Full Transcript for Vineet Edupuganti (Cogent Security) - First Principles in the AI Era

This is the Entrepreneurial Thought Leaders series, brought to you by Stanford eCorner and the Stanford Technology Ventures Program. Welcome to the Entrepreneurial Thought Leaders Seminar at Stanford, the seminar for aspiring entrepreneurs brought to you by STVP, the Entrepreneurship Center and the Engineering School at Stanford, and MS &E, the Management Science and Engineering Department at Stanford.

You know, it is a real treat whenever we can have a speaker here who is an alum. It's a full circle moment. Those are always the most special ETLs that we have, and that's what we get this week. Before I formally introduce Vineet, who's going to be spotlighting our ETL this week, let me explain the intention, which is really, I think what I want you to get from Vineet is how to be a successful founder even before you start your company.

I think that's going to be the lesson that we're going to try to impart in the next 50 minutes. Vineet is a Stanford alum. He's a recent grad, which we'll talk about in a moment. He grew up in Portland. He comes from a family of generations of entrepreneurs. He played competitive tennis also in high school. How many people played tennis in high school? Okay, awesome.

He came to Stanford. The air got a little bit more rarefied, and he excelled in certain things, but he couldn't make other things, so he wasn't on tennis at Stanford. But he did get a bachelor's in EE at Stanford. He was the president of the Venture Capital Club at Stanford. Any people who are pursuing EE or thinking about pursuing EE? All right.

And then Vineet went on to get a master's in computer science. Any CS majors that people are being pursuing? Okay, awesome. From there, he joined Abnormal Security for four years. That we'll talk about. That was his first work experience post -Stanford. And then after Abnormal, he's now become the CEO of Cogent Security, which is really using AI to understand vulnerabilities and address security vulnerabilities, which is a huge, obviously, emerging space at the intersection of AI and security.

Cogent's gone on to raise over $50 million from some of the best funds in Silicon Valley, from Greylock and Bain Capital. And so we're incredibly excited to welcome Vineet to ETL. And if everybody's in the back, please have a seat. But please join me in welcoming Vineet to ETL. We're going to do a fireside chat. Vineet.

We got your mic on? Yes. Okay, we're awesome. We're awesome. We're good to go. Well, Vineet, welcome back to the farm. Thank you for having me. And Vineet took ETL, gang. I did. I did. I took this class probably seven or eight years ago. So it's great to be back and thank you for having me. And so, you know, literally, it wasn't that too long ago that, Vineet, you were sitting in one of these seats.

Correct. You now have raised $53 million of venture capital and you've had a bunch of experiences. I'm curious just to kick things off. What advice you would give your younger Vineet from who, if you were to talk to the Vineet that was at Stanford at this time, what advice would you give them to do differently? I think there's a couple things.

I think, number one, there's just so much that's always happening at Stanford, right? You have your peers. Everybody's chasing after something. They're interning at some startup. They're in some random club. You can get stuck very often in this rat race. I think what I've learned over the years, both at Stanford and even beyond, is you just have to have tunnel vision, right? There's just so much chaos that's always happening.

And especially now with AI and all the hype around AI. So that's one thing. I think just like focusing, picking your path and then really going deep on that thing. We can talk a little bit about my journey with that too. Normal and now cogent and really writing the right exponentials. I think the second thing, and this is probably going to get me barred from campus.

I came into Stanford very much with an academic kind of orientation, right? Going to all these classes and grinding very hard on problem sets. And that was terrific and that's important. And so continue doing that. But if I were to look back and think about what the most critical aspect was of my several years at Stanford, it's actually the relationships, right? And those folks were my bosses, my investors, my peers, my direct reports now, the folks that I've hired in.

And so I think that is the thing that has compounded tremendously. And I significantly underestimated that actually when I was at Stanford, right? It was like, everybody's just like, you know, a normal person. They're your friends. And then five years later, everybody's doing incredible things. So, you know, I don't know if that's the proper takeaway for this group, but I'd say, you know, don't lose sight of the other aspects of university outside of classes.

Make sure you're enjoying and participating in the upside of everything else Stanford has to offer. And I think people is absolutely at the center of that. It is so amazing. And I think we totally under indexed for it because you understand it objectively how valuable it can be. And it's so fascinating to see if Anita has a case study of how that's played out.

I want to sort of talk about then how you decided to do what you did before you started coaching. So one thing that I think students wrestle with is when is the right time to be a founder? You know, in two weeks, we're going to have Karina. She stopped out of Stanford and she raised a quarter billion dollars and she's starting a company. So some people start their companies immediately after their startup journeys.

Some of the students, you know, may want to do Teal Fellows and drop out or join Y Combinator after Stanford. You chose to work for four years before being a founder. Can you walk us through how you made that decision? And in hindsight, do you think that was the right decision? And what advice do you have for founders that are wondering what experiences they should get before or students on what experiences they should get before they become founders? Yeah, absolutely.

And I have many weaknesses, but I think one strength is self -awareness. And one of the best pieces of advice I received actually when I was at Stanford, because I was always thinking that, hey, being a founder could be interesting. You have that orientation that comes from everything from your family background, from I think like it's in the Stanford era a little bit as well.

