well in extremely good markets. And then of course there's the magical great team, great market, and then you have a Meta or a Google or a Stripe or a Databricks or one of these companies.
Turner Novak:
Yeah. And you just think about what was Google's... Their business model, I guess. At the end of the day, the market was their customers, helping businesses sell things. That's ultimately what was happening. Same with Facebook. It was just, you're helping other businesses make money. That's a pretty good market.
Elad Gil:
I think that's where it ended up. I mean, really Google's market in some sense initially was helping people find things, right? Because it really started off as search. And for four years they struggled to find a business model. They went through three business models before they landed on ads.
Turner Novak:
What were the business models? They licensed their technology originally, right?
Elad Gil:
Yeah. So they tried to syndicate out search, and that turned out to not be a very meaningful revenue stream, although it helped them later with ads.
And then they tried to do a search appliance. So they'd actually sell these boxes for enterprise search. That was one of their main lines of business that eventually turned into Google Enterprise. I mean, they still have a business there, but they thought, "Oh, maybe we just package up search as a search engine on a piece of hardware and sell it to companies for internal search."
And then they morphed over a few different types of ads and then they ended up with these sort of a CTR, a CPC-based ads auction where you're doing ROI based advertising as a starting point. But they went through multiple business models before they actually got to that, but the key thing was the consumer stuff just kept growing. And consumers were looking for things.
So then usually when I look at the monetization of consumer products, the way that you monetize it as a natural extension of what it's doing for the user. And so in the context of Google, it was like, "Help me find stuff." And some of the stuff I want to find is commercial. And therefore ads make sense as a monetization mechanism. It just fits well with what people were using the product for.
Turner Novak:
So if you're building a consumer product and you're trying to make money, the best way to monetize is the core problem that you're solving that you go even deeper on, if that makes sense. Or just make sure that the monetization really...
Elad Gil:
Yeah, the monetization is usually a natural extension of the product itself. That isn't always true because sometimes I'll see banner ads or whatever. But for the really large scale, best working products, that tends to be the case.
For market places, it's like, "Hey, I'm just going to cut out the transaction because I'm helping create liquidity between buyers and sellers. And so that's how I monetize."
But if you look, for example, at the early ads products that Twitter launched for example, it was promoted follows, "Hey, help me uncover interesting people." It was promoted tweets. It was a variety of things, right? There were different types of ad units, but there were ad units that were natural to the platform.
Turner Novak:
That makes sense.
And then in terms of just analyzing a market worth building a startup in, how do you kind of think about that? And maybe another way to tee it up is, you have this framework that I've heard you mentioned, just that value accrues to different types of companies at different waves of technological cycles. Can you just talk through that whole process and then maybe how would you approach that today? Or think about a market, tear it down and figure out what to do?
Elad Gil:
I think there's a few different views on it. And I don't think there's a one size fit all for every market or every approach. So for example, in AI right now, there's just a lot that d you can do in different types of services and you just go through them one by one and say, "Where does AI have an impact? Where there income at workflows or not, et cetera, et cetera? Where is the technology capable?"
There may be other situations where you just stumble upon something because you need it and your friends need it, and so you just build it for yourself. That's kind of the traditional story of how Dropbox got started or at least what's sort of reported in the press. I wasn't around Dropbox really, so I just don't know. But that's kind of the story that I've heard about it. Sometimes you see people build the same thing over and over again for a company internally and then that turns into a company externally. I think that was the origins of PagerDuty for example. And then sometimes people just have a theoretical hypothesis and they're like, "Here's what I'm going to build and here's why" and it turns out to be correct. And so that was how we had Airtable, or Ivan at Notion, where they're like, "Here's what I think needs to exist in the world and I'm just going to go build the thing."
Turner Novak:
One concept that you have in your blog, probably one of my favorite blog posts you've ever written, is this, a concept of most companies started as a good product and then they switched to being distribution vehicles. Do you have any favorite examples of that throughout history of a good product building up distribution and then you almost layer on other products and it becomes the product is your distribution?
