Jeremy Stein: Could you tell me more about your background, especially with Nvidia?
Ambuj Kumar: I'm Ambuj Kumar, co-founder and CEO of Simbian. We are using AI to automate security.
I started my career at Nvidia back when it was a tiny company that was one of dozens of companies trying to make gaming video cards. But what differentiated Nvidia from all others was that we were building GPU (graphics processing unit) and our goal was always to do parallel computing and gaming cards was just one application.
So that foundation gave us that edge, and eventually, all other video card companies died one after the other or were bought by somebody else, ATI being the last one to get acquired by AMD. And as we know right now, Nvidia is ruling the world.
So I was privileged to work there for eight years and led multiple generations of GPUs there. I then became interested in cyber security so I worked for a company called Cryptography Researching Inc. CRI founded by Paul Kocher, the creator of TLS, so a very elite set of colleagues there and in 2016, inspired by what I had seen at Nvidia, Jensen Etc I started my own company Fortanix.
At Fortanix, I created a new type of security called confidential computing, raised 135 million dollars, hired hundreds of employees, obtained customers etc.
So I ran that for seven years. Last year, I started Simbian, a seed-stage company backed by some of the best investors that works with some really incredible partners and customers.
Jeremy Stein: That's pretty exciting. Why is Nvidia one of the largest companies in the world? Why did they get it perfectly right?
Ambuj Kumar: Nvidia has a very strong culture. Intellectual honesty, I would hear that 5 times a day. I would hear the term “speed of light execution”, 10 times a day. I would have people replying to my emails at 2 am in the morning. Intellectual honesty is pretty critical. What it means is that you understand things from first principles. You are very clear on what you know, what you don't know. And you are not, pardon my french, BS’ing others, by telling something that you don't believe in. And so it created an open environment where conflicting voices can come together, collaborate, discuss and find what's the real truth. That's been the guiding factor behind Nvidia since day one. And if you accumulate that right over 30 years, it does magic. Most companies run into a ceiling, where people at the top, or nearly at top, they're trying to protect their jobs or they are trying to protect their viewpoints. Nvidia is just the opposite. You are always seeking the real truth.
So that intellectual honesty, seeking the ultimate truth, trying to see, believing that parallel computing is the future. That's the hypothesis and then trying to insert it at as many places as possible. Taking feedback from the market. Correcting. And the other one is the speed of light execution. So you are always benchmarking yourself against what's the best possible theoretical limit. So the speed of light is finite. It's not infinite. So it's finite and you want to match that. It’s not about we are 20% better than yesterday. Or we are 30% better than our competition. That does not cut it at Nvidia. You have to compare yourself against what's the best possible scenario.
Jeremy Stein: How are you translating this culture to Simbian?
Ambuj Kumar: We are also strongly focused on our culture and values. So every company is slightly different because of different personalities.
There are three things that are important for Simbian.
One is what I call excellence. Excellence is being the best in the world at what we do. So we can choose our craft, we can choose what we want to do. But once we have chosen our lane, then we want to be the fastest and best. So that is excellence.
Second one is replication. Replication to me is taking individual excellence, translating into organizational excellence. So how do I learn from you, how do we collectively learn and figure out things.
Third one is accountability. Accountability and ownership, they go closely together.
Accountability is again comparing yourself against what's the reality of the universe. What's the best we can do. Not pointing fingers because it doesn't really matter whether it's your fault or my fault. It is, we should collectively do better. Companies should improve. Individuals should improve. And if you can create a right culture that is moving in a certain direction and the right culture will make sure that you accelerate and this is when magic happens when compounded over many years.
Jeremy Stein: Would you say culture is one of the key ingredients of growth?
Ambuj Kumar: Yeah hundred percent, because everybody has the same ideas, everybody lives in the same world. Everybody knows that AI is big. Everybody knows for me that security is huge. Everybody has nice ideas. It’s the execution that sets it apart. And why would you want to work hard on something if you don't believe you are not working with the right people. If you are not feeling inspired, if you are not inspiring others. So it always comes down to personal values and personal values translated into a company is the culture.