So I was always thinking that that could be an interesting path for me. But I was actually speaking to a VC and asked him that question. I said, hey, when is it the right time for me to start this company? And I saw a lot of my friends going down that path and, you know, trying to do things of their own. He said, it's when you would actually invest in yourself, right? Imagine you cloned yourself and that person was running the company.

Would you invest a ton, maybe all of your own capital in that person? And I was like, that was very insightful. Because the answer was obviously not. I have no skills. You know, I would never work for the clone of myself, right? And so I think for me that was the trigger point. I just want to go and I want to see excellence. I want to absorb that for a couple of years, learn the ropes, and then go do it and actually be successful.

And I think sometimes people can be a little bit too short -sighted, right? Of just wanting to do the next locally optimal thing. I want to be a founder. So I'm going to go start a company tomorrow. In reality, by working backwards from where you want to go, taking the right intermediate steps, you can actually get further along. And it's as if you're actually parallelizing and making progress ahead of time, right? And so I came into the progress of starting Cogent and that whole journey actually feeling I was much further ahead than just starting from a blank slate.

So I want to double -click into that. I want to dive a little bit deeper because I really want the students to grok this. About when do you know that you can predict that you're going to actually have the best chances of success as a founder? And what I was struck when we talked earlier was this intentionality that you took before you even incorporated Cogent.

Can you share in as much detail as possible about how you built conviction before you actually even decided to officially start the company? Absolutely. I think one thing I advise to folks, I mean, there's one school of thought, which is like, you know, start a company for the sake of starting a company and, you know, you'll figure it out.

And I have seen that work, actually, in some cases. I think the better approach and the philosophy that I ascribe to more is actually solving a problem that you've experienced yourself or are familiar with. So for me, I had spent four years in the cybersecurity field. I talked to hundreds, maybe even thousands of security practitioners. I just saw how broken everything was, right? You have a lot of people running the same processes and doing the same things they were doing 20 years ago.

At the same time, every hacker can just feed prompts into LLMs and start exploiting vulnerabilities and launching social engineering attacks at scale and much faster than ever before. So clearly this is not going to work, right? Being attacked by armies of bots and humans are running around in spreadsheets and doing the same things they were doing in 2003.

Thought Leaders, and I think that's a great idea. Priorities and role levels, aggregated those data points, and the intention was really to become encyclopedic about the market, right? Where I wanted to know, for the people who have tried this type of thing in the past, why did those companies fail? For the chief information security officer, who potentially was going to be the person adopting our product, what was it that they were looking for? How do they evaluate products like ours? How much would they be willing to pay? What do we need to show them in a pilot or proof of concept to earn their trust and earn their business? So those were all really critical aspects.

Did all that work, and then we ultimately had so much confidence. We're like, hey, we know who's going to be our first set of customers in parallel, who's going to be the first set of recruits, who's going to go join us. Very clear what the market is, very clear what the strategy is. All right, let's go raise money and get after it. So I want that checklist just to become clear.

And by the way, gang, if you're in the back, there's seats. Try to come in if you can and fill out the seats that are on the sides. But just so that checklist is clear, it sounds like what you wanted to, what you did was you understood why the competition failed. You had that clear in your mind. You also had customers that, I don't know, did you have LOIs pre -signed? Or how did you know that you would have definitive customers? Not pre -signed, but verbal.

So you had people that said that if you did with you. They said, hey, if this thing actually works, then I'd be interested in working with you. Right? And that was the risk that we wanted to take on. Right? Execution risk of can we go and build it? And the thought was, hey, we're going to go raise capital. We're going to assemble the world's best engineering team in cybersecurity.

We're going to go do this. But the litmus test with the customers is you ask them, would you actually put money down against this vision? Exactly. Exactly. And how much? Right? How much? And for what capabilities? And really understanding all those nuances. Okay. So getting it very specific in terms of what they would want for a first initial engagement.

Correct. Okay. And then you also got talent secured, it sounded like, too, before you even started the company. Yes. And that's obviously soft because people might still be locked away in their jobs. And when push comes to shove, there's a little bit of, you know, additional convincing you have to do. But that's just as important. I think one learning from the company journey so far, I actually thought at the beginning fundraising was going to be the hardest aspect.

That turned out to be the easiest aspect. Recruiting has been the hardest aspect. Yeah. Especially now. Especially now. So what were you looking for to check off that box on talent? Was it just a soft circle yes that I would join your company? Or was it something more definitive? That's all it was. Okay. Just kind of a soft yes. And by the way, some people that give us a soft yes ended up backing up because things changed for them and all that.

Yeah. And so that's another learning that you always have to have backup options and insurance and more pipeline than you think you need. But that was what we looked for, basically. And then how do you communicate that also to investors when they're not joining your team, but they're about to join your team, but you don't want to out them because they're working somewhere else? Is that, was there a nuance there? It wasn't too much of a consideration.