Elad Gil:
There's a ton of those. I mean Rippling is a great example of that today. Ashby to a lesser extent on the ATS side because they're bundling different ATS-related products and cross-selling them. Datadog did that really successfully. HubSpot is a great example of that where they're doing multiple sales and marketing tools for SMB.
If you actually go back to the '90s, that was much of tech. I mean that was Microsoft, right? So Office. All the Office products were actually things that were originally built on the Microsoft platform that Microsoft then either bought or out competed, right? Excel and Word and PowerPoint. These were all third party apps initially. So that's an example of a bundle. Cisco. Incredibly aggressive at cross-selling and cross bundling.
So tech used to be that way. And then there was this odd period of time where it stopped being that way. I think there's all sorts of potential reasons for that and I think it's back and vogue again. You could also argue Facebook has done that really well with WhatsApp and Instagram and the core Facebook app and everything else. When they bought Instagram, they started really accelerating it by cross promoting it on their platform or just taking some of the features that Snapchat has developed and cross-selling them effectively.
Turner Novak:
Yeah, that's probably one of my favorite examples, the Facebook one, where I think if you just look at Instagram's growth over time, there's a very clear it's growing like this and then when Facebook acquires it, literally the slope increases by three times of how vertical it goes.
And it's pretty simple, just cross promoting it in the Facebook app. I mean they have insane distribution. I don't know what time spent is, 30 minutes a day, 50 minutes a day, whatever it was at the time. And you just use that to grow more products. Pretty straightforward.
Elad Gil:
Yeah. Yeah, there's a lot of examples of that over time. And so I think that really is a traditional almost economical approach, which is, you now have market share, why would you not cross sell a bunch of stuff against that share? Particularly if it's to the same user or customer?
Turner Novak:
Yeah. How would you think about that?
If I'm a founder and maybe let's say I have a product where I don't know what the cutoff is, 10 million users, customers, 100 million users, customers, I don't know where this point is, but how would you dissect, "Okay, now is the time. It's probably the right time to start to think about it from this way instead of just being a product company"?
Elad Gil:
Yeah, usually I think there's two things to say.
One is most companies that innovate early keep innovating. The companies that innovate late actually never innovate again, right? Yeah, it's just a common pattern. So eBay or PayPal had amazing core businesses, but they never did anything with it again, right? They never really built out that much new stuff. And you can contrast that directly to something like Stripe where they innovated it early and now they have a dozen different SaaS style products and other things that they built just even on top of payments that connect and other things. I mean, they just launched meter billing and all this. They just keep innovating and I think they're actually accelerating now on that stuff. If you just watch what they did at Sessions the last two days, it's been very impressive.
Usually I think a good time to start to ask what's the next product, and it could be just a new set of features or bundle or whatever it is. I think it's usually when you have say, I'm making this up, for each company it's going to be different, load, tens of millions of revenue.
It's repeatable. So you're turning a crank. Every single thing is bespoke. You actually have repeatable sales and repeatable product motion and repeatable everything, and you have the organizational bandwidth to add the next thing.
Turner Novak:
So it's almost on autopilot?
Elad Gil:
I wouldn't say it's on autopilot. I would just say that it's reaching reasonable scale and it has strong repeatability. And usually the tension then is, how much do we invest in the core and we just accelerate the core versus starting new things? A new thing may just mean some new product thing that is part of the core but is a little bit to the site, right? It doesn't have to be a radically different product line. It's just making sure that you're thinking of things through this innovative lens where you're willing to expand the aperture a bit.
Often what you see is the companies that are trying to diversify their product lines really early, they have $500k of revenue or a few million of revenue. It's usually not always, but it's often because the founder just doesn't know if the core thing is actually really working and it's almost like a hedging strategy.
That's what I've observed. It's not always the case, but it's frequently the case. Because if the thing is working, you're going to be spending all your time and attention on getting it to $10 or $20 or $30 million in revenue, right?
