Transcript: Vinod Khosla - How to build and manage a team

February 4, 2018

We have put together another transcript - This time a great talk by Vinod Khosla, Founder of Khosla Ventures and previously the founding CEO and co-founder of Sun Microsystems. The focus is on his belief systems around hiring, and how to manage your company’s growth internally.

 He was interviewed by Anu Hariharan, Partner at YC Continuity.

 The full original video can be found here:

Anu Hariharan, Partner at YC Continuity, sits down with Vinod Khosla, Founder of Khosla Ventures and previously the founding CEO and co-founder of Sun Micros...




AH: You've talked a lot about company building and hiring, and I know that you've often said that a company becomes the people that it hires, not the plan it makes. Can you elaborate? What should founders focus on in hiring?

 VK: This is one of my favourite topics. When you start you know very little about what you're doing because if you did know enough, you wouldn't start the company. If you were really sensible and pragmatic, you wouldn't start companies.

 I'm an entrepreneurship zealot - that's self-described - and I'm a real fan of people trying companies and trying to build things, especially when you're younger and can afford to take risks. There's no reason to go work for someone else in my book, and I never did.

 What I find is people have plans, but it is very, very rare for a company to achieve that plan. So the way to think about where you're starting is a collection of ideas in a relatively rich space that you find that you can be passionate about. That's the way to think about a plan.

 I have a separate talk on flexi-planning which don't plan, just plan to plan.

 But the critical part of that is the team you build ends up making all the iterative decisions about where you're going to end up. Whether you're going to hire the next person from this area or that area, or that strength or some other strength, or take path A or B, is very critical to where a company ends up.

 There's a huge path dependence, and the path depends so much on the people you hire. Your starting plan becomes much less relevant than the people you hired initially - especially the first 10 people. Those 10 people will then go on to hire the next 50, and they will then hire the last 100. You can fire the original team, and its gene/DNA stays in the company. So that's why it's so important, and why the people you hire determine the company it becomes. We can come back later to talk about where I see mistakes made around functional hiring.

 People always ask me, and I say - this gets tweeted a lot - there's a huge difference between a $0 million company, and a $0 billion company. And the difference is attitude and ambition, and mostly the kind of team you build, which will reinforce certain directions for you, and who you end up hiring. It's far more critical than people believe. And it's really hard to hire for those skills later.

 AH: You touched upon the first time hiring being so important, and especially defining the cultural and belief system of the company. So if a lot of founders here have one or two-member teams, how should they think about who should be those first few hires? How do they go about finding those candidates? I know you have a strong view about how equity should be split among the first 10 employees versus the next 30 employees, so how should a founder think about that?

 VK: Many startups are very tactical, and they should be. But in doing the tactics, you can sometimes forget what the big vision is and you may even exclude the big vision. I was talking to a company fairly recently in our portfolio, and they had hired a bunch of people to do particular tasks that needed doing.

 There was a new hire - potentially a big hire, a very attractive person to have join the team - and I asked the founder who will interview him? The founder told me, and I said if I was a really great candidate, and these 3 people interviewed me first, I would drop out of the recruiting process. So I said, why don't I be the first interviewer? Because I want him to grab the big vision first, and not just be interviewed. Let me sell him on the big idea, and why he can be a part of forming it.

 The best people almost can't be interviewed. They have to be sold and evaluated. Evaluation is important, but evaluation and interviewing are different things in my mind. Once you've hired the first 5 people, and they're interviewing other people, you might actually exclude the best candidates. That's very tricky for a founder.

 At the same time, sometimes you can't afford the big names so you don't have enough to attract them. Those are very tricky times, and that's why I think whether you're a $0 million company or a $0 billion company sort of gets set by the time you're at 15 or 20 people.

 There are people who have done it very poorly - I won't give you examples - but there are people like Stripe (Patrick Collisson) did an amazingly great job of his early hires, and has continued to do that. You can completely change the trajectory of a company by that hiring, and how big you'll become.