Okay. And I think they respected the confidentiality. Yeah. Of course. Okay. I want to move into thinking about first principles thinking, because I think it's a buzzword that we all hear about, but it's rare to actually see it in practice. And I know you're a fan of first principles thinking. Can you explain what first principles thinking means to you? And can you give us some specifics of how that actually manifested in Cogent and how you benefited from first principles thinking? Absolutely.

And I think from a definition perspective, whatever Jeb and I are chat to be, it's probably a higher quality definition. But I think the framework that I have in my mind is you kind of want to start when you're tackling a set of problems with either the facts or the assumptions that are just incontrovertibly true. Right. And then you build up from there.

And so I'll walk you through that in the context of Cogent, which was actually some of the things I was mentioning earlier. AI is exploding. Software is exploding. Software has a lot of vulnerabilities. Therefore, there's going to be a huge surge in vulnerabilities. Hackers can use AI to generate exploit code to take advantage of those vulnerabilities way faster than before.

And so therefore, unless you have good AI on the defensive side that can fight off the bad AI that's going to be coming after you, we're going to be in a state of near constant breach. Right. And so that was the thought process. And then there were other aspects. Right. So actually, when we started the company, many folks did not want fully autonomous AI defense.

And it can break things. It can go rogue. There's, you know, potential issues that could result. And so we actually faced a lot of pushback. And so part of what we had to do was figure out how to meet people where they were at that point in time, but be very dogmatic and clear about what the end state was and what the vision was. And that all played out actually in a really powerful way earlier this year.

Yeah. Two years ago, cyber was like, you know, this niche thing. You just have to do it. And now obviously it's at the forefront with mythos coming out right in the April timeframe, the hugging face hack a couple months ago to now even the stodgiest of Fortune 500 enterprises, biggest energy companies in the world as an example, are coming to us saying, hey, we're going to have fully autonomous AI on defense.

Right. Right. And so it took some time for that maturation to happen. It took some time for us to evangelize. But now we're kind of in the right place at the right time. Yeah. I think I just being in that magic spot when the market opens, it's too late now to almost even try to compete if you act reactively instead of with first principles.

So I love that. By the way, I'm curious, just, you know, everybody's talking right now about the existential threat of AI. You're living in this world between AI and security. You know, the anthropic researchers and Daniel Amodez famously said that he thinks there's a 10 to 25 percent chance that humanity will die, will become extinct from AI.

I'm curious where you stand on that spectrum. How, you know, given everything that you know, how worried are you actually that AI poses an existential threat to humanity?

Nonzero. I wouldn't go 10 to 25 percent. So I'd say nonzero but small. And my model around that is, I think, to end humanity. That's a very extreme claim, by the way. And I'll talk a little bit about what I think is more likely to happen. There's just a lot of things that have to happen in the probability chain, right? You have to have rogue AI that does something and can't be stopped and then does another thing.

And safeguards don't work. And, you know, by the time all that multiplies, you know, it's just the probability that whole sequence of events being true is going to be pretty low. That could change. We're continuing to learn more about AI's capabilities and what's happening in the wild. The thing that actually terrifies me is AI in the hands of bad people.

Right. And I think that's what you're seeing. And the consequences can be crazy. Right. Power can get shut off for entire countries. Right. Financial futures of people can be erased. And so that's the part that I think is in the here and now, because what AI has done is it's raised both the floor and the ceiling for any hacker. So some person in their parents basement can now operate like a very sophisticated nation state.

Very sophisticated nation states can scale the things that they're doing across basically the whole world. And that's again why there's so much urgency. People are talking about regulation and some of those different aspects. And, you know, there's a lot of dimensions to that conversation. The part that I'm not hearing enough people talk about is adoption of defensive technologies that enable you to fight back.

And I think that's the only way in the future. Okay. Okay. I want to move into one area that I know Cogent is sort of making a name for itself around, which is sort of attracting founder level talent as employees. And before I get to the founder Academy, which I know you guys have been pushing and creating, I'd be curious about your own co -founder journey, especially for advice for solo founders that are thinking about finding the right co -founders.

Can you talk about how you found your co -founder, how you knew they were, you guys were the right team or the right fit and any advice for people that are thinking about finding co -founders? What are the most important lessons to have successful co -founder partnerships? Definitely. The nice thing about the co -founders that I have now is that we worked together in the past.

And so I think having that level of familiarity and that level of trust in the workplace was a huge advantage.

It's not true that you want to start a company with the people that you're social best friends. In fact, that can be a very bad idea because now, you know, the personal side of things can actually get in the way of making hard decisions and getting into things from a work context. So I actually like to keep those things a little bit separate, work and personal.

Are you a fan of living together before you actually start a company so that you can work out the kinks? No. No, okay. You know, that's my take. I mean, obviously there could be other perspectives, but my view is keep the personal side and the work side separate. Okay. And I think like, you know, the friendships and the social bonds, those evolve from very close work collaboration.

And I think that's what's kind of happened with me and my group. And it was great that we had worked together and we experienced collaborating and solving hard problems in that context. And then I think the other thing that is really important when coming up with your co -founder pairings is making sure that you have complementary skill sets that span the set of things that need to be done.