But then once you're on this breakout traction at real scale, the question is, what do you start adding to it so that you can start cross selling? And again, cross-sell may just mean features or clear adjacencies. It's like, you're Vanta and you have SOC 2 and you're going to add HIPAA. That's just a cross-sell, but you're innovating, right?
Turner Novak:
And it generally works best when it's the same customer? You're just selling different products to almost the same sales channel?
Elad Gil:
It's much easier if it's the same customer, because otherwise, say you're selling into the sales team and then you have to go sell to the CFO.
Turner Novak:
Yeah, a different muscle.
Elad Gil:
It's just a different customer and dynamic. And the cross-sell is harder and it's a different product suite. So usually, I'm just going to cross-sell to the existing customer base, whatever that may be.
Turner Novak:
So I have a question just in terms of non-obvious things for startups to get right early on, you have a blog post about this. I think it's from probably 15 years ago. It's pretty old, but I thought it was interesting.
What are some things, just non-obvious things that startups should really try to nail early on?
Elad Gil:
I mean, I think again, the most important thing is product market fit. So just focusing on that and not doing anything else is really important.
You don't need a co-founder. Conventional wisdom as unique co-founder, I don't think that's true at all. And if you look at the biggest companies in the world, many of them were solo or in-charge founders. That's Amazon. And Microsoft was an unequal thing with Bill Gates really driving the company eventually. You just go through the biggest companies by market cap and it's almost always unequal with maybe Google being the exception. And even then Larry had a little bit more equity than Circuit. So you don't need co-founder. You need somebody in charge. So you need somebody who can make decisions versus consensus.
As the company starts to grow, it's okay not to delegate everything. People always say, "Oh, you have to delegate everything." That's false. It's perfectly fine to micromanage certain areas. You just have to choose the right ones.
I think one piece of contrarian advice now maybe is, I've seen a lot of companies that take enormous pride in being very, very lean, and I think capital efficiency is very important, but I think there's almost been this flip into, "I'm not going to hire and go after this opportunity," or "Hey, isn't it great how small we are when you see companies just miss things because of that?"
That kind of happened to Sketch, I don't know if you remember them, the design tool, and it was like, I don't remember what it was, 40, 50, 60 people, whatever it was in the Netherlands. They were the tool that every designer was using. They just were like, "We don't want to grow. We don't want to be aggressive." I don't know if they still exist as a company. They might, but I think they really missed the opportunity because they were so focused on being small and lean, and "Look at how bespoke we are" versus, "I want to win the market."
That actually supposedly was part of the backstory on Microsoft in the OS world. I think there was an OS company.
Turner Novak:
CPM?
Elad Gil:
I don't remember the name of it. Yeah, it was like a small team and they were kind of happy to be small. And then Bill Gates shows up and I think even licensed it from them. That was the original OS. He sold IBM before they built their own, right?
So I just think there are these examples like that.
Turner Novak:
This one that I thought was super interesting, I just wanted to make sure we get in. Giving your family stock in your company.
So the way you phrase it is, you might be super successful, make a bunch of money, you want to give money to your parents, your close family members, but that's a lot harder to do at the end than the beginning. Is that another thing you'd recommend founders to do?
Elad Gil:
Yeah, I think it definitely is worthwhile if you want to set aside or grant some equity and you could still have the votes on it or control it in different ways, but give it to different family members early and a small amount of stock may end up being incredibly valuable. And if you gift it to them at founding for example, it's worth a few cents. There's no tax event and it just falls under natural exemptions, but that could become quite meaningful five or 10 years later.
And so it's a nice way to spread the potential success out to people who may benefit from it that you care about.
Turner Novak:
I like that idea. I really had not seen or heard much about that, but it makes ton of sense. I have one founder, he gave some to his parents and his girlfriend, now wife, and thought it was an interesting strategy.
Another thing that I've kind of heard you mention, and you kind of touched on this a little bit earlier with companies that innovate or don't innovate usually persists. In terms of sticking with an idea that's not working, how would you recommend a founder, let's say they're working on something, not working, middling traction, is it okay to give up? How do you think about that?