AH: What are some of the things that Patrick at Stripe or some of these successful founders have done in the first year? Is it that they meet candidates in other roles? Because the founder often doesn't know a lot about the other roles, or doesn't have domain knowledge or the roles they're looking to hire.

 VK: Let's talk about each of those separately, because they're quite distinct.

 First, I think of company-building as going after Mount Everest. But nobody got to Mount Everest without getting to Base Camp first.

 Here's the misstep. You can't get to Everest without Base Camp. And many people forget that Base Camp has to lead to Everest. So they set up Base Camp where it doesn't help reduce the risk of Everest. So in building a company, you should never shoot for the big enchilada in one step, you should go after Base Camp 1, and then Base Camp's a very good analogy.

 However, especially with bad investors, and I have a lot of beef about investors, they want revenue, or they want break-even, or they want management...all the things that can lead to setting up business here, that then excludes the possibility that you'll learn more about how to get to Base Camp 2 and 3 - so you are out in some other local minima or maxima. And that's really dangerous early on too.

 I was just talking to somebody in the YC class and started talking about that. I said, if you get too tactical and forget the strategic, you set up Base Camp in the wrong place. You want to decide which is the right space with the technology in this case, before doing all the tactical stuff. So there's this uncomfortable balance between tactics and strategy.

 I say be obstinate about your vision, be flexible about your tactics. Another way of saying it is to think big and act small. These are really difficult things to navigate - much easier said than done.

 But back to this issue. Let me talk about a process I call gene pool engineering. If you're going after a space, assuming you've picked the right space - or the one you're passionate about, it doesn't have to be right for others, just right for you - the first thing I suggest people do is to see all the ways you can fail. Think about it. You want to identify the key risks in whatever you're trying to do. Then you hire for those risks.

 I always say to people, find your top 5 risks. For each of those risks, say risk 1: who are the companies that have dealt with that risk? Where does that expertise on that risk lie? Make a list of 3-5 companies. By the way, there's a document on our website called Gene Pool Engineering that I wrote just for this purpose. If you've identified the 5 companies that have dealt with that risk, then you have at least the places that can reduce your risk.

 Within each company, identify 3 names. Now you have 15 names for risk number 1. If you do that for your top 5 risks, you essentially have a library of names that will help you reduce the risk in your business. That is risk-oriented management. For start-ups, after you've decided what space you're going after, it is all about risk management. There are lots of risk trade-offs. Everytime you hire somebody, you increase your burn so you increase financial risk but you're reducing some other risk. Maybe it's product risk, or development risk, or feature's a series of trade-offs. That's how start-ups should think about what they're trying to do.

 So this is a process I call gene-pool engineering. You engineer 15 names for each of your major risks then go after all of them, and hopefully hire 1 or 2 of the 15. Then you've engineered the gene pool of your company for the risks that are in the plan you're going to pursue. It's way more precise than the platitude of "Let's hire great people." I find that really disturbing because I've never actually had a start-up come in to me and say they're only hiring bad people. Never quite happens...

 So how do you make platitudes into something concrete and actionable? That's a big thing to me.

 The second part of that is the reality that start-ups seldom pursue the original plan they started with. You have to hire for not only for the risks you know, but for all the other things that you might pivot to, change to, adapt to, iterate to. Building a business plan in a new area is a series of iterations. So suddenly, this risk requirement goes in parallel with this other requirement - I call this smart adaptation.

 You adapt very quickly to whatever the market is teaching you. YC is particularly good at that - teaching you how to iterate quickly and learn. Essentially, very similar to what I call this flexi-planning process - where you iterate your way to your plan versus "This is my plan." I always worry when people make plans 7 years out, and give me forecasts, because it tells me they don't have a clue what they don't know.

 That's one of the great benefits from YC - this iterative approach. Some people call it MVP, I call it flexi-planning. But the team you built, in addition to this risk approach, also has to be open to new ideas, grab new ideas, ask a lot of questions, adapt quickly, brainstorm with others constructively, so build on each other's ideas. These are the parallel set of requirements around hiring that are very hard to quantify.