Jeremy Stein: And then what are some other learnings for building a really strong company culture?
Ambuj Kumar: There is no one path to success. What I mean by that is that Nvidia culture worked great for Nvidia. Simbian culture will work great for us. And these are closely related right because of my experience, but there might be some other company that might treat their people like commodities and they still may be successful. So I'm not here to judge them. At the same time you should not be judged by others. So culture and values are very personal, there's no one path to success. But you have to remain true to yourself. So what it means for me is that when I interview somebody, it don't really matter how smart they are, but if I feel that they are somebody who won't get along with others, or they would have this thing that hey just because I have been doing something for 20 years, I don't want to talk to somebody junior or somebody new cannot have ideas that will match mine. I will never ever work with them. They might be very successful, but Simbian is not the place for them. Conversely, you want to make sure that you advertise your culture, express your culture correctly to candidates, let them self select out if they don't feel that they're the right fit. So you have to be true to your culture and that's hard.
Jeremy Stein: And then how do you make sure you're hiring the correct people for your company versus the wrong person. Are there red flags that you’ve noticed?
Ambuj Kumar: Yeah, bunch of things. Junior people or people with not much experience, you can interview them on the basis of their work, their ideas. What great things they have done, what great things they want to do. People with lots of experience, you also have to then look at their track record. You can also do reference calls on them. But it's very interesting in my view, the right people reveal what they want to do. So, if you ask somebody, one of my favorite questions is that hey, why do you want to work for a company like Simbian where there's lots of work, we can't pay you enough salary, you are going to take a salary cut compared to large companies, very intense culture, because we want to do the best the speed of light execution.
We want to be the best in our thing, generate a constant barrage of ideas from different people. and just see how they respond. Some people like this and they are the ideal for us. They feel inspired. They feel that they have this gift, and they are looking for a venue where they can apply their gift into the best creation. So they want to work with other people who will push them to their limit. They feel Simbian could be the greatest company one day and they are fully motivated, passionate and that's the primary driver and those are the kind of people I want.
Other people, they get terrified. I'm looking for a nine to five job. I want to do the minimum required to satisfy the job. And again, they might be great candidates for other companies, not for us, because we are in this constant mode of if I want to build the world’s greatest security company, right, I don't even know what job requirements I would have a couple of years from now. So somebody who is preparing for a static job description, they will not fit. So I think if you express what you want to do, what you want from them, people tend to either love it or hate it and it's really clear.
Jeremy Stein: And then how are you building autonomous security or how are you getting such a technology advantage over other companies? Why are you doing so well in the market?
Ambuj Kumar: Yeah, good question. So look, cyber security is very important and 10 years from now collectively we will be spending more money on cybersecurity than we are spending today. Just because it will be more important. Ten years from now we'll be spending more money on education and healthcare than we are doing today because those things will become more important. So people want to invest in cybersecurity, that's not the issue. but worldwide, there are only six million people working in cyber and so these six million people, they are protecting the digital well-being of 8 billion individuals.
And if you look at the number of cops, number of police in the world, number of people in armies or firefighters, the six million number is very small. And to make it worse, when you are a firefighter, you know how to fight the fire once, you train for it, and then you more or less set for the bulk of your time. 90% of your skills are done. In cyber security it’s very difficult. Right, the kind of attacks that work today will not work tomorrow because there is an active adversary on the other side who's trying to find new ways of breaking you. So it’s about a small team constantly trying to reinvent new things, constantly trying to counterbalance new attacks, it’s very difficult.
Simbian is building AI agents that can operate your security. So you have 10 people on your security team. You can take the equivalent of another 10 from Simbian or hundred or thousand or ten thousand and these AI agents, they have the skills to protect your security, they are autonomous. If you have 20 things to do, they will take care of ten of them. Very different from security today, if you take a security product in the beginning of the morning, you have 10 things to do after you start to use a new security product in the evening, you will have 20 things to do.