Elad Gil:
Yeah, I think there's three forms of giving up.
And I would say that the advice to keep going no matter what usually is actually bad advice in my opinion. As long as again, you're seeing reality and you're actually seeing what's really working or not. Because sometimes things are working and you don't think they are, and so obviously you shouldn't give up then.
Or you sell the company too early or whatever it may be, right? But I think often, at least in recent times, it's been the opposite where people keep going no matter what and there's money there because capital's been so abundant and so people just keep going and going. And suddenly you look back and seven years of your life are gone and you don't have that much to show for it. And those are the most, at least, potentially risk seeking years that you can have. Those may be your best years to do things that are big and take big swings.
And so you really want to optimize for making the most of that time that you have because it goes away. You have kids, your family members get sick, whatever it is. And so life tends to constrain you more and more overtime for many people. Not all obviously. There's people who are constrained early on. But for many people, that's the pattern.
And so taking advantage of that time is really important. And so that means sometimes you just have to decide or realize, "Hey, this thing is not going to work and I need to switch what I'm doing." Sometimes that means switching what you're doing within a local area. Most pivots that happen pivot within a market versus across markets. But in some cases it's like, "Let me just rethink the whole thing. Should I sell the company now and start another one? Should I shut it down? What should I do?" People undervalue exiting now for some reason.
Turner Novak:
They undervalue exiting?
Elad Gil:
Yeah. Now when people decide to give up, they're like, "I'm just going to shut it down" versus saying, "Okay, I'm going to land it somewhere and then go do something else."
I think actually landing the company somewhere is important. And it's important for your employees. It's important just as a sign that you care about wrapping things up in a certain way. But I also think it helps you as a founder for your second time of starting a company because you're now perceived as somebody successful, even if it's an acqui-hire, if you're like, "Hey, I sold my company to Stripe," people are like, "Oh wow, you sold your first company to Stripe, so your second one's likely to be successful too.'
Turner Novak:
There's just some bias in that, yeah.
Elad Gil:
Yeah. And it makes hiring easy. It makes fundraising in some cases easier. So I just think it's one of those things that's almost become undervalued in some ways, even if it's a small asset.
Turner Novak:
So you mentioned it makes fundraising easier. Any specific advice you'd give of founders in terms of fundraising, some non-intuitive things that you've kind of picked up on?
Let's say I'm raising my very first round. We'll say I have a product, I have a thesis, I have a market. How would you recommend a founder just approach fundraising for the very first round?
Elad Gil:
I think I would kind of view it through the lens of who do I want to be working with for the next decade, because some of these are such long-term relationships, right?
If co-founder is your work spouse or whatever, then your investors are like your in-laws. Do you want to see them at Thanksgiving? Do you want to spend the holidays with them? You're going to see them all the time. They're going to show up for the kids' birthdays. If you don't like your in-laws, it's kind of miserable, right? And that's kind of like your investors.
And so often I see people rush the first round fundraise process, like, "I'm just going to get done in a week and whoever gives me money..." And you're like, "Well, you're going to be stuck with that person for 10 years. Hopefully you're optimizing for somebody who's both useful, but that you also want to see because otherwise it may be a painful couple years."
So I kind of view it through that lens. Who do you want to work with for the next decade if you can choose? If you can't choose, then you go with where there's money pragmatically. But if you can choose, you may want to take a little bit of time and choose.
Turner Novak:
Yeah. How would you recommend a board member? Just if somebody was like, "Hey, I want to add someone to my board," what advice would you give them to think through that?
Elad Gil:
Again, it comes down to how many options you have and is it an investor board seat or an independent board seat?
If it's an investor board seat, sometimes you don't have as many options. If it's the only investor making you an offer, you take it, right? Or you figure out some alternative.
But either way, I would suggest writing up a job rec, because it's kind of weird. If you look at every other thing in your company, you'll write up a job rec. And then you'll hire against that rec. But for somebody with the board members, you just kind wing it, right? And that's the person you have to talk to all the time. And you can't get rid of if it's an investor board seat. It's hard to get rid of.