 They're also full of danger. They're full of danger for the following reasons. Sometimes people don't hire for this - that is a danger. At other times, and one of the biggest risks if you decide you want to be aggressive about hiring, is going after names or titles. Big names and titles in big companies are really, really dangerous and misleading for start-ups.

 My biggest mistakes early on in my career was getting too impressed with somebody's title at IBM or some big company. Now I take the view, if somebody has been at a big company for 10 or 15 years - if they've been at Cisco for 10 years - they can't work effectively in a start-up. They must get re-brainwashed by failing once or twice in a start-up - then it'll work.

 Hiring for this other stuff and not falling into the trap of just big names is very important. It also depends upon your domain. Some domains need a lot of expertise. If you're doing a blood test like Theranos, you can drop out of school and start that. But eventually, it's a serious physics and chemistry problem. And you have to get that talent on board.

 If you're doing something like Stripe did, it was a very simple problem so you don't need 3 PhDs. If you're going to solve healthcare with AI, you're going to need expertise. So it depends on the domain. You don't always need the gurus. It depends on your problem. You don't always need the big names.

 Another very counterintuitive thing - we were recruiting for a CEO for our company, AliveCor. They essentially do cardiac care, and I insisted I would not talk to any candidate who had ever worked in healthcare. And that's because healthcare executives tend to apply old rules, because their domain hasn't changed very much. People are so used to applying old rules and applying good process, as opposed to inventing this new future or new path. That's very dangerous. We should come back and talk about how founders figure all this out when they don't have a clue about what these other functions are.

 It's seldom that a person out of a domain can innovate in a domain. And if you're a young founder, and you hire somebody with deep experience in the domain, it is logical as human beings to defer more to their knowledge - that is a real problem. It's important to understand the problems they raise, but not the solutions they propose. If you take that, you keep doing what traditional people in that industry have done. This is why CEOs hired into companies don't work very well and why founder vision becomes so important and so critical.


AH: Before we touch on the functional hiring and how to look for domain knowledge, I wanted to touch upon one other element that you talked about in the gene pool: hiring people - which is about diversity - but not in the literal sense, but you talked about the different dimensions of diversity that are important. Can you elaborate on what the different dimensions of diversity that start-ups should focus on, and why it is important to focus on that early on?

 VK: If you buy the thesis that most of what you end up executing on isn't part of your plan today, then the single most important thing becomes people who can help you evolve your plan. The more diverse the set of people - and I don't mean diverse as in ethnically - I mean backgrounds and places they've worked, areas they know, age, and level of experience.

 There are some really great things about experience - you want experience. But you don't want experience to guide you into doing what others have already done in the same business, and not innovate. That is the result of too much experience.

 Mixing the right experience, where you can identify the problems, with first-principles thinking, from fresh new ideas from people who have never worked in the domain, or people who are just fresh graduates and have no idea, but can ask fundamental questions - those are the best kinds of founders. Put differently, in big companies you need people with good process. In start-ups you need people with good iteration and good adaptation. Those are very different skills because in start-ups you invent 90% of stuff, and in big companies 90% of what you do this year will be what you did last year. So there's little innovation and it takes a different kind of person.

 Every start-up, if they're successful, will need some process people later. But process people early in a start-up can be a disaster. Too good a manager is a bad thing for early start-ups. But at some point, you really need to add that.

 So that's what I mean by diversity. If you have a bunch of PhDs, how many are experienced PhDs and how many are fresh? How many are from within your area, and how many are from completely outside your area? Every dimension of diversity will help because they'll bring a different point of view. As long as they have this ability to think from first principles - not say, this is done this way because we did it this way before.

 I should also say this field's way more complex. This is why I've written about it extensively in two articles. One is called the Art and Science of Hiring, one is called Gene Pool Engineering. There's a third one that I actually wrote , which talks about what Boards should and shouldn't do. Boards do more damage to start-ups than anything else I know. That's why I don't go to Board meetings anymore.