So I think oftentimes I see these like copycat profiles, right? Where it's like two people that are like effectively twins on paper and starting a company. And it's like, you know, maybe two or three really business -minded people. Nobody can write code, right? Or build systems and understand architecture. Not ideal, right? Or it's like three people that are super technical in the weeds, but nobody knows how to sell, right? So I think like having, you know, a co -founding group where we're covering all the capabilities.

And ideally by default, there's zero stepping on people's toes. Everybody just kind of has that instinctive set of here's what I'm good at and here's what I should be responsible for. That's probably the ideal setup. And I think that's what we have. Yeah, okay. And how many people are you for a co -founder? How many co -founders? We have three in total.

So myself and two others. Is there an ideal architecture, do you think of that? Or is it just mainly just separating out roles so that people aren't stepping on each other's toes and you can move quickly? There's different configurations that can work. You know, in our case, it's a VP engineering, runs the engineering team, handles all the people, process, aspects of things, builds the org.

We have a chief technology officer, technical strategy, making sure the right things are getting built. And then I started off handling a lot of things, you know, product and go to market oriented. So it's one configuration that can work. Sometimes you have a couple people on the other side. Okay. And then I know you have championed something called the Founders Academy at Cogent, which is sort of unusual because you're trying to attract talent that you're deliberately trying to encourage to start something and then leave Cogent.

Can you speak about that? Or what's the rationale behind that? And I wouldn't say that we're actively encouraging them to leave Cogent, but I think organically, they will leave Cogent at some point in time to go start companies. And why do we do this? I think for us, having founders that work at Cogent is a really deliberate part of the talent strategy.

And it's because founders and aspiring founders tend to have a lot of very positive reinforcing characteristics, work ethic, ambition, growth mindset. Typically, they're optimistic because if you're not optimistic, you probably don't want to start a company. So those are all traits that are obviously very advantageous for us in terms of being able to get work done and move faster.

And then also as an employer, I think that's something that's differentiating for us and also just very important to me as a person is creating these environments where people can do their best work and then have this launching pad that takes them to the next point in their career. Because the reality is, we all think in these one to three year increments, careers are 40 years long, right? And as life spans extend, maybe careers will extend.

And so it's not just about the current job that you're in. It's about thinking about your next job and your next, next job, right? And we want to make sure that in like 10, 20 years, people look back and they say, hey, Cogent was really that launching pad for me and the thing that inflected the shape of my career trajectory. Okay, awesome.

I'm going to move to doing something that we're trying, this is going to be the first time we're doing it, which is signature questions. So the class has put together a prioritized list of nine questions that we want to sort of rapid fire through. I mean, ideally like spend like a minute or two on each question. But these are sort of our bold questions for the class.

So I want to kick off with if you were 20 years old and starting again today, and I know you've already talked a little bit about some advice, but what would you spend the next five years learning or building?

I think two archetypes I'm starting to see emerge right now is there's one, which is the been there, done that. They know what excellence looks like. They've been in the industry for decades. And then there's just like the super AI pilled, you know, fast mental horsepower types of folks, right? That are just like very in tune to how to like work in this new era that we're in.

I would say making yourself kind of mold into that category is probably very impactful. So building a lot of things, being up to speed with the tools,

you're trying to deliver as much impact as possible per unit of time, which basically you as an individual would be the bound on that. That's probably the number one piece of advice that I'd have. Okay. Okay. So everybody build right now and have impact. What do you be, well, this is a good segue. What do you believe about building companies today that you wish you'd understood when you were sitting with where these students are? I stumbled into some of these insights, but I think like, you know, hopefully this group can take my learnings and employ them.

There's a couple of things that just really matter, right? Like market. You pick the wrong problem to work on. You could be the best entrepreneur ever. You could be the best person at executing. There's probably no hope, right? Because ultimately you're bounded by just what that market will go and support. Similarly, your team, right? You could be working on an interesting problem.

If you don't have the right people around you, it's going to be really tough to make forward progress. A lot of the other stuff is fixable. And so I think if you were to think about the company journey and what makes companies successful, getting those couple core things right is just so, so crucial. And I wish I kind of like internalized that a little bit more ahead of time.

Okay. Well, thank you. I think this is the full circle moment, so we appreciate it. What's the biggest professional failure you've had and what did you learn from it? So many failures.

It's hard to pick one. I think one theme though,

one theme and one learning. I think sometimes I'm a very individual person and an independent person. And so that can mean trying to run solo on a lot of things and do everything myself, right? I think something that I'm trying to get better at is being a little bit more vulnerable, right? And so leaning on mentors, right? They can be incredibly helpful.

There's a bunch of people who have been through similar scenarios in the past. Being better at delegation, right? Working with the team. So a million different failures associated with that, but that's part of the broader umbrella. And this is a similar question, but what's the biggest mistake you've ever made?

Biggest mistake I've ever made. Biggest mistake. There's a lot of hiring mistakes. I won't go to the details on those ones. And what's the general lesson on hiring? Just for people who are going to go down that path. You know, I think there's like a saying. I don't know who actually put it out there. It's like, if you're not a hell yes, it's a no.