And so I would just write up a job rec. What do I actually want from the person? What are the skill sets, the background, the attitude, the perspective, the things that they can help with, how they actually help out with that? Let me background or reference check them. So I kind of treat it like a job almost, like I'm hiring somebody in that I couldn't hire otherwise.
Reed Hoffman always puts it as like, "The optimal board member is somebody that you would've wanted as a co-founder, but you never would've been able to get him." That's his view of what a good board member is. I think that's a very interesting framing of it.
Turner Novak:
Yeah, I kind of see it... I mean, I'm only on one board, but I basically see it as, I don't do anything. That's my default, is I don't do anything, except I'm very responsive whenever you need me to. And I'll help with anything. There's a lot of things I'm not good at, a few things I feel like maybe I can help with. That's interesting framing, the co-founder you couldn't get.
In terms of hiring, so let's say I raised my first round, I got $2 million in the bank, whatever the number is, we'll say it's a reasonable number. I didn't raise too much, I didn't raise too little. I need to hire my first three to five employees, whatever the number is, how would you recommend somebody to go out and do that?
Elad Gil:
Yeah, I think the first employee for some reason is often the hardest to convert unless they're already in the queue. Do you have past relationship like you work together really closely at something?
Once you bring on the first person, I feel like it starts to roll a bit more. I think very tactically, you obviously want to check for a mixture of skill set, culture fit, drive, ability to get things done. There's kind of a series of attributes that I think are important in early stage companies. And so usually you end up constructing an interview process against that.
I think the other mistake people make, particularly when they're hiring ancillary roles... Or ancillary is the wrong word, there's core engineering. And then usually early on you bring on say a designer and you bring on say a chief of staff type or somebody that's a business generalist, and then sometimes you'll hire an early sales or BD person. That's kind of the main components of an early team.
What people do is say that they haven't worked with a machine learning person before, an AI person, they need to hire them, they'll suddenly cast this very broad-based net or a designer. They're like, "I'm going to talk to every type of designer from every type of company" instead of just saying, "What are the two or three companies that do the exact type of design that you really love?" and just talk to every designer out of those three companies. That's going to give you a much better, more targeted search.
And so I think part of your search function should really be geared towards specific companies for specific roles versus, "Hey, it's just the role, so let me just interview designers from anywhere." And so that's one place where I think people tend to go wrong in terms of early hiring.
And honestly, late hiring. You do the same thing when you're hiring executives. "Where are the best sales team, that sale and the customer type that I like or that I'm looking for the type of product that I have? I need mid-market sales for developer products. Oh, there's like three companies I should recruit out of because I know they do it very well."
Turner Novak:
Very straightforward. It makes a ton of sense when you talk through it like that.
I feel like you mentioned one time the issue of having too many bureaucrats. I don't know if that's the right wording, but how do you avoid hiring people that don't actually do work? What would you advise for that?
Elad Gil:
It's a good question. It depends on the level that you're hiring at and what is considered the nature of work. But in general, you want to, in startups, hire people who have scrappiness to them. Irrespective of role and stage.
Part of that is what sort of exercise or activities are they willing to engage in as part of the interview process even? And part of that is just asking questions about like, what do they do today? How do they spend their time? What do they focus on? What are the actual activities that they do versus people who work for them or with them? And so just digging into the specifics, I think, and some articulation of that. And then of course, reference checks or back channels make a huge difference on that kind of thing.
Turner Novak:
How would you recommend doing a reference check just for somebody who's never done that before?
Elad Gil:
It depends of whether it's a function that you've hired for before or not.
If it's a function you've hired before and that you understand well, then you kind of know the questions to ask and then you have a series of questions to go through. And those are usually a mix of functional expertise, speed, output, diligence in their work, collaboration with others, et cetera, right? You tend to have your own list.