 Applying tradition just really hurts. When you're looking for an investor, it's not money. It's helping build the team that's way more important. Way, way more important to where you get to, than anything around the amount of money raised, or valuation, or other things. We should come back and talk about founder equity. But I spend a third of my time - maybe more time on recruiting for our companies than any other single thing that I do. Because of all the things I've talked about.

 In fact, at Start-up Grind, the interviewer said you couldn't possibly spend that much time interviewing. That's because he just fundamentally has a different belief system than me. I do - it is most of what I do. I'll go through 100 LinkedIn resumes by myself on a weekend, if I'm looking for a critical hire for a company. A few people can make or break a company.


AH: You've talked about how a CEO's job initially has to be the Chief Recruiting Officer, right? Can you elaborate on what the founder's role should be? We see a lot of start-ups where they say I'm looking for a VP Sales, and they write down the skills. But they forget about all of the elements needed for what it takes to lead a start-up.

 VK: Let's come back to this notion of functional hiring. Let me talk about what founders need to do.

 A founder's role...and when I was a founder I spent probably 40% of my time recruiting. It was not coding. You're much better off hiring the people who can code, than coding.


AH: Was that throughout the life of the start-up, or did it change?

 VK: Both in the beginning and throughout the life of the start-up. I think it's true of very early, mid-stage, and late-stage companies. Really good founders do most of that. If you can hire the people, you should generally be able to hire the people who can do any given job better than you can. I tell founders about the role, which is: any job you can hire for, you should hire for that role so you don't have to spend time on that.

 Founders have some unique characteristics. They have sort of this entrepreneurial sprit, this first-principles thinking, this attitude that says: Yeah it's done this way, but why? This "why" - and I recommend a book by Simon Sinek called "Start With Why." It is a really good book about questioning everything. But yet staying true to your vision, even if it is not precise. If you're doing something radically new, it's going to be ambiguous. Very, very few normal people are comfortable with that ambiguity. "Let's charge ahead in this direction even though we don't know precisely what the business model is, or why, or what it will take" - that leadership is a founder's role. That's something that is almost impossible to hire for.

 If you hire a senior executive from another company, there's 0 chance you'll have that. They can systematize what you're doing and get much more boost to your efforts, but not without your vision. So that is a founder's role. I call it leadership. It's a belief system in the face of ambiguity, it's a figuring-it-out kind of attitude, it's assembling diversity.

 I was very lucky as a young founder. I loved awesome people. I didn't know why we hired many of the people. If you look at it, the first 15 people we hired at Sun probably started more than 10 companies worth a billion dollars later on.

 We hired Carol Bartz who became CEO of Yahoo, we hired Eric Schmidt who became CEO of Google, we hired Bill Joy - in fact Bill Joy became a founder 6 months after we started. I so badly wanted him, we made him founder retroactively. Anything is OK in this game of assembling great talent. In retrospect, that is why we were successful. I started believing the company becomes the people you hire.

 By the way, Andy Bechtolsheim was a student in Margaret Jacks Hall not far from here, where the computer science department used to be. He had licensed his technology to 6 companies. The only one that succeeded with it is us, and we started 2 years after some of the companies had licensed the technology. But we got all the people. I hounded Andy till he dropped his PhD to join us. I hounded Bill Joy till he dropped his PhD to join us. So both of them dropped their PhDs when they were 6 months from finishing. Selling is also a founder's job. And it's gut-wrenching because you don't know if you're going to be successful, and you're trying to convince somebody to drop their PhD. It's gut-wrenching. Hiring is gut-wrenching. Or you're convincing them to drop a comfortable job at Google and join many stories.

 Anyway, back to what we were talking about.


AH: You have said when it comes to selling for hiring, no means maybe, and maybe means yes.

 VK: When a candidate says no, I convince them it's a maybe. Once they say maybe, then my job is to say maybe is a yes. I always so no is a maybe, and maybe is a yes.

 And it doesn't matter if you fail. Occasionally when you succeed it is worth it. People give up on recruiting way too early and way too quickly. I just hound people.


AH: I wanted to touch on founder's equity because you brought it up twice. A great team is a magnet for other people to join the company. How should founders think about equity, especially in the early hires?