I think that's what I'm coming around to. Okay. And so we've tried to keep the talent part incredibly high and just have full conviction on everybody that we bring on board. And can we make that a little bit more concrete? You don't have to mention names, but can you walk us through your thinking and making that decision or that moment when you hired somebody and when you realized it was a mistake? So I think like oftentimes

the place where a lot of us can succumb is doing the thing that's easy versus the thing that's right. Yeah. And so for many folks, including myself, we've raised capital, right? And even in my last job, right? We're a very high growth startup. There's a lot of pressure to fill roles, right? And so if you don't fill roles, binary people can look from the outside and say, hey, you haven't hit your metrics on hiring or this or that.

And so as a function of that, you know, you're like, okay, I want to check the box, right? And like, maybe it'll work out. Maybe it won't. And the problem then is like you get these people in and then, you know, maybe their performance is not what it should be. And it weighs on other people around the organization. And then there's additional folks that they may hire in themselves if they're managers, right? That might be lower quality.

And so it just creates this whole negative cascade. And I think, you know, it's painful, but just holding the line is really, really impactful. Okay, terrific. What are you seeing right now that you think most smart people are missing?

Most smart people are missing. I think patience is a virtue, right? Kind of goes back to what we were talking about earlier. Yeah. I think there's like a whole wave of people, right? That are just doing things that are maybe like locally optimal or make sense right now, but don't make a lot of sense for the future. I think even beyond the career decision side of things, that's also true of startups.

There's a lot of startups you see right now getting funded, people working on problems, where if you play things forward over the next two to three years, it's actually not super clear why or how that makes sense and how that's defensible given everything else that's happening. I think the people that are slow and steady with the right kind of intentionality and plan behind them actually will have a lot of success over time.

And can you walk through then how do you make a tradeoff? If the virtue is patience and the world is changing so rapidly and you know that in two years, the pain points or use cases that people want you to solve today aren't going to exist, can you walk us through how you're making those tradeoff decisions between what to build or focus on given the fast, rapid pace of development? It's absolutely the wrong thing to be building for the future two years out and have that ready to go right now and nobody's there to adopt it.

That's just as bad as the other direction. And so I think what you have to do is you have to think about the North Star, where do you want to be? You have to decompose that. What are all the constituent parts that need to be true? And then ideally, there's some subset of those things where you can do that now and it still delivers value. And then you're kind of building this bridge between current state, future state, and you're bringing everybody along for the ride.

That's my mental model, at least for me to deal with it. Okay. What have you refused to compromise on and what has that cost you? I go back to the hiring point. I think that that's the one that comes to mind. It's cost us some speed, for sure. There's a lot of short -term wins that we've definitely had to sacrifice because of keeping the talent par high, but I think on that it's probably the right thing to do.

Okay. And what belief do you hold that you think others would disagree with you on, that other smart people would disagree with you on? People would disagree with me on.

These are all these curveball questions. Yeah, I know. It's making me think. Yes. This was not like this eight years ago.

Here's one. I think a lot of people think that AI is transforming everything, which it is. So in a way, I kind of agree with that statement. I think being very close to the enterprise and how businesses are using AI, I actually believe that the ROI realized by AI is quite low. Right? And I talk to so many organizations right now and CIOs and CEOs, and the number of failed pilots and experiments and all these things, I mean, I have hundreds of examples.

Right? And so I think the learning from that, right, is like, you could have the foundation, you can have a lot of the core technical pillars, and maybe they're good for consumer use cases. But when you're talking about high complexity, when you're talking about big companies and making advanced use cases work, oftentimes in mission -critical settings, that last mile is oftentimes the bottleneck.

And so I think that's one takeaway. I think a lot of people think, oh, AI solved everything, and we are still scratching the surface. There's so much more left to go. Okay. Okay. That's great. What part of you has contributed most to your success that people wouldn't expect?

I think it's actually humility, right? I think for me, I'm very comfortable and eager to be the dumbest person in the room.

And humility coupled with curiosity. And so even going back to before we started the company, all that research, right, I would just pull aside security practitioners and be like, hey, explain these concepts to me, right? And, you know, go deep. And even now, it's like all the time. I'll ask people on the team, like, you know, IC is like, hey, like, how does this thing work? Or what was the intention behind this? And so I think having the desire to constantly learn, but not being, like, worried about what other people think or not having to, you know, posture and all of that, that's actually been a very helpful asset, I think.

And did you always have that ability to just not worry about what people think and just ask the dumb question? Or does that just come out through experience?

I think some of it's come from experience, for sure. I think Stanford was very helpful for that because, you know, you come in and you realize there's all these great people around you and so much to learn from. So I think that was a very formative aspect of it. Okay. And the final one is, what's something important about entrepreneurship that you had to learn the hard way?

Execution is the hard part, right? You can have all these grand ideas and a bunch of plans that are laid out, and that's wonderful. But then to actually make things happen, right? As an example, you could have an incredible plan for your go -to -market, right? Here's how I'm going to go acquire customers, win the market. Actually getting your first customer, first five customers, first ten customers, it's like pushing a boulder up a hill, right? Incredibly difficult.