If it's a function you've never hired for before, then usually what you do is you go and talk to three or four people who are very good at that function. And you both see what they think is important, but then you also ask them like, "What questions do you ask when you interview and what questions would you ask for a reference check?" I always just ask people who are experts at it like, "What do you care about? And usually that converges pretty quickly." If you've never hired a CFO before, go talk to five really good CFOs and see what they would look for.
Turner Novak:
Yeah, makes sense.
Elad Gil:
Versus just making it up, you know?
Turner Novak:
Yeah, that's the easiest way. Just make it up.
So I've heard you mention before that ship velocity, just speed is one of three key attributes for any early stage company. I mean, really any company really at the end of the day. What would you say are kind of the three, and then maybe good examples of that?
Elad Gil:
Yeah, I mean, everything is in the service of product market fit, right? Ship velocity is really important.
Some form of salesmanship or capability is very important. Salesmanship may mean convincing people to join the team. It may mean raising money. It may mean getting it really customers, but you need the ability to sell.
And related to that, you need some form of clarity of thought, roadmap, direction, communication, et cetera. And usually that really helps with ability to sell.
Turner Novak:
So it's speed, ability to sell, and clarity on the customer or the mission?
Elad Gil:
Yeah, clarity of thought and action and direction. Yeah.
Turner Novak:
Makes sense. And I think you mentioned Perplexity earlier. You led their first round. I think I saw you tweet that that was one of the things that really stood out to you when you invested, was just speed of execution.
Elad Gil:
They had extremely high shipping velocity, yeah. They're very good at execution.
Turner Novak:
Why do you think they're so good at execution? What makes them so good?
Elad Gil:
Aravind, the CEO, is just exceptional at that. And then he co-founded with a bunch of people who similarly had the same outlook of velocity as being important.
In my first startup, we used to say that, if it took as long to debate something as to build it and try it, you should just build it and try it because you end up in these multi-day debates. And "What if we do this and what if..." And you're like, "Fuck it. Just take two hours and ship the feature and let's see." Or a day, whatever the timeframe is. But often that's the best way to solve for something, is just go do it.
Turner Novak:
Yeah, because something might not even work. And if you spend a month planning it out, it took you a couple days to do it.
Elad Gil:
Yeah, planning and debating and "Should we really do this?" and "Let's have another meeting." It's just, you end up wasting a lot of time so you either want to kill stuff fast or try them really fast, right?
If you think about it, the only advantage an early stage company has against an incumbent, there's probably two advantages. The biggest advantage is speed. The incumbent is going to have more resources, it's going to have more customers, it's going to have more people, it's going to have more everything. The only advantage you have is speed.
The other advantage you may have is business model. You can change the business model significantly enough that it destroys them to keep up. Those are your two advantages. What else do you have?
Turner Novak:
Yeah, that's fair.
Elad Gil:
That's it.
Turner Novak:
I mean, bringing it back to the AI discussion, I always think it's fascinating that basically all this technology that's coming out in AI, it was basically created by Google and they had it all. They've maybe fumbled a little bit, if that'd be a fair way to say it. What's kind of been your observation about how that's played out over the last couple of years?
Elad Gil:
I've been really impressed by how quickly, once Google decided to go after this, they've gone after it. So I think it's been nine months or something, or maybe a little bit longer than that now since they decided to really focus on Gemini and Bard and all these other things. And Gemini 1.5 I think was a really interesting step in capabilities for them. I think the long context window was super interesting.
So I think to your point, they have all the assets and capabilities. And once they've directed them, they've actually made really fast progress.
Turner Novak:
Yeah, I've been surprised or I've noticed, I wrote a blog post probably about two years ago. It was before ChatGPT. Just the Google homepage, all the things they could do with that. If you open up Chrome on your phone, Chrome mobile, there's kind of this little For-you page of different articles you might want to click on.
I always thought they had so much surface area and so much distribution really to utilize a lot of this technology that's coming out. So I'm interested in seeing how they take advantage of it. I think that I'll be interesting.
Elad Gil:
I guess like at a very high level, we're going through a generational shift in technology right now due to the AI world. And part of it is because we've built these really performant models with new capabilities.