 VK: I find, by and large - and there are exceptions and plenty of them - founders keep too much equity for themselves. They don't realize the following fact: anybody they hired who is exception will, a year later, figure out they want to start their own company, unless they are a founder. So you self-select for OK people and good people, but not awesome people.

 My son is starting a company - actually two companies - right now. He graduated in June from Stanford last year. I convinced him that him and his co-founder should just take 15% each, leave 15% for a third founder, and 15% for 3 junior founders. And then, a 20% pool after that. So they have this humongous pool. I fundamentally believe a company becomes the people it hires. I said, you will be able to hire great people.

 So for one of his companies, I think it's public and if it's not, stop recording! He just hired the VP of Engineering from Quora. Why would that person join a fresh graduate? Because, one, he helped define a fairly interesting and large vision. But two, he had this huge pool and he has been talking to people who would take that kind of thing. Any of them could start their own company.

 Let me invert it and say, when a team comes in to present to me, one of the questions I ask myself in judging how good a team is (yeah, do they have expertise in all the areas they need expertise) but the single most important question I ask: if some of these people left and started a company, would I want to invest in that company? If somebody has 2 people like that on their team, 0 chance I won't want to invest.

 Independent of whether I like the plan or not. Because they are collecting great talent. If I could have great talent or a great plan, I'd rather have great talent. Especially a founder who loves great talent.

 So it's a great way to judge a start-up - what kind of people do they hire? I was telling Sam in the last YC batch, there were 2 companies we made an offer to if they hired in a critical area. They were both AI start-ups. If they left co-founder equity - equal co-founder equity - for an AI start-up. I said I'll write you a check at valuation you want if you do that. If not, we don't want to invest. One of them took our offer and the other one didn't, which is OK.

 But I really religiously believe this idea. If you get a Bill Joy, or Andy Bechtolsheim - and they want experience because they were graduate students...they made the whole company. Eric Schmidt wanted to come work for me. By the time we had the first 10 people, we were getting so many resumes. Everybody wanted to join us, because we had so many awesome people. Other than...well me and McNealy were just fresh business school graduates and we didn't count.

 I always say, and I keep telling my son, hire some magnets. I tell my son, you graduate from Stanford, you're young and nobody is going to come join you just to work with you. Think about who you could get that everybody wants to work with. That's how you'll hire numbers 10 through 30 or 50.

 A rough rule of thumb I use on equity, and I think Sam said at least 10% of the company should be reserved for the first 10 hires and I think that number is too low. I flip it around and say, when a company has 50 people, the founders should have 1/3 of the common, their direct reports should have - say 5 or 6 direct reports - should have 1/3 of the common, and everybody else, the engineers and worker bees, should have 1/3 of the common. If you do that, people don't leave. People stay with you.

 Now, Sam has a good blog post on how to attract the team. One of those is growth, but you can't always close on that. When things get tough, when you into trouble, why will people stay? They will stay because they believe in your value system. What are you going after, and what is the vision you are defining? Define an interesting and ambitious enough vision. If your vision is just to make money, at the first dip, the people will say well I joined to make money and now it doesn't look like that will happen, so they'll leave.

 The single biggest reason people stick with a company is because they have people around them who are really awesome, that they can learn from, who are better from them - everybody should feel that. That is a great culture to set up, and every time I've seen that done it has turned out really well. This Gene Pool Engineering article I mentioned - I put that together in 1995 when we were starting Juniper.

 Nobody had ever challenged Cisco and nobody wanted to challenge Cisco. And no customer in the world wanted to buy TCP/IP routers. None, for the public network. It sounds incredulous, but go look up articles in 1995. Not a single carrier, AT&T which was then Pac Bell, what is now Verizon...they all said they would never introduce TCP IP in the public network.

 We defined this large vision. We built this process of engineering the gene pool of the company, and everybody got excited about changing the course of the internet, from what was ATM and nobody remembers this. A very, very senior executive told me they would never use TCP IP. This is what telecom carriers do, this is how traditional wisdom goes. But you need this awesome team to change that.