And so I think that's something I have newfound appreciation for. And oftentimes in the early days, it's just pure hustle, right? It's just effort you put in. And none of these things seem like they have ROI on the surface, right? Going to that event where you might come away empty -handed and not meet any interesting people, hopping on the Zoom call with a candidate who may not close, just keep chipping away at it.

Eventually things land and it gets a lot easier over time. At what point does it get easier? At what number of customer does it get easier? Or is there a milestone? I'd say for us, it really started to shift around maybe like 10 paying customers and then 15 to 20 people on the team. Okay. And then now there's kind of like this baseline legitimacy that you've established for yourself.

And then, you know, people can tell their friends and all those different pieces. Okay. I'm going to open it up, gang, for Q &A in one moment. So we're going to get the mic set up. Can you walk through the reality of how that actually works to close $53 million from Greylock and Bain and unpack that for us? It was actually a pretty tight process.

So we had a target list of firms. We wanted people who had experience in our domain, who had experience in enterprise software, AI, well -regarded, good brand, people who could really push the company forward. It was a super tight process. It was less than one week. So we had first meeting, the term sheet signed in about five days, I think.

But let's just be mistaken. And it's the reason why that was so quick. I know the technical process lasted a week. But why were you able to get it done so quickly? Was it a process that had actually begun with prior relationships way before? So prior relationships, I'd known actually the folks at Bain for many years before starting the company.

Right. But from an actual investor contact perspective, we were just heads down building, right? So I didn't take any investor meetings, kept a healthy distance. I think these investors, their job is to keep tabs on the company. So they kind of had an initial sense of what was going on. We had some prior relationships. That definitely smoothed things over.

But I just want to make sure, because I think this dovetails well to this whole point on relationship building.

Just because, was the experience at Abnormal related to getting funded for Cogent? Without a doubt. Without a doubt. So can you give more details around that? Yeah, absolutely. And I think the Abnormal experience definitely was very helpful from a perception standpoint. But were the VCs involved in Abnormal? For a very first round, not the most recent round.

Okay. But for the first round, that was Greylock. And so that round was all relationships. We didn't even have a pitch deck. So can we unpack that? And that was an $11 million round. Correct. Is that right? I just want the students to understand the realities, like the specifics of how that actually happens. Because you couldn't just go into Greylock stores and just get a check in a day for $11 million.

That was a multi -month process, right? So we had relationships for many years. Greylock was actually the firm that introduced me to Abnormal in the first place. That's where it all started. And how did that introduction happen? How did Greylock introduce you to the employer that you were going to work for? How did you initially get connected to them? I had actually met some of those folks back from the Stanford days.

Okay. So that's what I'm trying to get to you. So what happened when you were at Stanford to get connected? I don't even remember what the initial interaction point was. But some of those relationship points were formed. They introduced me to Abnormal, got into Abnormal, stayed close over time. And a lot of these people, they want to be working with people who they know and who they trust.

Yes. Over their self -experience. Yes. So I was leaving Abnormal, working on a company, formed a relationship, we're working closely together. And then it kind of organically happened. And were you proactively staying connected to them? Or are they proactively reaching out to you just saying, hey, what are you doing now, Vanit? After I left Abnormal, it was much more bi -directional.

I was reaching out, them reaching out as well. But definitely in close contact. And is there any tactical tips that you would recommend for students that want to get funded by VCs in the future on specific tactical things they should be doing in advance of starting their companies? I think the relationship piece is definitely helpful. So I think forming connections with those VCs such that at the time you raise, it's a relationship that's been budding for six months to five years, whatever it might be.

And then I think secondly, it's just like work back from what it would take. I think one thing that I even do to this day is I think about my future rounds is almost write the investment memo that I anticipate them selling internally. And so when you have that, then there's a lot of things that you can now work backwards from and think like how do I go make that true? A similar piece of advice I give to all our employees is like write your LinkedIn profile for two years in the future as well.

And so I think based on that you're like, oh, okay, it would be great if I had XYZ experience or ABC person on my team. I'm going to actually make those things true. Okay, awesome. Okay, Caden, we'll tee you up for the first question. And then gang, if anybody wants to ask any questions, just cue behind Caden. We'll try to get there as many student questions as we can.

Awesome. Yeah, you're good. Okay, awesome. Hi, I'm Caden. Thank you so much for coming to speak here today. So obviously you're in the cybersecurity field. I mean, with regards to that, what do you trust agents to do kind of end to end versus where do you need more human reliance and more traditional methods? And then where do you see the field of cybersecurity as a whole going in the future? It's a great question.

It's something I spend a lot of my time thinking about for obvious reasons. So I think the way we kind of think about it, right, is there's almost like a two by two where one axis is security benefit and security impact. And another axis Access is operational impact. So in most use cases with customers where there's no potential issue that could happen, right? Maybe it's like an agent that's just gathering context or returning an answer, right? Read -only type of situation, totally fine, right? Now, as soon as you turn into right actions, especially things that can have negative downside.