With Simbian you have 10 things to do in the morning, you start to use Simbian during the day, in the evening you will have three things to do. So our AI agents are so powerful and strike at the core of the problem, which is the exploding number of attacks, talent shortage, overwhelmed teams, that's why we're doing so well.
Jeremy Stein: How did you find your value proposition? How did you sharpen your language to what it is now?
Ambuj Kumar: Yeah, good question. So it is always a combination of internal intrinsic things you want to do versus discovering what the market needs, so you have to have some original thesis, without that, you won't go anywhere. For us, the original thesis was that there are six million people who are overwhelmed, have lots of work, and it’s going to get much worse.
So what is our solution? Okay, we'll build AI agents, it will do this this this, so these are the intrinsic things, then we start to go talk to customers, and we said that we can do this this this, a customer says that okay, these five things are not that burning need for us, but these things are, and additionally, if you can do something for these three things, it will be more, so it's okay, like we started with 10 things. Now, we have eight things, then we go to some other customers, and they come up with their own requirements. After talking to them you start to see some patterns, so you want to have a common set of features, products, capabilities, whatever you want to call it, that will appeal to your segment. So you want to have a product that does something for a particular segment and the segment could be a small company, large company, healthcare, finance, some vertical and once you've figured out what you can build, that will deliver value to a set of customers and be repeatable. Then you want to go full speed. So that is the common way of finding product market fit.
Jeremy Stein: When you’re advising one of the startups you’re working with, what would be your two or three main suggestions in terms of expanding their growth and finding product market fit?
Ambuj Kumar: Yeah, I think it depends from company to company so sometimes you look at a company, they have great unit economics, meaning that they are selling something 400 dollars while it costs them five dollars to build and service, then you have to ask questions about what is holding them back. Is the market they are going after a small market or let's say it's a large market then okay, so there is enough to do there then maybe something they are doing that they should be doing more. So maybe they have five sales reps in the company, what would happen if we add another one? If you go from five to six what happens if we go to 5 to 10. What happens when you go from five to 50 or 500? So let's say that you want to go from five to ten, what other things do we need? Do we need to create more leads? So how are we generating leads today? Maybe you are doing this advertisement. Maybe we can do more. So maybe we're doing events. So looking at a problem. So advisors are bringing their experience from the outside world to this company. So I look at a problem. What is holding them back from beginning to end right maybe the thing is that the product team, sales team, marketing team, engineering team & everybody is great in your company, but your product is not differentiated enough. There are 20 other companies doing exactly the same thing that you are doing.
So then what you want to do is right, you want to make your product more complex, not by adding more complexity, but by solving even bigger and harder problems.
So you look at adjacent areas. Maybe you want to make your product wider so you can do more things. Or maybe you want to make it go deeper. So it was going after 70% of a bigger problem now it can do 90% of that problem. So you need to work on differentiating your product.
Jeremy Stein: How do you know when to go deeper vs. wider?
Ambuj Kumar: Yeah, good question. So you want to go deeper if people are willing to give you more money. Deeper is always easier because you are selling the same thing to the same buyer.
For example, we had 20 features yesterday, now we have 30 features, you are giving us $20,000 yesterday. Now, you can give us $30,000. So that's easier. Wider is selling the same thing to multiple people, a wider group.
But if your goal is to add differentiation, then you look at what are the things that are hard, because you don't want to do something again that there are 100 other companies. So you have to have some background on where the market is going.
Right and what are your unique strengths? Maybe your team is great at building the world's fastest database, that's distributed, that has less than five megabytes of stuff stored in them. Then you can think of embedding their database in the browser.
But It proves the point that you have to look at what are your unique strengths. And you want to play to them and then look at the market, what the market wants you to do and you have to combine those two.
Jeremy Stein: What are some other learnings on making the product deeper vs. wider? What are the ways to do this most effectively?
Ambuj Kumar: I think I mean first have ideas, think carefully. If you think that you have done it, think again, because there is no dearth of information in the world right now, so, just go read all the blogs, go watch YouTube, go read what people are saying about your competition etc. Once you have done your homework, then pick five, ten people who are knowledgeable in the field and say this is what I’m thinking, what do you think? And so this is an increasing level of cost when you are doing your own research. You can spend a hundred hours, one hundred years, whatever, collectively but that's the easiest to do.