 Pradeep Sindhu, who was one of the greatest founders I've run into, not only built a great company but actually changed the trajectory of the protocols of the internet, in my view. We would have ended up with something else. Most likely ATM if all the carriers had their way. Every single carrier from Deustche Telekom in Europe to AT&T here, had decided they would go with ATM and not TCP IP.

 So great missions, great people to work with - all that is why you want people in your team.


AH: I want to touch upon functional hiring, because I think it is a component of hiring a great team as well. Too often founders tend to prioritize a lot of skills in a particular role, versus really figuring out the complete package or fit. Can you talk about how founders should think about hiring?

 VK: Let me answer that question, but later. If you're a founder, and you need a VP Marketing or CFO. You should come back to the question - and I covered this in my art of hiring - how you judge if somebody's a good VP Marketing or CFO. Because the first thing you have to realize is that you're not qualified to judge. Knowing what you don't know is critical and one of the most important things a founder does is to decide who's judgement to trust on a topic. Not just ask their friends, because they may not be qualified. In fact, founders just hire their friends instead of gene-pool engineering a start-up. That's wasting slots and talent too.

 But let me go back to functional hiring. If you need a VP of Engineering, you can look at somebody who has managed a project. But if you're a founder, and especially an inexperienced founder with a big breakthrough idea in a new space that nobody knows, you clearly need to build an engineering team and execute on that. But you need more than that. You need somebody who is going to help you evaluate all the pivots and iterations in your plan.

 Now let's say you hire a VP of Engineering who is good at engineering, a VP of Marketing who's good at marketing, a VP of Sales who's good at sales, a CFO...and they do their job well. Who's going to help you figure out what path to take?

 I always say you need functions, but way more important, while hiring those functional people, you're getting something else whether you recognize it or not. You're getting somebody who's going to participate in all the more critical decisions. Even more critical than having engineering run smoothly, which is, "which path should we take?"

 Ask the following question: When I hire a VP of Engineering, how will it make my VP of Marketing better? That's by asking fundamental first principles questions.

 Elon Musk's book is worth reading, because he knew so little about building cars or launching rockets. He made so many screwy mistakes that could have been avoided. But if he had had all that experience and knowledge, he wouldn't even have attempted what he did. So I'd rather have somebody like Elon Musk who just fundamentally asks a lot of questions. Why are we doing it this way? Not that we've done it this way for 50 years at GM. Remember, GM spent more money and more time and started earlier, trying to do electric vehicles. And Elon, with no experience, by asking a lot of questions, and adapting rapidly as he screwed up - he got to a place where Tesla is a much more interesting company than GM today.

 When you look for that functional hire, realize you are also getting this other thing which is much more important. Look for how they think, how fast they change, how defensive they become when you challenge their idea versus saying, "Oh, why didn't I think about that?" Or, someone who is much more brainstorm-oriented. This sort of horizontal hiring that you get while doing functional-vertical hiring, is much more important and a much more important dimension. And by the way, most VCs don't look for it, because they have not gone through building companies themselves. So they give you bad advice.

 The number of times they say, "Go get somebody with distribution channel expertise." I'd rather have somebody figure it out and not have a source of inertia in this fluid, dynamic, evolving, thinking method in a company. So when you're hiring, more than anything, you're hiring a brain trust. They just happen to also have skills in doing functional jobs, but the brain trust is more important than the functional job, and evaluating them is very tricky. It's tricky because partly you don't know how to evaluate somebody outside your own area of competence.

 Do you call a VP of Marketing at Cisco? I don't mean to pick on Cisco, but it's a good company to pick on. They will look at marketing very, very differently than what a start-up needs. So they'll give you the wrong advice.

 This is where the tricky part comes. Not being overly-swayed by titles or position, and understanding why people did certain things and how thoughtful they were about it. What did they do that wasn't usual is very hard. It's also tricky to go back to this question of how you evaluate people. Most experienced interviewees know how to give you the right answer. They're going to fool you.