So as an example, if you try to patch a vulnerability and you do it at the wrong time on the wrong system, you know, you might like take down a business, right? And they may lose like hundreds of millions of dollars of revenue over the course of the time. So a month, that's actually probably worse than getting breached, right? And so, you know, what we do is we build in by design a bunch of controls on these agents.

So there's certain things that they can do, there's certain things they can't do. And we also have this human governance layer on top of it. And so as an example of how that plays out is a customer may say, Cogent, you can auto -remediate vulnerabilities for people's laptops, because if you reboot someone's Google Chrome, you know, not the end of the world, or someone's development environment, but you're not allowed to touch our production environment.

And so those sorts of things are helpful. Now, where do I see this all going? To the point from earlier on, we're getting attacked by swarms of AI agents. I think there's literally no choice but to get to fully autonomous defense. So I see it unfurling kind of like, you know, self -driving cars and Waymos, right? They're running all around San Francisco.

They're probably not in Chicago, right? So it's like, let people slice by slice, kind of adopt this technology, build their comfort. But widespread adoption, it's not even optional, right? It's just the only way that people can survive. Awesome. Thank you so much. Thank you, Cogent.

Thank you so much for being here. My name is Will. I was wondering, you talked about AI use in businesses having like low ROI or no ROI. And that's because the last mile is a particular bottleneck. I was wondering if you could explain more like what that bottleneck is, and maybe an example or two of what that would look like. Yeah, absolutely.

I think a lot of what I've seen bottlenecking AI adoption is twofold. I think one, people in process, right? So, you know, oftentimes, like these tools and these projects that require people to stand them up. And if those people don't have the right expertise, they don't have the right desire. If you're trying to automate something that maybe is not even working properly in the first place, it becomes very difficult.

So I'd put that under the category of maybe like ill -formed use of AI, right? Where AI sounds like a good idea on paper, but maybe there's not the right prerequisites in place to make it successful. And then more from the technical side, I think the fuel to AI and ultimately the thing that is one of the highest order impacts in terms of how effective AI can be is data in context, right? So, for example, if you want to automate financial fraud detection, right, in a consumer lending company, and you don't have any insight on what fraudulent claims might look like, right, that's going to be very difficult.

And so I think oftentimes people don't have the right data foundation in place. They're just tossing AI at some stuff and hoping it'll work. You have the smartest model in the world. You don't have the right data. It's not going to work out. We found that actually for us, this whole context graph that we've built that models out the enterprise and on top of which we have a bunch of agents taking actions, getting that first piece right unlocked a huge wave of outcomes.

So that's another thing that I've seen as well. Thank you.

Hi, thank you so much for being here. My name is Charles. So when you were at Stanford, you participated in venture capital groups. How heavily would you recommend that to people in this room who will eventually be on the other side and receiving investment?

I think it was really helpful, primarily just from a connection standpoint. I think I got to learn a lot more about how the industry worked, meet a lot of people who ended up being very useful in the future. I think my broader advice would just be explore. Dabble in everything. See what you like. See what makes sense. And then from there, choose what to double down on.

And now's the perfect time to get that exposure. And who knows, maybe after you kind of explore some of these things, you realize that's the side of the table that is more interesting to you. So that would be my general framework for thinking about it. But I thought it was quite helpful just from a standpoint of understanding the broader ecosystem and meeting a lot of really cool people.

Thank you so much. Hi. I think it's a lot of really cool things.

I think it's a lot of really cool things. I think it's a lot of really cool things. I think it's a lot of really cool things. I think it's a lot of really cool things. I think it's a lot of really cool things. I think it's a lot of really cool things. I think it's a lot of really cool things.

I think it's a lot of really cool things. I think it's a lot of really cool things. I think it's a lot of really cool things. I think it's a lot of really cool things. I think it's a lot of really cool things. I think it's a lot of really cool things. I think it's a lot of really cool things. I think it's a lot of really cool things. I've done a lot of companies before, had worked at startups, very strong pedigree.

And then culturally, I just got along very well with the people there. I could see it being an environment where I could thrive. So I think all those aspects were things that I looked at. And you never know, right? The most likely scenario when I joined Abnormal is going to be like, you know, average outcome for any other startup. And then obviously through good execution over time, it really took off.

So you could never get it 100%. But I think if you do your homework upfront, you can maximize the odds. Thanks.

Hello, I'm Athena. Thanks for being here. I wanted to know if you were thinking about like post -quantum migration and like if that's even a discussion amongst your team or if like any of your customers are thinking about things like that. Just because like fully autonomous defense, I mean, feels like something that could incorporate that.

So I was just wondering, I know it's like further out, but I want to know your thoughts. Yeah, absolutely. And that's going to be, I think the next big wave as it comes to cybersecurity and even technology more broadly. I think not many customers are thinking about it now because it's still a little bit away. Cybersecurity sometimes is reactive, right? Where you need that like big explosion or bad thing to happen, right? And then people start to concentrate.