So you do that you build an idea first or slide, run it by people, get their validation, build an MVP, a little bit more, then you can play with it. You said that this could be useful if you build it, now we have built it. What do you think? Oh, this is great. Then you will start to ask them for money. Right once you get that then you figure out what is common between them and just run with it.
Jeremy Stein: Can a company go deeper or wider at the same time, or do they have to choose one?
Ambuj Kumar: I think normally they have to choose one over the other, at least in the beginning phase. Typically the way you go wider is by making your product more generic, so google search engine, right? What do you search for? You don't even think that, you can search for anything, you can search for a product. You can search for blogs. You can search for individuals, you can shop there.
GPUs allow you to do AI, what your AI does it's up to you, you can use it for cyber security. You can use it for healthcare, you can use it to write a poem, so you go wider by making your product more generic. And go deeper by making it more use case driven.
You can have multiple products that are use case driven, but that's like an ensemble approach. But normally you have to typically choose between the two.
Jeremy Stein: How quickly would you want people to pay for your product,, or could you go with more of a freemium model?
Ambuj Kumar: I mean that's a conversation for another hour, but you can do either, my general advice there is to ask for money sooner rather than later. Because money is a good representation of value, objective measure of value. So if people are saying all the nice things but they're unwilling to give you money you would learn something. Conversely nobody thinks that they like your product but everybody is giving you money that means that you are certainly going after an important problem.
Jeremy Stein: Should the founder be asking for money from customers, or should it be the sales hire?
Ambuj Kumar: I don't think it matters as much but the way these things work out that founders become the first ones to ask for money because they are still building the company, they don't have a salesperson, but if you have a salesperson early on I think they are literally trained to ask for money so you can use them but in either situation founders must be very closely tied to what's happening. They can't be hands off.
Jeremy Stein: How many deals do you think it takes to find product market fit? With Simbian, how did you know when you had product market fit?
Ambuj Kumar: When you build your product for the first time, you spend let’s say 300 hours researching the topic, researching the problem, and maybe let's say another thousand hours building the product. So you did 300 hours of research, thousands of hours of building a product. That's quality step 1. When you have perfect product market fit, your product is already ready. So it requires 0 hours of additional work. And when you talk to a customer it takes three minutes.
Do you want to use AI? Yes. Are you interested in large language models? Here is my GPU, take it. So you didn't need to ask lots of questions so the research required to figure out whether somebody is your prospective customer has gone down, from 300 hours to 30 seconds. So that is perfect product market fit and then the whole continuum, and so the sign of product market fit is that you don't have to know as much as you did yesterday about your customer to tell them whether they can use your product and you don't have to constantly write new code in your product to make it appeal to that customer.
Jeremy Stein: If a 10th person from the same industry/same characteristics comes on, you know most likely they’ll become a customer and most likely you don’t have to write new code?
Ambuj Kumar: Exactly exactly or they don't even have to be from the same industry. So we just talked about Google search engine, right, everybody uses it from different industries or your product can be suited for SMBs, small to medium businesses of any industry, right, or it can be suited for CFOs of big companies or small companies, right? So something where maybe you scan all the SEC reports on financing and use AI to crystallize some insights and educate, what's to look for etc. that could be a product that might interest a CFO of any company. So people can come from, you're buyer can be from different verticals and have different job titles, can have different roles, can have different seniority but there is something in common. So if I have built this product that watches all the things that SEC is doing and new guidance around revenue recognition etc and that's my product. And I know that any CFO will buy this product then I will not ask you that hey, are you interested in biology? Do you work for a small company or what does your company do? I'll simply ask you, are you CFO and if you say yes, then I know that you’re my buyer.
Jeremy Stein: In terms of scaling, aside from culture, what would be your key suggestions?
Ambuj Kumar: Yeah, work with winners. There are people who have a chip on their shoulder. They want to create the world’s best company. Work with them. That takes care of lot’s of problems.