 You ask somebody, "Why did you do here?" They'll retroactively look at what was great about the strategy and say, "I did this." I've actually had people pitch me on why they did Sun's marketing strategy. Being very, very indirect - kind of a list of questions in my paper on hiring - the way you get to the bottom of things...

 I'll ask the same thing 3 different ways. I'll ask at the beginning, "What do you love doing? What are you good at?" Then while talking later about a job, I'll ask "What did you do really well there?" If those two don't coincide - if somebody's making it up - they won't. I'll ask that question 3 different ways. Third time I might ask, if you were to join this company, what would you think your strength would be? I've asked the same question 3 different ways in 3 different contexts. People who are making it up will make up 3 different answers. People who fundamentally have the truth will give the same right answer.

 Then at the end of the interview, I might ask "What would your boss say about you?" Now I've asked the same questions 4 times. Then I'll ask, "Hey can I get your last performance review that your boss did?" If they say no, I know they don't want to share their performance review. I have 100 tricks like this that I use. You can never rely on the direct answer you get when you interview someone.

 Then it's good to get together with each other and see what each person looked at. One of the dangers in hiring, is if I find like I'm rushed today and I'm interviewing somebody, I'll evaluate them different than if I'm doing it over coffee on a Saturday morning. Another thing to keep in mind, depending on how your day went or what mindset you are in, how you evaluate people changes. So for critical positions, I'll interview the same person 3 times on 3 different days, just so I am not biasing myself. Humans, including yourself and myself, are really biased. And so there are so many pitfalls, which is why doing a start-up is so hard.

 Anyway I'll let you ask more questions, I could go on for 6 more hours. 


AH: I'll ask one more question then open it up for Q&A. I know in that paper you have a lot of questions, but you also make the point that even if you follow all those processes, mistakes are inevitable. And at least 1/3 of the time, the hires don't work out. What, in your view, is the most common reason why the hires don't work out? How should a founder handle that? Sometimes start-ups wait too long to fix it.

 VK: If somebody wants a job, in the interview their job is to fool you. Question is, are they better at fooling you, or are you better at getting behind the words? I find, after hiring many, many people - I probably do hundreds of interviews a year - I have a success rate of about 65%. I'm wrong and I find most people less experience are wrong more often. Usually if you're a founder, the people you're interviewing are more experienced than you at fooling you. OK, so it is a game.

 It is really not being able to get behind the scenes and dig deep. So you should plan on that happening. I always tell my teams, if you want 4 people on your team, hire 6. Even make up a title you don't need. So I always overhire. Because you're never successful and hiring is a long process. To hire a critical person in a start-up, whether it's a critical technologist or a senior executive, it can easily take 6 months or longer. Then they take 6 months longer getting up to speed. So if you've made a mistake, you lose a year. And they will have hired other crappy people during that time, so it is very expensive.

 This is why you really need to watch new people, and you need to really evaluate how they're doing. This is where I think Board members and investors should be much more about assisting a venture. I never call myself a venture capitalist, I say venture assistant. Even our tagline for our firm says venture assistant.

 I like to do all this, and spend all my time doing all this. I never go to VC functions, I never go to financial conferences, I just spend my time with founders and people they're trying to hire. By the way, it's a lot more fun and you learn a lot more, and get excited about it. But you should overhire.

 Flip side, if you find somebody great, make up a title and hire them. Two things. No great talent is worth giving away because you don't have a slot. Make up a slot. They will pay for themselves. The flip side is, with few exceptions and there are some, a great talent is never too expensive to hire.

 I've gone through all kinds of ridiculous things. We're trying to hire an exec today, and I was arguing with another VC this morning. I said, so his company offered him this package and they're about to go public - why don't we guarantee him the first $15 million out of this start-up? Give him a preference over the preferred stock. I'll go to any length for great talent. Nobody heard of putting $15 million of preference for an employee, above the preferred stock of the investors' money. But I've done this more than once. So a couple things around that. The flip side is to plan for failure and evaluate new people.