Part of it, which I think is under discussed, is the fact that you can do anything through an API now. So you just ping an API and instead of dealing with MLOps and set up and training, you just have a general purpose model that you can ping. And so that's created this explosion of new things to do around AI and technology.
And on my side, there's a series of incubations, funds, operating companies, other things for the technology future that I'm working on. And so there's this big surface area of stuff, incubating, companies like Braintrust leading big rounds, and companies like Applied Intuition, Harvey, Anduril and others leading first rounds in companies like Perplexity. And maybe doing AI driven buyouts. So there's just so much that I've been doing.
And so I'm starting to add slowly to my team. And so that includes, I guess two types of people. One is builders, former CEOs, COOs, product people and engineers. And there'll be a small number of people, but I really want to add some people on that side.
And then I'm also adding one or two people with more traditional investor backgrounds. with a bias towards people actually with private equity or related backgrounds. So Blackstone, Silver Lake, Warburg, MS, things like that. In part because I want to have this complementary skill set, but in part I want to take a builder-centric approach towards engaging with this broader big shift in ecosystem.
And so my hope is eventually to have people who've started companies, product managers and engineers who built really interesting things, and then people from the investment world who can help compliment and supplement that.
So that's kind of how I'm thinking about shaping things up over time.
Turner Novak:
Okay. So it's kind of like your own take on like a platform team almost maybe, but just less generic?
Elad Gil:
It's not really a platform team, it's just people who are going to go build and invest in things. And so we do very technical work in terms of going and trying out different AI infrastructure platforms and benchmarking them, just performance benchmarking or trying different tools or building different things.
Like two people on my team, and it's like a three person team so it's not a lot of people, went in and built an embedding playground where you could swap in and out any vector DB and then any embedding framework and upload data session, just look at performance. And that way you can really tell, "Oh, what's the difference between the vector DBs and what's the best one?" And so it's a lot of very hands-on kind of stuff on that side.
But then sometimes you really want to also say, "Okay, what do the cohorts look like and what do the numbers look like and how do we think this thing will evolve?", the more commercial aspects of what's going on as well.
Turner Novak:
So if I think this sounds interesting and I'm listening to this, how does one get a job working for you? Do I need to message you? Reach out? Are you posting a job application somewhere?
Elad Gil:
Yeah, it's usually reaching out through somebody we both know is the best way to do it.
Turner Novak:
That's fair. That's usually a good way to get a job somewhere. I have actually one more question. So you've gotten really into gardening it seems, but you've also trimmed down the types of gardening that you do. I was curious. I mean, I think you've actually described it as a "Darwinian death match with nature and a battle for scarce resources."
Do you do a lot of gardening? Do you like gardening?
Elad Gil:
No. I started doing a little bit of gardening during Covid. We were trapped at home and it was something to do and the kids got to participate and plant things and see them grow and eat things and all that kind of stuff.
And then last year I kind of overdid it and I think I planted a dozen different things. I think Hobs has this quote from Leviathan on the State of Nature. He said something like, "Life is solitary, poor, nasty, brutish, and short and gardening really drives that home."
Turner Novak:
That's very grim.
Elad Gil:
Yeah, I know. Gardening is a grim experience.
Turner Novak:
Yeah, we have one. Or we had one kind of, basically planted, I don't know, a lot of stuff. And we had some groundhogs that were kind of... they had a bunch of kids.
It was basically, once everything started to fruit and flower, just the entire garden was just ravaged by the groundhogs one night. And it was just all gone. And it was probably a full two weekends of building it all out, planting everything. That was rough. That was very sad. So we're taking a different approach this summer.
Elad Gil:
Yeah, it's the state of nature. It's the state of nature. Everybody forgets.
Turner Novak:
Well, this is awesome. Thanks for taking the time to come on.
Elad Gil:
Yeah, thanks for having me.
Stream the full episode on Apple, Spotify, or YouTube.
Find transcripts of all other episodes here.