The one thing that we've built that I think is super powerful is to get this context graph that enables us to very deeply model how the business actually works, all of its assets. And so we are actually seeing some folks who are on the bleeding edge come to us and say, hey, can you map this and kind of figure out what our weak points might be and attack paths might be.

So starting to think about it, we're scratching the surface. I think this is going to be a huge thing in the two year timeframe. Okay, thanks.

Hi, my name is Lisa. So nice to meet you and thank you for a great talk. I need to go closer to the mic just because it's hard for people to hear. Okay, sorry. Okay. So during your talk, you mentioned how finding the right market is very important for when you're building a startup.

Let's say, for example, I have like an idea that I know that not many people are talking about right now, but I would be very interested in exploring. How would you distinguish between knowing whether that's like a good niche market that just hasn't really been established yet and could generate a lot of good public interest and like it'd be a good market to pursue versus like whether it's just a market that isn't being talked about because it's not a good way to pursue.

I think it all comes down to data collection. So what I would do, and again, it depends on the nature of the market that you're thinking about and some of the dynamics there, is that I talk to your potential users. I would do kind of what I was describing earlier, understand. I'm sure people have tried to solve the problem in some ways in the past.

Maybe it was a very legacy manual type of approach. Maybe it's something that's more comparable to kind of what you're thinking. So I would try to map all that stuff out and kind of get a sense and then really build conviction that, hey, if this were to work, it could be big. And then it's totally fine if it's non -consensus. It's actually better if it's non -consensus, but you want to avoid going down the path of, hey, finger in the wind, I think this could be cool.

And then now you've invested so much. And then a year later, you've got this big freight ship, right? Because your company turns from a speedboat to kind of like this big barge over time, and it's harder to maneuver. The earlier you kind of figure out what it is that you want to go after, just keep kind of focused on that direction, the better that's going to be ultimately as an outcome.

Thank you. Hi. Hi. Okay. My name is Roman. Thank you for being here. So my question is, so you talked about how you find the idea, like you identified the problem. Part of it was your experience in the industry, right? I was wondering whether for you it was more about this experience, like talking to people and customers, or did you have like this intuition that with all these big AI models, like the market would open like way wider than it's? It was a combination of both, right? I think it was really three things, right? I think first of all, I'd seen some of these problems myself and literally dealt by hand with some of the things that Kojin is now solving for.

So I knew, okay, this is a very big, painful area. I think number two, talked to a lot of organizations, understood how they felt about this, what the appetite would be. And then thirdly, it goes back to that first principles point of just really kind of like playing things out, right? It's like, it's untenable, right? Like we just like sit back and do the same thing that we're doing.

It's not going to work. This must exist in the future. I think that's another framework by the way to think about. A lot of the best companies, it's like you project out five years and this thing just like has to exist, right? There's no way in which it could not exist. And I think that's sort of the thought that we had with Kojin as well.

Okay. Thank you. Is there another company that you would start today that has to exist that nobody's starting? I've got to keep those ideas to myself. Oh, this is a safe community. And, but it's tough. And I think like to really come with, come up with those sorts of things, those sorts of great ideas, it takes a lot of time. And, you know, we talked about it, right? It's like a months long process.

It's immersion in this space. I'd say if I wasn't working on Kojin insecurity, which obviously I have a lot of depth in,

identity and access for AI agents is something that's very interesting, right? So again, first principles in the span of a couple of years, agents are going to outnumber humans 1 ,000 to one, if not more. How do agents work together? And how do they collaborate, right? In a safe and secure way. We have technologies that support that for humans.

But if humans are no longer the predominant force within a company or within an institution, I think that's a very interesting problem to track down. Okay. Well, that's a nice little jewel to give everybody. I want to give you Vinit final words. If there's any final words that you want to impart. This is the next generation of Vinits here.

Is there anything else? Is there anything? I know you've talked about a lot already, but anything that you want to finally impart to Stanford? I mean, I think we talked about a lot of, you know, frameworks and pieces of advice. And I think those are all things to keep in mind for sure. But also just have fun. I think like the future is so bright.

It's such an exciting time to build. You know, this is also a really exciting place to be. So I would just, you know, make sure however many years you have left, whether it's like, you know, six, seven months or, you know, you're just starting and you're a freshman, lean into it, make the most out of it. Some of the fondest memories that I have are from what I was listening to.

Did you have fun at Stanford? I had a lot of fun. Were you also stressed? Yeah. You know, there's things I think sometimes go hand in hand. Okay. So it's not a mutually expensive view. But don't stress too much. Enjoy campus while you're here. Okay. On that note, let's all thank Vineet. Thank you very much.

You can find this talk and others at the Stanford eCorner website or on our YouTube channel. And join us next week when we're going to have Michael Bernstein. Thank you all for coming. Thank you all for coming.

The Entrepreneurial Thought Leader Series is an original production of Stanford eCorner and the Stanford Technology Ventures Program. To learn more, please visit us at stvp .stanford .edu. To learn more, please visit us at stvp .stanford .edu slash eCorner. Thank you all for coming. Thank you. Thank you.

Thank you.