 One other thing, when you hire somebody, try and assimilate them. People can fail because they're the wrong fit. But people can also fail because you didn't assimilate them well. If you have someone really experienced in a business, and you hired somebody from that business, they can advise you on how to assimilate this person. Get advice on how to assimilate people into your business, because they are going to be different. They are the diversity you're trying to build, and you don't want to be too uniform.

 So getting help in assimilating, getting lots of advice on how to make them successful, as opposed to reacting because they're acting so differently and doing things so differently - that's bad too. But then of course, if you make a mistake, you have to be quick. Good investors can help you with that.

 That, to me, is the principal value good investors bring. Anyway, let's open it up for questions for a few minutes.


Question: I'm going to ask about Space-X, Tesla, and then Theranos. Both of them are college dropouts, but why do you think one requires good domain expertise, and the other one doesn't?

 VK: They are different questions. One, I think...Stripe hired really awesome teams. In fact, the Strip co-founder is now running OpenAI. He hired somebody so strong, even though he was a college dropout, he hired awesome people around him.

 I don't think Theranos did this early. There's a culture thing we didn't talk about, and culture is very important. I think good companies acknowledge their problems and say, what do we need to solve this? Theranos, early on, tried to hide their problems. They didn't try to encourage discussion around their problems. I think that was the difference. I think Elizabeth has learned a lot, and if she were to do another start-up she would do much, much better. But sometimes, in trying to show the outside world you're a great start-up, you tend to now acknowledge your weaknesses and that's your danger.

 In fact, most start-ups spend way more time on their assets and not liabilities. Your liabilities can kill you faster than your assets can make you successful. Because you're a cohesive group when you are starting out, you tend to have groupthink and then tend to lose sight of your risks. You want to encourage that behaviour of asking a lot of questions. That's culture - we won't go in too much into that today - but establishing the right values...why people are here, what the mission is, what kind of culture.

 A number of people have made the point, having goodies in your lunch menu isn't culture. Allowing dogs in your workplace isn't culture, though I love dogs. Culture is about fundamental values, and how you behave, and what you encourage. The more people, at every level in the organization, that can ask questions, the more you will identify your risks sooner, the more you'll have a constructive culture and the more likely you'll increase your success.

 Let me finish that by saying one other thing. Sam said something else I really liked, which is start-ups are a pass/fail game.

 I like to say start-ups are not about dilution, they are about the probability of success. If you can increase the probability of success by hiring great people, by asking a lot of questions, by finding your risks early, by having people surface problems, you're going to increase your probability of success. Either you're successful or you're not. So it's about probability. People worry way too much about dilution, and not giving more equity to a co-founder or new hire who may by needed to made a new co-founder.

 I'm seeing a lot of this in the AI start-ups now. They're AI start-ups but they don't the AI talent because, hey, the average AI machine-learning person at Google - the median income I think is like $750,000 comp. What are you going to do get that person? You can pretend you can do without that, but you are fooling yourself.


Question: You touched on this earlier about spending Saturdays going through resumes. For early stages, could you go through what you're looking for? Let's say for a technical or early business person.

 VK: What do I look for? First, people who can think originally about their area, whatever their area is. Sometimes I just ask somebody what interesting papers they've read in their area of expertise. If it is narrow diversity, it tells me one thing. If it is broad diversity, it is one of my 100 questions. I'll ask what books have you read - it tells you how they think. That is critical to start-up in addition to knowing encryption algorithms.


Question: For an early stage start-up that is bootstrapping, do you see the value in leveraging technical advisors or advisory boards?

 VK: There is always value in leveraging technical advisors in an advisory board. What is important is to get their opinions but not respect it too much. You want them to surface the issues, but you as the founder, using first principles, want to say do I believe this or not, or how much should I weigh this versus other things?

 That is the critical part, and that is the hard part. That's what makes entrepreneurship hard. Because it is a lot of fun, but it is really hard too. I have a great presentation - I came up with the title in 1986 and it's still on our website - the entrepreneurial rollercoaster where the highs are high and the lows are low. And the lows are really low in every start-up.

 AH: Thank you for your time.